2010-01-05 02:06:57 +01:00
|
|
|
#----------------------------------------------------------------------
|
|
|
|
#
|
|
|
|
# Catalog.pm
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
# Perl module that extracts info from catalog files into Perl
|
2010-01-05 02:06:57 +01:00
|
|
|
# data structures
|
|
|
|
#
|
2023-01-02 21:00:37 +01:00
|
|
|
# Portions Copyright (c) 1996-2023, PostgreSQL Global Development Group
|
2010-01-05 02:06:57 +01:00
|
|
|
# Portions Copyright (c) 1994, Regents of the University of California
|
|
|
|
#
|
2010-09-20 22:08:53 +02:00
|
|
|
# src/backend/catalog/Catalog.pm
|
2010-01-05 02:06:57 +01:00
|
|
|
#
|
|
|
|
#----------------------------------------------------------------------
|
|
|
|
|
|
|
|
package Catalog;
|
|
|
|
|
|
|
|
use strict;
|
|
|
|
use warnings;
|
|
|
|
|
2018-05-04 00:06:45 +02:00
|
|
|
use File::Compare;
|
|
|
|
|
|
|
|
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
# Parses a catalog header file into a data structure describing the schema
|
|
|
|
# of the catalog.
|
|
|
|
sub ParseHeader
|
2010-01-05 02:06:57 +01:00
|
|
|
{
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
my $input_file = shift;
|
2012-07-05 03:47:49 +02:00
|
|
|
|
|
|
|
# There are a few types which are given one name in the C source, but a
|
|
|
|
# different name at the SQL level. These are enumerated here.
|
|
|
|
my %RENAME_ATTTYPE = (
|
2023-05-19 23:24:48 +02:00
|
|
|
'int16' => 'int2',
|
|
|
|
'int32' => 'int4',
|
|
|
|
'int64' => 'int8',
|
|
|
|
'Oid' => 'oid',
|
|
|
|
'NameData' => 'name',
|
2017-12-21 23:07:32 +01:00
|
|
|
'TransactionId' => 'xid',
|
2023-05-19 23:24:48 +02:00
|
|
|
'XLogRecPtr' => 'pg_lsn');
|
2012-07-05 03:47:49 +02:00
|
|
|
|
|
|
|
my %catalog;
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
my $declaring_attributes = 0;
|
2023-05-19 23:24:48 +02:00
|
|
|
my $is_varlen = 0;
|
|
|
|
my $is_client_code = 0;
|
2018-01-12 15:21:42 +01:00
|
|
|
|
2023-05-19 23:24:48 +02:00
|
|
|
$catalog{columns} = [];
|
|
|
|
$catalog{toasting} = [];
|
|
|
|
$catalog{indexing} = [];
|
|
|
|
$catalog{other_oids} = [];
|
Rethink method for assigning OIDs to the template0 and postgres DBs.
Commit aa0105141 assigned fixed OIDs to template0 and postgres
in a very ad-hoc way. Notably, instead of teaching Catalog.pm
about these OIDs, the unused_oids script was just hacked to
not show them as unused. That's problematic since, for example,
duplicate_oids wouldn't report any future conflict. Hence,
invent a macro DECLARE_OID_DEFINING_MACRO() that can be used to
define an OID that is known to Catalog.pm and will participate
in duplicate-detection as well as renumbering by renumber_oids.pl.
(We don't anticipate renumbering these particular OIDs, but we
might as well build out all the Catalog.pm infrastructure while
we're here.)
Another issue is that aa0105141 neglected to touch IsPinnedObject,
with the result that it now claimed template0 and postgres are
pinned. The right thing to do there seems to be to teach it that
no database is pinned, since in fact DROP DATABASE doesn't check
for pinned-ness (and at least for these cases, that is an
intentional choice). It's not clear whether this wrong answer
had any visible effect, but perhaps it could have resulted in
erroneous management of dependency entries.
In passing, rename the TemplateDbOid macro to Template1DbOid
to reduce confusion (likely we should have done that way back
when we invented template0, but we didn't), and rename the
OID macros for template0 and postgres to have a similar style.
There are no changes to postgres.bki here, so no need for a
catversion bump.
Discussion: https://postgr.es/m/2935358.1650479692@sss.pgh.pa.us
2022-04-21 22:23:12 +02:00
|
|
|
$catalog{foreign_keys} = [];
|
2023-05-19 23:24:48 +02:00
|
|
|
$catalog{client_code} = [];
|
2012-07-05 03:47:49 +02:00
|
|
|
|
2017-03-27 04:24:13 +02:00
|
|
|
open(my $ifh, '<', $input_file) || die "$input_file: $!";
|
2012-07-05 03:47:49 +02:00
|
|
|
|
|
|
|
# Scan the input file.
|
2017-03-27 04:24:13 +02:00
|
|
|
while (<$ifh>)
|
2012-07-05 03:47:49 +02:00
|
|
|
{
|
|
|
|
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
# Set appropriate flag when we're in certain code sections.
|
|
|
|
if (/^#/)
|
2018-04-25 20:00:19 +02:00
|
|
|
{
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
$is_varlen = 1 if /^#ifdef\s+CATALOG_VARLEN/;
|
|
|
|
if (/^#ifdef\s+EXPOSE_TO_CLIENT_CODE/)
|
2012-07-05 03:47:49 +02:00
|
|
|
{
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
$is_client_code = 1;
|
|
|
|
next;
|
2012-07-05 03:47:49 +02:00
|
|
|
}
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
next if !$is_client_code;
|
2018-04-25 20:00:19 +02:00
|
|
|
}
|
2012-07-05 03:47:49 +02:00
|
|
|
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
if (!$is_client_code)
|
2018-04-25 20:00:19 +02:00
|
|
|
{
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
# Strip C-style comments.
|
|
|
|
s;/\*(.|\n)*\*/;;g;
|
|
|
|
if (m;/\*;)
|
|
|
|
{
|
|
|
|
|
|
|
|
# handle multi-line comments properly.
|
|
|
|
my $next_line = <$ifh>;
|
|
|
|
die "$input_file: ends within C-style comment\n"
|
|
|
|
if !defined $next_line;
|
|
|
|
$_ .= $next_line;
|
|
|
|
redo;
|
2012-07-05 03:47:49 +02:00
|
|
|
}
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
|
|
|
|
# Strip useless whitespace and trailing semicolons.
|
2018-04-25 20:00:19 +02:00
|
|
|
chomp;
|
|
|
|
s/^\s+//;
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
s/;\s*$//;
|
|
|
|
s/\s+/ /g;
|
2018-04-25 20:00:19 +02:00
|
|
|
}
|
|
|
|
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
# Push the data into the appropriate data structure.
|
Create a script that can renumber manually-assigned OIDs.
This commit adds a Perl script renumber_oids.pl, which can reassign a
range of manually-assigned OIDs to someplace else by modifying OID
fields of the catalog *.dat files and OID-assigning macros in the
catalog *.h files.
Up to now, we've encouraged new patches that need manually-assigned
OIDs to use OIDs just above the range of existing OIDs. Predictably,
this leads to patches stepping on each others' toes, as whichever
one gets committed first creates an OID conflict that other patch
author(s) have to resolve manually. With the availability of
renumber_oids.pl, we can eliminate a lot of this hassle.
The new project policy, therefore, is:
* Encourage new patches to use high OIDs (the documentation suggests
choosing a block of OIDs at random in 8000..9999).
* After feature freeze in each development cycle, run renumber_oids.pl
to move all such OIDs down to lower numbers, thus freeing the high OID
range for the next development cycle.
This plan should greatly reduce the risk of OID collisions between
concurrently-developed patches. Also, if such a collision happens
anyway, we have the option to resolve it without much effort by doing
an off-schedule OID renumbering to get the first-committed patch out
of the way. Or a patch author could use renumber_oids.pl to change
their patch's assignments without much pain.
This approach does put a premium on not hard-wiring any OID values
in places where renumber_oids.pl and genbki.pl can't fix them.
