(why bother dropping individual objects in a just-created database?)
as well as dangerous (as the code stands, the drops will be issued in
the wrong database, namely the one you were originally connected to).
never overwrite adjacent pages with copied data, even if page header
and/or item pointers are already corrupt. Change inspired by trouble
report from Alvaro Herrera.
environment strings need to be moved around, do so when called from
initial startup (main.c), not in init_ps_status. This eliminates the
former risk of invalidating saved environment-string pointers, since
no code has yet had a chance to grab any such pointers when main.c
is running.
in .:/home/postgres/testversion/bin:/opt/perl5.6.1/bin:/home/postgres/bin:/usr/local/bin:/usr/bin:/opt/ansic/bin:/usr/ccs/bin:/usr/contrib/bin:/opt/nettladm/bin:/opt/pd/bin:/usr/bin/X11:/usr/contrib/bin/X11:/opt/upgrade/bin:/opt/CC/bin:/opt/langtools/bin:/opt/graphics/phigs/bin:/opt/java/bin:/bin:/opt/imake/bin:/opt/hparray/bin:/opt/aCC/bin:/opt/lrom/bin:/usr/local/nmh/bin:. (I suppose the only common case for this is '.').
per suggestion from Peter. Simplify several APIs by transmitting the
original argv location directly from main.c to ps_status.c, instead of
passing it down through several levels of subroutines.
Mask both typmod subfields for INTERVAL to avoid setting the high bit,
per dire warning from Tom Lane.
Clear tmask for DTK_ISO_TIME case to avoid time zone troubles.
Symptom reported by Tom Lane.
Clean up checking for valid time zone info in output routine.
This should now work for both SQL99 and Unix-style time zones.
Put in explicit check for INTERVAL() typmod rounding to avoid accumulating
cruft in the lower bits. Not sure that this helps, but we'll need to do
something. The symptom is visible with a query like
select interval(2) '10000 days 01:02:03.040506';
Regression tests are patched to repair the Tom Lane symptom, and all pass.
subprocesses; perhaps this will fix portability problem just noted by
Lockhart. Also, move test for bad permissions of DataDir to a more
logical place.
bootstrap) check for a valid PG_VERSION file before looking at anything
else in the data directory. This fixes confusing error report when
trying to start current sources in a pre-7.1 data directory.
Per trouble report from Rich Shepard 10/18/01.
just after receipt of the startup packet. Now, postmaster children
that are waiting for client authentication response will show as
'postgres: user database host authentication'. Also, do an
init_ps_display for startup/shutdown/checkpoint subprocesses,
so that they are readily identifiable as well. Fix an obscure race
condition that could lead to Assert failure in the postmaster ---
attempting to start a checkpoint process before any connections have
been received led to calling PostmasterRandom before setting random_seed.
authentication failed' and a 'send() failed: Broken pipe' message
on every connection from psql in password auth mode. Problem is
that psql doesn't ask user for a password until it sees a password
challenge failure, and libpq just closes the connection unceremoniously
if it's challenged for a password when it hasn't got one to send.
Accordingly, EOF from the client after asking for a password is
normal behavior and should not result in postmaster log entries.
the entered password would get echoed on some platforms, eg HPUX.
We have enough copies of this code that I'm thinking it ought to be
moved into libpq, but that's a task for another day.
Modified the parser and the SET handlers to use full Node structures
rather than simply a character string argument.
Implement INTERVAL() YEAR TO MONTH (etc) syntax per SQL99.
Does not yet accept the goofy string format that goes along with, but
this should be fairly straight forward to fix now as a bug or later
as a feature.
Implement precision for the INTERVAL() type.
Use the typmod mechanism for both of INTERVAL features.
Fix the INTERVAL syntax in the parser:
opt_interval was in the wrong place.
INTERVAL is now a reserved word, otherwise we get reduce/reduce errors.
Implement an explicit date_part() function for TIMETZ.
Should fix coersion problem with INTERVAL reported by Peter E.
Fix up some error messages for date/time types.
Use all caps for type names within message.