Project practice in that respect seems to be pretty good already,
but a follow-on patch will sand down some rough edges.
John Naylor and Tom Lane, per an idea of Peter Geoghegan's
Discussion: https://postgr.es/m/CAH2-WzmMTGMcPuph4OvsO7Ykut0AOCF_i-=eaochT0dd2BN9CQ@mail.gmail.com
2019-03-12 15:50:48 +01:00
|
|
|
# Caution: when adding new recognized OID-defining macros,
|
|
|
|
# also update src/include/catalog/renumber_oids.pl.
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
if (/^DECLARE_TOAST\(\s*(\w+),\s*(\d+),\s*(\d+)\)/)
|
2018-04-25 20:00:19 +02:00
|
|
|
{
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
push @{ $catalog{toasting} },
|
2018-04-25 21:19:44 +02:00
|
|
|
{ parent_table => $1, toast_oid => $2, toast_index_oid => $3 };
|
2018-04-25 20:00:19 +02:00
|
|
|
}
|
2022-04-21 18:02:23 +02:00
|
|
|
elsif (
|
|
|
|
/^DECLARE_TOAST_WITH_MACRO\(\s*(\w+),\s*(\d+),\s*(\d+),\s*(\w+),\s*(\w+)\)/
|
|
|
|
)
|
|
|
|
{
|
|
|
|
push @{ $catalog{toasting} },
|
|
|
|
{
|
2023-05-19 23:24:48 +02:00
|
|
|
parent_table => $1,
|
|
|
|
toast_oid => $2,
|
|
|
|
toast_index_oid => $3,
|
|
|
|
toast_oid_macro => $4,
|
2022-04-21 18:02:23 +02:00
|
|
|
toast_index_oid_macro => $5
|
|
|
|
};
|
|
|
|
}
|
2021-01-30 19:14:31 +01:00
|
|
|
elsif (
|
2021-06-29 07:57:16 +02:00
|
|
|
/^DECLARE_(UNIQUE_)?INDEX(_PKEY)?\(\s*(\w+),\s*(\d+),\s*(\w+),\s*(.+)\)/
|
|
|
|
)
|
2018-04-25 20:00:19 +02:00
|
|
|
{
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
push @{ $catalog{indexing} },
|
2021-01-30 19:14:31 +01:00
|
|
|
{
|
2018-05-09 16:14:46 +02:00
|
|
|
is_unique => $1 ? 1 : 0,
|
2023-05-19 23:24:48 +02:00
|
|
|
is_pkey => $2 ? 1 : 0,
|
|
|
|
index_name => $3,
|
|
|
|
index_oid => $4,
|
2021-06-29 07:57:16 +02:00
|
|
|
index_oid_macro => $5,
|
2023-05-19 23:24:48 +02:00
|
|
|
index_decl => $6
|
2018-05-09 16:14:46 +02:00
|
|
|
};
|
2018-04-25 20:00:19 +02:00
|
|
|
}
|
Rethink method for assigning OIDs to the template0 and postgres DBs.
Commit aa0105141 assigned fixed OIDs to template0 and postgres
in a very ad-hoc way. Notably, instead of teaching Catalog.pm
about these OIDs, the unused_oids script was just hacked to
not show them as unused. That's problematic since, for example,
duplicate_oids wouldn't report any future conflict. Hence,
invent a macro DECLARE_OID_DEFINING_MACRO() that can be used to
define an OID that is known to Catalog.pm and will participate
in duplicate-detection as well as renumbering by renumber_oids.pl.
(We don't anticipate renumbering these particular OIDs, but we
might as well build out all the Catalog.pm infrastructure while
we're here.)
Another issue is that aa0105141 neglected to touch IsPinnedObject,
with the result that it now claimed template0 and postgres are
pinned. The right thing to do there seems to be to teach it that
no database is pinned, since in fact DROP DATABASE doesn't check
for pinned-ness (and at least for these cases, that is an
intentional choice). It's not clear whether this wrong answer
had any visible effect, but perhaps it could have resulted in
erroneous management of dependency entries.
In passing, rename the TemplateDbOid macro to Template1DbOid
to reduce confusion (likely we should have done that way back
when we invented template0, but we didn't), and rename the
OID macros for template0 and postgres to have a similar style.
There are no changes to postgres.bki here, so no need for a
catversion bump.
Discussion: https://postgr.es/m/2935358.1650479692@sss.pgh.pa.us
2022-04-21 22:23:12 +02:00
|
|
|
elsif (/^DECLARE_OID_DEFINING_MACRO\(\s*(\w+),\s*(\d+)\)/)
|
|
|
|
{
|
|
|
|
push @{ $catalog{other_oids} },
|
|
|
|
{
|
|
|
|
other_name => $1,
|
2023-05-19 23:24:48 +02:00
|
|
|
other_oid => $2
|
Rethink method for assigning OIDs to the template0 and postgres DBs.
Commit aa0105141 assigned fixed OIDs to template0 and postgres
in a very ad-hoc way. Notably, instead of teaching Catalog.pm
about these OIDs, the unused_oids script was just hacked to
not show them as unused. That's problematic since, for example,
duplicate_oids wouldn't report any future conflict. Hence,
invent a macro DECLARE_OID_DEFINING_MACRO() that can be used to
define an OID that is known to Catalog.pm and will participate
in duplicate-detection as well as renumbering by renumber_oids.pl.
(We don't anticipate renumbering these particular OIDs, but we
might as well build out all the Catalog.pm infrastructure while
we're here.)
Another issue is that aa0105141 neglected to touch IsPinnedObject,
with the result that it now claimed template0 and postgres are
pinned. The right thing to do there seems to be to teach it that
no database is pinned, since in fact DROP DATABASE doesn't check
for pinned-ness (and at least for these cases, that is an
intentional choice). It's not clear whether this wrong answer
had any visible effect, but perhaps it could have resulted in
erroneous management of dependency entries.
In passing, rename the TemplateDbOid macro to Template1DbOid
to reduce confusion (likely we should have done that way back
when we invented template0, but we didn't), and rename the
OID macros for template0 and postgres to have a similar style.
There are no changes to postgres.bki here, so no need for a
catversion bump.
Discussion: https://postgr.es/m/2935358.1650479692@sss.pgh.pa.us
2022-04-21 22:23:12 +02:00
|
|
|
};
|
|
|
|
}
|
Build in some knowledge about foreign-key relationships in the catalogs.
This follows in the spirit of commit dfb75e478, which created primary
key and uniqueness constraints to improve the visibility of constraints
imposed on the system catalogs. While our catalogs contain many
foreign-key-like relationships, they don't quite follow SQL semantics,
in that the convention for an omitted reference is to write zero not
NULL. Plus, we have some cases in which there are arrays each of whose
elements is supposed to be an FK reference; SQL has no way to model that.
So we can't create actual foreign key constraints to describe the
situation. Nonetheless, we can collect and use knowledge about these
relationships.
This patch therefore adds annotations to the catalog header files to
declare foreign-key relationships. (The BKI_LOOKUP annotations cover
simple cases, but we weren't previously distinguishing which such
columns are allowed to contain zeroes; we also need new markings for
multi-column FK references.) Then, Catalog.pm and genbki.pl are
taught to collect this information into a table in a new generated
header "system_fk_info.h". The only user of that at the moment is
a new SQL function pg_get_catalog_foreign_keys(), which exposes the
table to SQL. The oidjoins regression test is rewritten to use
pg_get_catalog_foreign_keys() to find out which columns to check.
Aside from removing the need for manual maintenance of that test
script, this allows it to cover numerous relationships that were not
checked by the old implementation based on findoidjoins. (As of this
commit, 217 relationships are checked by the test, versus 181 before.)
Discussion: https://postgr.es/m/3240355.1612129197@sss.pgh.pa.us
2021-02-02 23:11:55 +01:00
|
|
|
elsif (
|
|
|
|
/^DECLARE_(ARRAY_)?FOREIGN_KEY(_OPT)?\(\s*\(([^)]+)\),\s*(\w+),\s*\(([^)]+)\)\)/
|
|
|
|
)
|
|
|
|
{
|
|
|
|
push @{ $catalog{foreign_keys} },
|
|
|
|
{
|
|
|
|
is_array => $1 ? 1 : 0,
|
2023-05-19 23:24:48 +02:00
|
|
|
is_opt => $2 ? 1 : 0,
|
|
|
|
fk_cols => $3,
|
Build in some knowledge about foreign-key relationships in the catalogs.
This follows in the spirit of commit dfb75e478, which created primary
key and uniqueness constraints to improve the visibility of constraints
imposed on the system catalogs. While our catalogs contain many
foreign-key-like relationships, they don't quite follow SQL semantics,
in that the convention for an omitted reference is to write zero not
NULL. Plus, we have some cases in which there are arrays each of whose
elements is supposed to be an FK reference; SQL has no way to model that.
So we can't create actual foreign key constraints to describe the
situation. Nonetheless, we can collect and use knowledge about these
relationships.
This patch therefore adds annotations to the catalog header files to
declare foreign-key relationships. (The BKI_LOOKUP annotations cover
simple cases, but we weren't previously distinguishing which such
columns are allowed to contain zeroes; we also need new markings for
multi-column FK references.) Then, Catalog.pm and genbki.pl are
taught to collect this information into a table in a new generated
header "system_fk_info.h". The only user of that at the moment is
a new SQL function pg_get_catalog_foreign_keys(), which exposes the
table to SQL. The oidjoins regression test is rewritten to use
pg_get_catalog_foreign_keys() to find out which columns to check.
Aside from removing the need for manual maintenance of that test
script, this allows it to cover numerous relationships that were not
checked by the old implementation based on findoidjoins. (As of this
commit, 217 relationships are checked by the test, versus 181 before.)
Discussion: https://postgr.es/m/3240355.1612129197@sss.pgh.pa.us
2021-02-02 23:11:55 +01:00
|
|
|
pk_table => $4,
|
2023-05-19 23:24:48 +02:00
|
|
|
pk_cols => $5
|
Build in some knowledge about foreign-key relationships in the catalogs.
This follows in the spirit of commit dfb75e478, which created primary
key and uniqueness constraints to improve the visibility of constraints
imposed on the system catalogs. While our catalogs contain many
foreign-key-like relationships, they don't quite follow SQL semantics,
in that the convention for an omitted reference is to write zero not
NULL. Plus, we have some cases in which there are arrays each of whose
elements is supposed to be an FK reference; SQL has no way to model that.
So we can't create actual foreign key constraints to describe the
situation. Nonetheless, we can collect and use knowledge about these
relationships.
This patch therefore adds annotations to the catalog header files to
declare foreign-key relationships. (The BKI_LOOKUP annotations cover
simple cases, but we weren't previously distinguishing which such
columns are allowed to contain zeroes; we also need new markings for
multi-column FK references.) Then, Catalog.pm and genbki.pl are
taught to collect this information into a table in a new generated
header "system_fk_info.h". The only user of that at the moment is
a new SQL function pg_get_catalog_foreign_keys(), which exposes the
table to SQL. The oidjoins regression test is rewritten to use
pg_get_catalog_foreign_keys() to find out which columns to check.
Aside from removing the need for manual maintenance of that test
script, this allows it to cover numerous relationships that were not
checked by the old implementation based on findoidjoins. (As of this
commit, 217 relationships are checked by the test, versus 181 before.)
Discussion: https://postgr.es/m/3240355.1612129197@sss.pgh.pa.us
2021-02-02 23:11:55 +01:00
|
|
|
};
|
|
|
|
}
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
elsif (/^CATALOG\((\w+),(\d+),(\w+)\)/)
|
2018-04-25 20:00:19 +02:00
|
|
|
{
|
2023-05-19 23:24:48 +02:00
|
|
|
$catalog{catname} = $1;
|
|
|
|
$catalog{relation_oid} = $2;
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
$catalog{relation_oid_macro} = $3;
|
2018-04-25 20:00:19 +02:00
|
|
|
|
2012-07-05 03:47:49 +02:00
|
|
|
$catalog{bootstrap} = /BKI_BOOTSTRAP/ ? ' bootstrap' : '';
|
|
|
|
$catalog{shared_relation} =
|
|
|
|
/BKI_SHARED_RELATION/ ? ' shared_relation' : '';
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
if (/BKI_ROWTYPE_OID\((\d+),(\w+)\)/)
|
2012-07-05 03:47:49 +02:00
|
|
|
{
|
2023-05-19 23:24:48 +02:00
|
|
|
$catalog{rowtype_oid} = $1;
|
2012-07-05 03:47:49 +02:00
|
|
|
$catalog{rowtype_oid_clause} = " rowtype_oid $1";
|
2023-05-19 23:24:48 +02:00
|
|
|
$catalog{rowtype_oid_macro} = $2;
|
2012-07-05 03:47:49 +02:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2023-05-19 23:24:48 +02:00
|
|
|
$catalog{rowtype_oid} = '';
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
$catalog{rowtype_oid_clause} = '';
|
2023-05-19 23:24:48 +02:00
|
|
|
$catalog{rowtype_oid_macro} = '';
|
2012-07-05 03:47:49 +02:00
|
|
|
}
|
2017-12-21 23:07:32 +01:00
|
|
|
$catalog{schema_macro} = /BKI_SCHEMA_MACRO/ ? 1 : 0;
|
2012-07-05 03:47:49 +02:00
|
|
|
$declaring_attributes = 1;
|
2018-04-25 20:00:19 +02:00
|
|
|
}
|
2012-07-05 03:47:49 +02:00
|
|
|
elsif ($is_client_code)
|
2018-04-25 20:00:19 +02:00
|
|
|
{
|
2012-07-05 03:47:49 +02:00
|
|
|
if (/^#endif/)
|
|
|
|
{
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
$is_client_code = 0;
|
2012-07-05 03:47:49 +02:00
|
|
|
}
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
else
|
2012-07-05 03:47:49 +02:00
|
|
|
{
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
push @{ $catalog{client_code} }, $_;
|
2012-07-05 03:47:49 +02:00
|
|
|
}
|
2018-04-25 20:00:19 +02:00
|
|
|
}
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
elsif ($declaring_attributes)
|
2018-04-25 20:00:19 +02:00
|
|
|
{
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
next if (/^{|^$/);
|
2012-07-05 03:47:49 +02:00
|
|
|
if (/^}/)
|
|
|
|
{
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
$declaring_attributes = 0;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2017-12-21 23:07:32 +01:00
|
|
|
my %column;
|
2018-01-12 15:21:42 +01:00
|
|
|
my @attopts = split /\s+/, $_;
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
my $atttype = shift @attopts;
|
2018-01-12 15:21:42 +01:00
|
|
|
my $attname = shift @attopts;
|
|
|
|
die "parse error ($input_file)"
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
unless ($attname and $atttype);
|
2018-04-25 20:00:19 +02:00
|
|
|
|
2012-07-05 03:47:49 +02:00
|
|
|
if (exists $RENAME_ATTTYPE{$atttype})
|
|
|
|
{
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
$atttype = $RENAME_ATTTYPE{$atttype};
|
2012-07-05 03:47:49 +02:00
|
|
|
}
|
2018-04-25 20:00:19 +02:00
|
|
|
|
2012-07-05 03:47:49 +02:00
|
|
|
# If the C name ends with '[]' or '[digits]', we have
|
Rationalize handling of array type names in bootstrap data.
Formerly, Catalog.pm turned a C array type declaration in the catalog
header files into a SQL type, e.g., 'foo[]'. Along the way, genbki.pl
turned this into '_foo' for the purpose of type lookups, but wrote 'foo[]'
to postgres.bki. During bootstrap, bootscanner.l had to have a special
case rule to tokenize this, and then MapArrayTypeName() would turn 'foo[]'
into '_foo' one more time.