Fix recently introduced side-effect bug disabling 'epoch' as a recognized
field for date_part() etc. Reported by Peter E. (??)
Bump catalog version number.
Rename "microseconds" current transaction time field
from ...Msec to ...Usec. Duh!
date/time regression tests updated for reference platform, but a few
changes will be necessary for others.
clause being added to a particular restriction-clause list is redundant
with those already in the list. This avoids useless work at runtime,
and (perhaps more importantly) keeps the selectivity estimation routines
from generating too-small estimates of numbers of output rows.
Also some minor improvements in OPTIMIZER_DEBUG displays.
view when using the aggregate function count() and function nextval
that returns an int8 value, but in python is represented like string:
>> db.query("select nextval('my_seq')").getresult()
[('2',)]
>> db.query("select count(*) from films").dictresult()
[{'count': '120'}]
Ricardo Caesar Lenzi
> ! $$ = cat_str(8, make_str("grant"), $2, make_str("on"), $4, $5,
> make_str("to"), $7, $8);
> ISTM your patch loses the opt_with_grant clause. (Of course the
> backend doesn't currently accept that clause anyway, but that's no
> reason for ecpg to drop it.)
My patch doesn't loose the option, it's never been passed on anyway:
opt_with_grant: WITH GRANT OPTION
{
mmerror(ET_ERROR, "WITH GRANT OPTION is not supported. Only relation owners can
set privileges");
}
| /*EMPTY*/
;
The existing code in ecpg/preproc/preproc.y to handle the WITH option
simply throws an error and aborts the processing... The patch below
prevents the segfault and also passes on the WITH option to the
backend, probably a better fix.
Lee Kindness
> - corrects a bit the UTF-8 code from Tatsuo to allow Unicode 3.1
> characters (characters with values >= 0x10000, which are encoded on
> four bytes).
Also, update mb/expected/unicode.out. This is necessary since the
patches affetc the result of queries using UTF-8.
---------------------------------------------------------------
Hi,
I should have sent the patch earlier, but got delayed by other stuff.
Anyway, here is the patch:
- most of the functionality is only activated when MULTIBYTE is
defined,
- check valid UTF-8 characters, client-side only yet, and only on
output, you still can send invalid UTF-8 to the server (so, it's
only partly compliant to Unicode 3.1, but that's better than
nothing).
- formats with the correct number of columns (that's why I made it in
the first place after all), but only for UNICODE. However, the code
allows to plug-in routines for other encodings, as Tatsuo did for
the other multibyte functions.
- corrects a bit the UTF-8 code from Tatsuo to allow Unicode 3.1
characters (characters with values >= 0x10000, which are encoded on
four bytes).
- doesn't depend on the locale capabilities of the glibc (useful for
remote telnet).
I would like somebody to check it closely, as it is my first patch to
pgsql. Also, I created dummy .orig files, so that the two files I
created are included, I hope that's the right way.
Now, a lot of functionality is NOT included here, but I will keep that
for 7.3 :) That includes all string checking on the server side (which
will have to be a bit more optimised ;) ), and the input checking on
the client side for UTF-8, though that should not be difficult. It's
just to send the strings through mbvalidate() before sending them to
the server. Strong checking on UTF-8 strings is mandatory to be
compliant with Unicode 3.1+ .
Do I have time to look for a patch to include iso-8859-15 for 7.2 ?
The euro is coming 1. january 2002 (before 7.3 !) and over 280
millions people in Europe will need the euro sign and only iso-8859-15
and iso-8859-16 have it (and unfortunately, I don't think all Unices
will switch to Unicode in the meantime)....
err... yes, I know that this is not every single person in Europe that
uses PostgreSql, so it's not exactly 280m, but it's just a matter of
time ! ;)
I'll come back (on pgsql-hackers) later to ask a few questions
regarding the full unicode support (normalisation, collation,
regexes,...) on the server side :)
Here is the patch !
Patrice.