This seems unnecessarily complicated, especially since nobody cares that
much about the readability of postgres.bki. Instead, make Catalog.pm
convert the C declaration into '_foo' to start with, and preserve that
representation of the type name throughout bootstrap data processing.
Then rip out the special-case code in bootscanner.l and bootstrap.c.
This changes postgres.bki to the extent that array fields are now
declared like
proconfig = _text ,
rather than
proconfig = text[] ,
No documentation update, since the SGML docs didn't mention any of this
in the first place, and it's all pretty transparent to writers of
catalog header files anyway.
John Naylor
Discussion: https://postgr.es/m/CAJVSVGUNao=-Q2-vAN3PYcdF5tnL5JAHwGwzZGuYHtq+Mk_9ng@mail.gmail.com
2018-04-18 00:29:11 +02:00
|
|
|
# an array type, so we discard that from the name and
|
|
|
|
# prepend '_' to the type.
|
|
|
|
if ($attname =~ /(\w+)\[\d*\]/)
|
2012-07-05 03:47:49 +02:00
|
|
|
{
|
2018-01-12 15:21:42 +01:00
|
|
|
$attname = $1;
|
|
|
|
$atttype = '_' . $atttype;
|
2018-04-25 20:00:19 +02:00
|
|
|
}
|
|
|
|
|
2023-05-19 23:24:48 +02:00
|
|
|
$column{type} = $atttype;
|
|
|
|
$column{name} = $attname;
|
2018-01-12 15:21:42 +01:00
|
|
|
$column{is_varlen} = 1 if $is_varlen;
|
2018-04-25 20:00:19 +02:00
|
|
|
|
2018-01-12 15:21:42 +01:00
|
|
|
foreach my $attopt (@attopts)
|
2018-04-25 20:00:19 +02:00
|
|
|
{
|
2012-07-05 03:47:49 +02:00
|
|
|
if ($attopt eq 'BKI_FORCE_NULL')
|
|
|
|
{
|
|
|
|
$column{forcenull} = 1;
|
|
|
|
}
|
Rationalize handling of array type names in bootstrap data.
Formerly, Catalog.pm turned a C array type declaration in the catalog
header files into a SQL type, e.g., 'foo[]'. Along the way, genbki.pl
turned this into '_foo' for the purpose of type lookups, but wrote 'foo[]'
to postgres.bki. During bootstrap, bootscanner.l had to have a special
case rule to tokenize this, and then MapArrayTypeName() would turn 'foo[]'
into '_foo' one more time.
This seems unnecessarily complicated, especially since nobody cares that
much about the readability of postgres.bki. Instead, make Catalog.pm
convert the C declaration into '_foo' to start with, and preserve that
representation of the type name throughout bootstrap data processing.
Then rip out the special-case code in bootscanner.l and bootstrap.c.
This changes postgres.bki to the extent that array fields are now
declared like
proconfig = _text ,
rather than
proconfig = text[] ,
No documentation update, since the SGML docs didn't mention any of this
in the first place, and it's all pretty transparent to writers of
catalog header files anyway.
John Naylor
Discussion: https://postgr.es/m/CAJVSVGUNao=-Q2-vAN3PYcdF5tnL5JAHwGwzZGuYHtq+Mk_9ng@mail.gmail.com
2018-04-18 00:29:11 +02:00
|
|
|
elsif ($attopt eq 'BKI_FORCE_NOT_NULL')
|
2012-07-05 03:47:49 +02:00
|
|
|
{
|
|
|
|
$column{forcenotnull} = 1;
|
|
|
|
}
|
2015-02-21 22:25:49 +01:00
|
|
|
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
# We use quotes for values like \0 and \054, to
|
|
|
|
# make sure all compilers and syntax highlighters
|
2017-12-21 23:07:32 +01:00
|
|
|
# can recognize them properly.
|
|
|
|
elsif ($attopt =~ /BKI_DEFAULT\(['"]?([^'"]+)['"]?\)/)
|
2018-04-25 20:00:19 +02:00
|
|
|
{
|
2017-12-21 23:07:32 +01:00
|
|
|
$column{default} = $1;
|
2018-04-25 20:00:19 +02:00
|
|
|
}
|
2018-09-20 21:14:46 +02:00
|
|
|
elsif (
|
|
|
|
$attopt =~ /BKI_ARRAY_DEFAULT\(['"]?([^'"]+)['"]?\)/)
|
|
|
|
{
|
|
|
|
$column{array_default} = $1;
|
|
|
|
}
|
Build in some knowledge about foreign-key relationships in the catalogs.
This follows in the spirit of commit dfb75e478, which created primary
key and uniqueness constraints to improve the visibility of constraints
imposed on the system catalogs. While our catalogs contain many
foreign-key-like relationships, they don't quite follow SQL semantics,
in that the convention for an omitted reference is to write zero not
NULL. Plus, we have some cases in which there are arrays each of whose
elements is supposed to be an FK reference; SQL has no way to model that.
So we can't create actual foreign key constraints to describe the
situation. Nonetheless, we can collect and use knowledge about these
relationships.
This patch therefore adds annotations to the catalog header files to
declare foreign-key relationships. (The BKI_LOOKUP annotations cover
simple cases, but we weren't previously distinguishing which such
columns are allowed to contain zeroes; we also need new markings for
multi-column FK references.) Then, Catalog.pm and genbki.pl are
taught to collect this information into a table in a new generated
header "system_fk_info.h". The only user of that at the moment is
a new SQL function pg_get_catalog_foreign_keys(), which exposes the
table to SQL. The oidjoins regression test is rewritten to use
pg_get_catalog_foreign_keys() to find out which columns to check.
Aside from removing the need for manual maintenance of that test
script, this allows it to cover numerous relationships that were not
checked by the old implementation based on findoidjoins. (As of this
commit, 217 relationships are checked by the test, versus 181 before.)
Discussion: https://postgr.es/m/3240355.1612129197@sss.pgh.pa.us
2021-02-02 23:11:55 +01:00
|
|
|
elsif ($attopt =~ /BKI_LOOKUP(_OPT)?\((\w+)\)/)
|
2018-04-25 20:00:19 +02:00
|
|
|
{
|
Build in some knowledge about foreign-key relationships in the catalogs.
This follows in the spirit of commit dfb75e478, which created primary
key and uniqueness constraints to improve the visibility of constraints
imposed on the system catalogs. While our catalogs contain many
foreign-key-like relationships, they don't quite follow SQL semantics,
in that the convention for an omitted reference is to write zero not
NULL. Plus, we have some cases in which there are arrays each of whose
elements is supposed to be an FK reference; SQL has no way to model that.
So we can't create actual foreign key constraints to describe the
situation. Nonetheless, we can collect and use knowledge about these
relationships.
This patch therefore adds annotations to the catalog header files to
declare foreign-key relationships. (The BKI_LOOKUP annotations cover
simple cases, but we weren't previously distinguishing which such
columns are allowed to contain zeroes; we also need new markings for
multi-column FK references.) Then, Catalog.pm and genbki.pl are
taught to collect this information into a table in a new generated
header "system_fk_info.h". The only user of that at the moment is
a new SQL function pg_get_catalog_foreign_keys(), which exposes the
table to SQL. The oidjoins regression test is rewritten to use
pg_get_catalog_foreign_keys() to find out which columns to check.
Aside from removing the need for manual maintenance of that test
script, this allows it to cover numerous relationships that were not
checked by the old implementation based on findoidjoins. (As of this
commit, 217 relationships are checked by the test, versus 181 before.)
Discussion: https://postgr.es/m/3240355.1612129197@sss.pgh.pa.us
2021-02-02 23:11:55 +01:00
|
|
|
$column{lookup} = $2;
|
|
|
|
$column{lookup_opt} = $1 ? 1 : 0;
|
|
|
|
# BKI_LOOKUP implicitly makes an FK reference
|
|
|
|
push @{ $catalog{foreign_keys} },
|
|
|
|
{
|
2023-05-19 23:24:48 +02:00
|
|
|
is_array => (
|
|
|
|
$atttype eq 'oidvector' || $atttype eq '_oid')
|
Build in some knowledge about foreign-key relationships in the catalogs.
This follows in the spirit of commit dfb75e478, which created primary
key and uniqueness constraints to improve the visibility of constraints
imposed on the system catalogs. While our catalogs contain many
foreign-key-like relationships, they don't quite follow SQL semantics,
in that the convention for an omitted reference is to write zero not
NULL. Plus, we have some cases in which there are arrays each of whose
elements is supposed to be an FK reference; SQL has no way to model that.
So we can't create actual foreign key constraints to describe the
situation. Nonetheless, we can collect and use knowledge about these
relationships.
This patch therefore adds annotations to the catalog header files to
declare foreign-key relationships. (The BKI_LOOKUP annotations cover
simple cases, but we weren't previously distinguishing which such
columns are allowed to contain zeroes; we also need new markings for
multi-column FK references.) Then, Catalog.pm and genbki.pl are
taught to collect this information into a table in a new generated
header "system_fk_info.h". The only user of that at the moment is
a new SQL function pg_get_catalog_foreign_keys(), which exposes the
table to SQL. The oidjoins regression test is rewritten to use
pg_get_catalog_foreign_keys() to find out which columns to check.
Aside from removing the need for manual maintenance of that test
script, this allows it to cover numerous relationships that were not
checked by the old implementation based on findoidjoins. (As of this
commit, 217 relationships are checked by the test, versus 181 before.)
Discussion: https://postgr.es/m/3240355.1612129197@sss.pgh.pa.us
2021-02-02 23:11:55 +01:00
|
|
|
? 1
|
|
|
|
: 0,
|
2023-05-19 23:24:48 +02:00
|
|
|
is_opt => $column{lookup_opt},
|
|
|
|
fk_cols => $attname,
|
Build in some knowledge about foreign-key relationships in the catalogs.
This follows in the spirit of commit dfb75e478, which created primary
key and uniqueness constraints to improve the visibility of constraints
imposed on the system catalogs. While our catalogs contain many
foreign-key-like relationships, they don't quite follow SQL semantics,
in that the convention for an omitted reference is to write zero not
NULL. Plus, we have some cases in which there are arrays each of whose
elements is supposed to be an FK reference; SQL has no way to model that.
So we can't create actual foreign key constraints to describe the
situation. Nonetheless, we can collect and use knowledge about these
relationships.
This patch therefore adds annotations to the catalog header files to
declare foreign-key relationships. (The BKI_LOOKUP annotations cover
simple cases, but we weren't previously distinguishing which such
columns are allowed to contain zeroes; we also need new markings for
multi-column FK references.) Then, Catalog.pm and genbki.pl are
taught to collect this information into a table in a new generated
header "system_fk_info.h". The only user of that at the moment is
a new SQL function pg_get_catalog_foreign_keys(), which exposes the
table to SQL. The oidjoins regression test is rewritten to use
pg_get_catalog_foreign_keys() to find out which columns to check.
Aside from removing the need for manual maintenance of that test
script, this allows it to cover numerous relationships that were not
checked by the old implementation based on findoidjoins. (As of this
commit, 217 relationships are checked by the test, versus 181 before.)
Discussion: https://postgr.es/m/3240355.1612129197@sss.pgh.pa.us
2021-02-02 23:11:55 +01:00
|
|
|
pk_table => $column{lookup},
|
2023-05-19 23:24:48 +02:00
|
|
|
pk_cols => 'oid'
|
Build in some knowledge about foreign-key relationships in the catalogs.
This follows in the spirit of commit dfb75e478, which created primary
key and uniqueness constraints to improve the visibility of constraints
imposed on the system catalogs. While our catalogs contain many
foreign-key-like relationships, they don't quite follow SQL semantics,
in that the convention for an omitted reference is to write zero not
NULL. Plus, we have some cases in which there are arrays each of whose
elements is supposed to be an FK reference; SQL has no way to model that.
So we can't create actual foreign key constraints to describe the
situation. Nonetheless, we can collect and use knowledge about these
relationships.
This patch therefore adds annotations to the catalog header files to
declare foreign-key relationships. (The BKI_LOOKUP annotations cover
simple cases, but we weren't previously distinguishing which such
columns are allowed to contain zeroes; we also need new markings for
multi-column FK references.) Then, Catalog.pm and genbki.pl are
taught to collect this information into a table in a new generated
header "system_fk_info.h". The only user of that at the moment is
a new SQL function pg_get_catalog_foreign_keys(), which exposes the
table to SQL. The oidjoins regression test is rewritten to use
pg_get_catalog_foreign_keys() to find out which columns to check.
Aside from removing the need for manual maintenance of that test
script, this allows it to cover numerous relationships that were not
checked by the old implementation based on findoidjoins. (As of this
commit, 217 relationships are checked by the test, versus 181 before.)
Discussion: https://postgr.es/m/3240355.1612129197@sss.pgh.pa.us
2021-02-02 23:11:55 +01:00
|
|
|
};
|
2018-04-25 20:00:19 +02:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
die
|
2018-05-19 22:04:47 +02:00
|
|
|
"unknown or misformatted column option $attopt on column $attname";
|
2018-01-12 15:21:42 +01:00
|
|
|
}
|
2015-02-21 22:25:49 +01:00
|
|
|
|
2018-01-12 15:21:42 +01:00
|
|
|
if ($column{forcenull} and $column{forcenotnull})
|
2015-02-21 22:25:49 +01:00
|
|
|
{
|
|
|
|
die "$attname is forced both null and not null";
|
|
|
|
}
|
2012-07-05 03:47:49 +02:00
|
|
|
}
|
2017-12-21 23:07:32 +01:00
|
|
|
push @{ $catalog{columns} }, \%column;
|
2012-07-05 03:47:49 +02:00
|
|
|
}
|
|
|
|
}
|
2018-04-25 20:00:19 +02:00
|
|
|
}
|
2017-03-27 04:24:13 +02:00
|
|
|
close $ifh;
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
return \%catalog;
|
2010-01-05 02:06:57 +01:00
|
|
|
}
|
|
|
|
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
# Parses a file containing Perl data structure literals, returning live data.
|
|
|
|
#
|
|
|
|
# The parameter $preserve_formatting needs to be set for callers that want
|
|
|
|
# to work with non-data lines in the data files, such as comments and blank
|
|
|
|
# lines. If a caller just wants to consume the data, leave it unset.
|
|
|
|
sub ParseData
|
Move bootstrap-time lookup of regproc OIDs into genbki.pl.
Formerly, the bootstrap backend looked up the OIDs corresponding to
names in regproc catalog entries using brute-force searches of pg_proc.
It was somewhat remarkable that that worked at all, since it was used
while populating other pretty-fundamental catalogs like pg_operator.
And it was also quite slow, and getting slower as pg_proc gets bigger.
This patch moves the lookup work into genbki.pl, so that the values in
postgres.bki for regproc columns are always numeric OIDs, an option
that regprocin() already supported. Perl isn't the world's speediest
language, so this about doubles the time needed to run genbki.pl (from
0.3 to 0.6 sec on my machine). But we only do that at most once per
build. The time needed to run initdb drops significantly --- on my
machine, initdb --no-sync goes from 1.8 to 1.3 seconds. So this is
a small net win even for just one initdb per build, and it becomes
quite a nice win for test sequences requiring many initdb runs.
Strip out the now-dead code for brute-force catalog searching in
regprocin. We'd also cargo-culted similar logic into regoperin
and some (not all) of the other reg*in functions. That is all
dead code too since we currently have no need to load such values
during bootstrap. I removed it all, reasoning that if we ever
need such functionality it'd be much better to do it in a similar
way to this patch.
There might be some simplifications possible in the backend now that
regprocin doesn't require doing catalog reads so early in bootstrap.
I've not looked into that, though.
Andreas Karlsson, with some small adjustments by me
Discussion: https://postgr.es/m/30896.1492006367@sss.pgh.pa.us
2017-04-13 18:07:47 +02:00
|
|
|
{
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
my ($input_file, $schema, $preserve_formatting) = @_;
|
|
|
|
|
|
|
|
open(my $ifd, '<', $input_file) || die "$input_file: $!";
|
|
|
|
$input_file =~ /(\w+)\.dat$/
|
2018-04-19 00:17:02 +02:00
|
|
|
or die "Input file $input_file needs to be a .dat file.\n";
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
my $catname = $1;
|
2023-05-19 23:24:48 +02:00
|
|
|
my $data = [];
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
|
2022-09-08 08:23:13 +02:00
|
|
|
if ($preserve_formatting)
|
|
|
|
{
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
# Scan the input file.
|
|
|
|
while (<$ifd>)
|
|
|
|
{
|
|
|
|
my $hash_ref;
|
2022-09-08 08:54:14 +02:00
|
|
|
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
if (/{/)
|
|
|
|
{
|
|
|
|
# Capture the hash ref
|
2018-05-07 21:21:00 +02:00
|
|
|
# NB: Assumes that the next hash ref can't start on the
|
|
|
|
# same line where the present one ended.
|
|
|
|
# Not foolproof, but we shouldn't need a full parser,
|
|
|
|
# since we expect relatively well-behaved input.
|
2022-09-08 08:54:14 +02:00
|
|
|
|
2018-05-07 21:21:00 +02:00
|
|
|
# Quick hack to detect when we have a full hash ref to
|
|
|
|
# parse. We can't just use a regex because of values in
|
|
|
|
# pg_aggregate and pg_proc like '{0,0}'. This will need
|
|
|
|
# work if we ever need to allow unbalanced braces within
|
2018-05-09 16:04:17 +02:00
|
|
|
# a field value.
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
my $lcnt = tr/{//;
|
2018-05-09 16:04:17 +02:00
|
|
|
my $rcnt = tr/}//;
|
2022-09-08 08:54:14 +02:00
|
|
|
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
if ($lcnt == $rcnt)
|
|
|
|
{
|
2012-07-05 03:47:49 +02:00
|
|
|
# We're treating the input line as a piece of Perl, so we
|
2022-09-08 08:23:13 +02:00
|
|
|
# need to use string eval here. Tell perlcritic we know what
|
2018-05-07 21:21:00 +02:00
|
|
|
# we're doing.
|
2022-09-12 06:59:43 +02:00
|
|
|
eval "\$hash_ref = $_"; ## no critic (ProhibitStringyEval)
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
if (!ref $hash_ref)
|
2022-09-08 08:54:14 +02:00
|
|
|
{
|
2018-04-19 00:17:02 +02:00
|
|
|
die "$input_file: error parsing line $.:\n$_\n";
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
}
|
2022-09-08 08:54:14 +02:00
|
|
|
|
2018-04-19 00:17:02 +02:00
|
|
|
# Annotate each hash with the source line number.
|
|
|
|
$hash_ref->{line_number} = $.;
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
|
2018-04-19 00:17:02 +02:00
|
|
|
# Expand tuples to their full representation.
|
|
|
|
AddDefaultValues($hash_ref, $schema, $catname);
|
2022-09-08 08:54:14 +02:00
|
|
|
}
|
2018-04-19 00:17:02 +02:00
|
|
|
else
|
2022-09-08 08:54:14 +02:00
|
|
|
{
|
2018-04-19 00:17:02 +02:00
|
|
|
my $next_line = <$ifd>;
|
|
|
|
die "$input_file: file ends within Perl hash\n"
|
|
|
|
if !defined $next_line;
|
|
|
|
$_ .= $next_line;
|
|
|
|
redo;
|
|
|
|
}
|
2022-09-08 08:54:14 +02:00
|
|
|
}
|
2018-04-19 00:17:02 +02:00
|
|
|
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
# If we found a hash reference, keep it, unless it is marked as
|
|
|
|
# autogenerated; in that case it'd duplicate an entry we'll
|
|
|
|
# autogenerate below. (This makes it safe for reformat_dat_file.pl
|
|
|
|
# with --full-tuples to print autogenerated entries, which seems like
|
|
|
|
# useful behavior for debugging.)
|
2022-09-08 08:54:14 +02:00
|
|
|
#
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
# Otherwise, we have a non-data string, which we need to keep in
|
|
|
|
# order to preserve formatting.
|
|
|
|
if (defined $hash_ref)
|
2022-09-08 08:54:14 +02:00
|
|
|
{
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
push @$data, $hash_ref if !$hash_ref->{autogenerated};
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2018-04-19 00:17:02 +02:00
|
|
|
push @$data, $_;
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
}
|
|
|
|
}
|
2022-09-08 08:23:13 +02:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
# When we only care about the contents, it's faster to read and eval
|
|
|
|
# the whole file at once.
|
|
|
|
local $/;
|
|
|
|
my $full_file = <$ifd>;
|
2022-09-12 06:59:43 +02:00
|
|
|
eval "\$data = $full_file" ## no critic (ProhibitStringyEval)
|
2022-09-08 08:23:13 +02:00
|
|
|
or die "error parsing $input_file\n";
|
|
|
|
foreach my $hash_ref (@{$data})
|
|
|
|
{
|
|
|
|
AddDefaultValues($hash_ref, $schema, $catname);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
close $ifd;
|
2018-09-20 21:14:46 +02:00
|
|
|
|
|
|
|
# If this is pg_type, auto-generate array types too.
|
|
|
|
GenerateArrayTypes($schema, $data) if $catname eq 'pg_type';
|
|
|
|
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
return $data;
|
Move bootstrap-time lookup of regproc OIDs into genbki.pl.
Formerly, the bootstrap backend looked up the OIDs corresponding to
names in regproc catalog entries using brute-force searches of pg_proc.
It was somewhat remarkable that that worked at all, since it was used
while populating other pretty-fundamental catalogs like pg_operator.
And it was also quite slow, and getting slower as pg_proc gets bigger.
This patch moves the lookup work into genbki.pl, so that the values in
postgres.bki for regproc columns are always numeric OIDs, an option
that regprocin() already supported. Perl isn't the world's speediest
language, so this about doubles the time needed to run genbki.pl (from
0.3 to 0.6 sec on my machine). But we only do that at most once per
build. The time needed to run initdb drops significantly --- on my
machine, initdb --no-sync goes from 1.8 to 1.3 seconds. So this is
a small net win even for just one initdb per build, and it becomes
quite a nice win for test sequences requiring many initdb runs.
Strip out the now-dead code for brute-force catalog searching in
regprocin. We'd also cargo-culted similar logic into regoperin
and some (not all) of the other reg*in functions. That is all
dead code too since we currently have no need to load such values
during bootstrap. I removed it all, reasoning that if we ever
need such functionality it'd be much better to do it in a similar
way to this patch.
There might be some simplifications possible in the backend now that
regprocin doesn't require doing catalog reads so early in bootstrap.
I've not looked into that, though.
Andreas Karlsson, with some small adjustments by me
Discussion: https://postgr.es/m/30896.1492006367@sss.pgh.pa.us
2017-04-13 18:07:47 +02:00
|
|
|
}
|
|
|
|
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
# Fill in default values of a record using the given schema.
|
|
|
|
# It's the caller's responsibility to specify other values beforehand.
|
2018-01-12 15:21:42 +01:00
|
|
|
sub AddDefaultValues
|
|
|
|
{
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
my ($row, $schema, $catname) = @_;
|
2018-01-12 15:21:42 +01:00
|
|
|
my @missing_fields;
|
|
|
|
|
2018-04-28 21:27:16 +02:00
|
|
|
# Compute special-case column values.
|
|
|
|
# Note: If you add new cases here, you must also teach
|
|
|
|
# strip_default_values() in include/catalog/reformat_dat_file.pl
|
|
|
|
# to delete them.
|
|
|
|
if ($catname eq 'pg_proc')
|
|
|
|
{
|
|
|
|
# pg_proc.pronargs can be derived from proargtypes.
|
|
|
|
if (defined $row->{proargtypes})
|
|
|
|
{
|
|
|
|
my @proargtypes = split /\s+/, $row->{proargtypes};
|
|
|
|
$row->{pronargs} = scalar(@proargtypes);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
# Now fill in defaults, and note any columns that remain undefined.
|
2018-01-12 15:21:42 +01:00
|
|
|
foreach my $column (@$schema)
|
|
|
|
{
|
|
|
|
my $attname = $column->{name};
|
|
|
|
|
2020-02-15 21:13:44 +01:00
|
|
|
# No work if field already has a value.
|
|
|
|
next if defined $row->{$attname};
|
|
|
|
|
|
|
|
# Ignore 'oid' columns, they're handled elsewhere.
|
|
|
|
next if $attname eq 'oid';
|
|
|
|
|
|
|
|
# If column has a default value, fill that in.
|
|
|
|
if (defined $column->{default})
|
2018-01-12 15:21:42 +01:00
|
|
|
{
|
|
|
|
$row->{$attname} = $column->{default};
|
2020-02-15 21:13:44 +01:00
|
|
|
next;
|
2018-01-12 15:21:42 +01:00
|
|
|
}
|
2020-02-15 21:13:44 +01:00
|
|
|
|
|
|
|
# Failed to find a value for this field.
|
|
|
|
push @missing_fields, $attname;
|
2018-01-12 15:21:42 +01:00
|
|
|
}
|
|
|
|
|
2018-04-28 21:27:16 +02:00
|
|
|
# Failure to provide all columns is a hard error.
|
2018-01-12 15:21:42 +01:00
|
|
|
if (@missing_fields)
|
|
|
|
{
|
2018-04-19 00:17:02 +02:00
|
|
|
die sprintf "missing values for field(s) %s in %s.dat line %s\n",
|
|
|
|
join(', ', @missing_fields), $catname, $row->{line_number};
|
2018-01-12 15:21:42 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2018-09-20 21:14:46 +02:00
|
|
|
# If a pg_type entry has an array_type_oid metadata field,
|
|
|
|
# auto-generate an entry for its array type.
|
|
|
|
sub GenerateArrayTypes
|
|
|
|
{
|
|
|
|
my $pgtype_schema = shift;
|
2023-05-19 23:24:48 +02:00
|
|
|
my $types = shift;
|
2018-09-20 21:14:46 +02:00
|
|
|
my @array_types;
|
|
|
|
|
|
|
|
foreach my $elem_type (@$types)
|
|
|
|
{
|
|
|
|
next if !(ref $elem_type eq 'HASH');
|
|
|
|
next if !defined($elem_type->{array_type_oid});
|
|
|
|
|
|
|
|
my %array_type;
|
|
|
|
|
|
|
|
# Set up metadata fields for array type.
|
2023-05-19 23:24:48 +02:00
|
|
|
$array_type{oid} = $elem_type->{array_type_oid};
|
2018-09-20 21:14:46 +02:00
|
|
|
$array_type{autogenerated} = 1;
|
2023-05-19 23:24:48 +02:00
|
|
|
$array_type{line_number} = $elem_type->{line_number};
|
2018-09-20 21:14:46 +02:00
|
|
|
|
|
|
|
# Set up column values derived from the element type.
|
|
|
|
$array_type{typname} = '_' . $elem_type->{typname};
|
|
|
|
$array_type{typelem} = $elem_type->{typname};
|
|
|
|
|
|
|
|
# Arrays require INT alignment, unless the element type requires
|
|
|
|
# DOUBLE alignment.
|
|
|
|
$array_type{typalign} = $elem_type->{typalign} eq 'd' ? 'd' : 'i';
|
|
|
|
|
|
|
|
# Fill in the rest of the array entry's fields.
|
|
|
|
foreach my $column (@$pgtype_schema)
|
|
|
|
{
|
|
|
|
my $attname = $column->{name};
|
|
|
|
|
|
|
|
# Skip if we already set it above.
|
|
|
|
next if defined $array_type{$attname};
|
|
|
|
|
|
|
|
# Apply the BKI_ARRAY_DEFAULT setting if there is one,
|
|
|
|
# otherwise copy the field from the element type.
|
|
|
|
if (defined $column->{array_default})
|
|
|
|
{
|
|
|
|
$array_type{$attname} = $column->{array_default};
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
$array_type{$attname} = $elem_type->{$attname};
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
# Lastly, cross-link the array to the element type.
|
|
|
|
$elem_type->{typarray} = $array_type{typname};
|
|
|
|
|
|
|
|
push @array_types, \%array_type;
|
|
|
|
}
|
|
|
|
|
|
|
|
push @$types, @array_types;
|
2018-09-20 22:10:23 +02:00
|
|
|
|
|
|
|
return;
|
2018-09-20 21:14:46 +02:00
|
|
|
}
|
|
|
|
|
2010-01-05 03:34:03 +01:00
|
|
|
# Rename temporary files to final names.
|
2018-05-04 00:06:45 +02:00
|
|
|
# Call this function with the final file name and the .tmp extension.
|
|
|
|
#
|
|
|
|
# If the final file already exists and has identical contents, don't
|
|
|
|
# overwrite it; this behavior avoids unnecessary recompiles due to
|
|
|
|
# updating the mod date on unchanged header files.
|
|
|
|
#
|
2010-01-05 21:23:32 +01:00
|
|
|
# Note: recommended extension is ".tmp$$", so that parallel make steps
|
2018-05-04 00:06:45 +02:00
|
|
|
# can't use the same temp files.
|
2010-01-05 02:06:57 +01:00
|
|
|
sub RenameTempFile
|
|
|
|
{
|
2012-07-05 03:47:49 +02:00
|
|
|
my $final_name = shift;
|
2023-05-19 23:24:48 +02:00
|
|
|
my $extension = shift;
|
|
|
|
my $temp_name = $final_name . $extension;
|
2018-05-04 00:06:45 +02:00
|
|
|
|
|
|
|
if (-f $final_name
|
|
|
|
&& compare($temp_name, $final_name) == 0)
|
|
|
|
{
|
2018-05-04 15:46:31 +02:00
|
|
|
unlink($temp_name) || die "unlink: $temp_name: $!";
|
2018-05-04 00:06:45 +02:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
rename($temp_name, $final_name) || die "rename: $temp_name: $!";
|
|
|
|
}
|
2018-05-27 15:08:42 +02:00
|
|
|
return;
|
2010-01-05 02:06:57 +01:00
|
|
|
}
|
|
|
|
|
2017-10-04 09:11:36 +02:00
|
|
|
# Find a symbol defined in a particular header file and extract the value.
|
2018-05-19 22:04:47 +02:00
|
|
|
# include_path should be the path to src/include/.
|
2017-10-04 09:11:36 +02:00
|
|
|
sub FindDefinedSymbol
|
|
|
|
{
|
|
|
|
my ($catalog_header, $include_path, $symbol) = @_;
|
2018-05-19 22:04:47 +02:00
|
|
|
my $value;
|
2017-10-04 09:11:36 +02:00
|
|
|
|
2018-05-19 22:04:47 +02:00
|
|
|
# Make sure include path ends in a slash.
|
|
|
|
if (substr($include_path, -1) ne '/')
|
2017-10-04 09:11:36 +02:00
|
|
|
{
|
2018-05-19 22:04:47 +02:00
|
|
|
$include_path .= '/';
|
|
|
|
}
|
|
|
|
my $file = $include_path . $catalog_header;
|
|
|
|
open(my $find_defined_symbol, '<', $file) || die "$file: $!";
|
|
|
|
while (<$find_defined_symbol>)
|
|
|
|
{
|
|
|
|
if (/^#define\s+\Q$symbol\E\s+(\S+)/)
|
2017-10-04 09:11:36 +02:00
|
|
|
{
|
2018-05-19 22:04:47 +02:00
|
|
|
$value = $1;
|
|
|
|
last;
|
2017-10-04 09:11:36 +02:00
|
|
|
}
|
|
|
|
}
|
2018-05-19 22:04:47 +02:00
|
|
|
close $find_defined_symbol;
|
|
|
|
return $value if defined $value;
|
|
|
|
die "$file: no definition found for $symbol\n";
|
2017-10-04 09:11:36 +02:00
|
|
|
}
|
|
|
|
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
# Similar to FindDefinedSymbol, but looks in the bootstrap metadata.
|
|
|
|
sub FindDefinedSymbolFromData
|
2017-03-10 00:17:58 +01:00
|
|
|
{
|
Replace our traditional initial-catalog-data format with a better design.
Historically, the initial catalog data to be installed during bootstrap
has been written in DATA() lines in the catalog header files. This had
lots of disadvantages: the format was badly underdocumented, it was
very difficult to edit the data in any mechanized way, and due to the
lack of any abstraction the data was verbose, hard to read/understand,
and easy to get wrong.
Hence, move this data into separate ".dat" files and represent it in a way
that can easily be read and rewritten by Perl scripts. The new format is
essentially "key => value" for each column; while it's a bit repetitive,
explicit labeling of each value makes the data far more readable and less
error-prone. Provide a way to abbreviate entries by omitting field values
that match a specified default value for their column. This allows removal
of a large amount of repetitive boilerplate and also lowers the barrier to
adding new columns.
Also teach genbki.pl how to translate symbolic OID references into
numeric OIDs for more cases than just "regproc"-like pg_proc references.
It can now do that for regprocedure-like references (thus solving the
problem that regproc is ambiguous for overloaded functions), operators,
types, opfamilies, opclasses, and access methods. Use this to turn
nearly all OID cross-references in the initial data into symbolic form.
This represents a very large step forward in readability and error
resistance of the initial catalog data. It should also reduce the
difficulty of renumbering OID assignments in uncommitted patches.
Also, solve the longstanding problem that frontend code that would like to
use OID macros and other information from the catalog headers often had
difficulty with backend-only code in the headers. To do this, arrange for
all generated macros, plus such other declarations as we deem fit, to be
placed in "derived" header files that are safe for frontend inclusion.
(Once clients migrate to using these pg_*_d.h headers, it will be possible
to get rid of the pg_*_fn.h headers, which only exist to quarantine code
away from clients. That is left for follow-on patches, however.)
The now-automatically-generated macros include the Anum_xxx and Natts_xxx
constants that we used to have to update by hand when adding or removing
catalog columns.
Replace the former manual method of generating OID macros for pg_type
entries with an automatic method, ensuring that all built-in types have
OID macros. (But note that this patch does not change the way that
OID macros for pg_proc entries are built and used. It's not clear that
making that match the other catalogs would be worth extra code churn.)
Add SGML documentation explaining what the new data format is and how to
work with it.
Despite being a very large change in the catalog headers, there is no
catversion bump here, because postgres.bki and related output files
haven't changed at all.
John Naylor, based on ideas from various people; review and minor
additional coding by me; previous review by Alvaro Herrera
Discussion: https://postgr.es/m/CAJVSVGWO48JbbwXkJz_yBFyGYW-M9YWxnPdxJBUosDC9ou_F0Q@mail.gmail.com
2018-04-08 19:16:50 +02:00
|
|
|
my ($data, $symbol) = @_;
|
|
|
|
foreach my $row (@{$data})
|
|
|
|
{
|
|
|
|
if ($row->{oid_symbol} eq $symbol)
|
|
|
|
{
|
|
|
|
return $row->{oid};
|
|
|
|
}
|
|
|
|
}
|
|
|
|
die "no definition found for $symbol\n";
|
2017-03-10 00:17:58 +01:00
|
|
|
}
|
2017-03-10 02:45:52 +01:00
|
|
|
|
2018-04-25 22:01:47 +02:00
|
|
|
# Extract an array of all the OIDs assigned in the specified catalog headers
|
|
|
|
# and their associated data files (if any).
|
2018-04-26 19:22:27 +02:00
|
|
|
# Caution: genbki.pl contains equivalent logic; change it too if you need to
|
|
|
|
# touch this.
|
2018-04-25 22:01:47 +02:00
|
|
|
sub FindAllOidsFromHeaders
|
|
|
|
{
|
|
|
|
my @input_files = @_;
|
|
|
|
|
|
|
|
my @oids = ();
|
|
|
|
|
|
|
|
foreach my $header (@input_files)
|
|
|
|
{
|
|
|
|
$header =~ /(.+)\.h$/
|
|
|
|
or die "Input files need to be header files.\n";
|
|
|
|
my $datfile = "$1.dat";
|
|
|
|
|
|
|
|
my $catalog = Catalog::ParseHeader($header);
|
|
|
|
|
|
|
|
# We ignore the pg_class OID and rowtype OID of bootstrap catalogs,
|
|
|
|
# as those are expected to appear in the initial data for pg_class
|
|
|
|
# and pg_type. For regular catalogs, include these OIDs.
|
|
|
|
if (!$catalog->{bootstrap})
|
|
|
|
{
|
|
|
|
push @oids, $catalog->{relation_oid}
|
|
|
|
if ($catalog->{relation_oid});
|
|
|
|
push @oids, $catalog->{rowtype_oid} if ($catalog->{rowtype_oid});
|
|
|
|
}
|
|
|
|
|
|
|
|
# Not all catalogs have a data file.
|
|
|
|
if (-e $datfile)
|
|
|
|
{
|
|
|
|
my $catdata =
|
|
|
|
Catalog::ParseData($datfile, $catalog->{columns}, 0);
|
|
|
|
|
|
|
|
foreach my $row (@$catdata)
|
|
|
|
{
|
|
|
|
push @oids, $row->{oid} if defined $row->{oid};
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
foreach my $toast (@{ $catalog->{toasting} })
|
|
|
|
{
|
|
|
|
push @oids, $toast->{toast_oid}, $toast->{toast_index_oid};
|
|
|
|
}
|
|
|
|
foreach my $index (@{ $catalog->{indexing} })
|
|
|
|
{
|
|
|
|
push @oids, $index->{index_oid};
|
|
|
|
}
|
Rethink method for assigning OIDs to the template0 and postgres DBs.
Commit aa0105141 assigned fixed OIDs to template0 and postgres
in a very ad-hoc way. Notably, instead of teaching Catalog.pm
about these OIDs, the unused_oids script was just hacked to
not show them as unused. That's problematic since, for example,
duplicate_oids wouldn't report any future conflict. Hence,
invent a macro DECLARE_OID_DEFINING_MACRO() that can be used to
define an OID that is known to Catalog.pm and will participate
in duplicate-detection as well as renumbering by renumber_oids.pl.
(We don't anticipate renumbering these particular OIDs, but we
might as well build out all the Catalog.pm infrastructure while
we're here.)
Another issue is that aa0105141 neglected to touch IsPinnedObject,
with the result that it now claimed template0 and postgres are
pinned. The right thing to do there seems to be to teach it that
no database is pinned, since in fact DROP DATABASE doesn't check
for pinned-ness (and at least for these cases, that is an
intentional choice). It's not clear whether this wrong answer
had any visible effect, but perhaps it could have resulted in
erroneous management of dependency entries.
In passing, rename the TemplateDbOid macro to Template1DbOid
to reduce confusion (likely we should have done that way back
when we invented template0, but we didn't), and rename the
OID macros for template0 and postgres to have a similar style.
There are no changes to postgres.bki here, so no need for a
catversion bump.
Discussion: https://postgr.es/m/2935358.1650479692@sss.pgh.pa.us
2022-04-21 22:23:12 +02:00
|
|
|
foreach my $other (@{ $catalog->{other_oids} })
|
|
|
|
{
|
|
|
|
push @oids, $other->{other_oid};
|
|
|
|
}
|
2018-04-25 22:01:47 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
return \@oids;
|
|
|
|
}
|
|
|
|
|
2010-01-05 02:06:57 +01:00
|
|
|
1;
|