--
Patrice HÉDÉ ------------------------------- patrice à islande org -----
-- Isn't it weird how scientists can imagine all the matter of the
universe exploding out of a dot smaller than the head of a pin, but they
can't come up with a more evocative name for it than "The Big Bang" ?
-- What would _you_ call the creation of the universe ?
-- "The HORRENDOUS SPACE KABLOOIE !" - Calvin and Hobbes
------------------------------------------ http://www.islande.org/ -----
Enabling this feature adds very light overhead of 1 select from pg_class on
first using of pl/tcl in backend if unknown suppport is really unused.
But pl/tcl with this support has very improved functionality.
Patch includes changes to documentation.
side encoding name. This is necessary for client API's such as JDBC
to perform correct encoding conversions. See my email "[HACKERS]
pg_client_encoding" 10 Sep 2001.
transformAlterStmt() use these routines, instead of having lots of
duplicate (not to mention should-have-been-duplicate) code.
Adding a column with a CHECK constraint actually works now,
and the tests to reject unsupported DEFAULT and NOT NULL clauses
actually fire now. ALTER TABLE ADD PRIMARY KEY works, modulo
having to have created the column(s) NOT NULL already.
> As you can see, psql reconnect as any user if the password is same as
> foo. Of course this is due to the careless password setting, but I
> think it's better to prompt ANY TIME the user tries to switch to
> another user. Comments?
Yeah, I agree. Looks like a simple change in dbconnect():
/*
* Use old password if no new one given (if you didn't have an old
* one, fine)
*/
if (!pwparam && oldconn)
pwparam = PQpass(oldconn);
to
/*
* Use old password (if any) if no new one given and we are
* reconnecting as same user
*/
if (!pwparam && oldconn && PQuser(oldconn) && userparam &&
strcmp(PQuser(oldconn), userparam) == 0)
pwparam = PQpass(oldconn);
regards, tom lane
current_timestamp, current_date for ODBC compatibility.
Add more functions to odbc.sql catalog extension, use new CREATE OR
REPLACE FUNCTION.
Document iODBC/unixODBC build options.
That patch broke the ability to read data from binary cursors.
--Barry Lind
Modified Files:
pgsql/src/interfaces/jdbc/org/postgresql/Connection.java
pgsql/src/interfaces/jdbc/org/postgresql/ResultSet.java
pgsql/src/interfaces/jdbc/org/postgresql/core/QueryExecutor.java
pgsql/src/interfaces/jdbc/org/postgresql/jdbc1/Connection.java
pgsql/src/interfaces/jdbc/org/postgresql/jdbc1/ResultSet.java
pgsql/src/interfaces/jdbc/org/postgresql/jdbc2/Connection.java
pgsql/src/interfaces/jdbc/org/postgresql/jdbc2/ResultSet.java
pgsql/src/interfaces/jdbc/org/postgresql/jdbc2/UpdateableResultSet.java
NOTICE added about error location (same method already used by plpgsql
executor). Add checking of pg_proc row xmin/cmin to ensure that
plpgsql functions will be recompiled after they've been modified by
CREATE OR REPLACE FUNCTION.
rather than having its own somewhat half-baked notion of what a type
declaration looks like. This is necessary now to ensure that plpgsql
will think a 'timestamp' variable has the same semantics as 'timestamp'
does in the main SQL grammar; and it should avoid divergences in future.
for nested typecasts. It now produces a column header of 'timestamptz'
for 'SELECT CURRENT_TIMESTAMP', rather than 'text' as it was doing for
awhile there.
lookup info in the relcache for index access method support functions.
This makes a huge difference for dynamically loaded support functions,
and should save a few cycles even for built-in ones. Also tweak dfmgr.c
so that load_external_function is called only once, not twice, when
doing fmgr_info for a dynamically loaded function. All per performance
gripe from Teodor Sigaev, 5-Oct-01.
on words as opposed to lines, which means that all of the
following work in psql:
\d foo \d bar
\d foo; \d bar
\d foo \d bar;;
\d foo; <space>
This one also uses "true and false" and strips semicolons
for the following backslash commands: \C \c \d \e \i \o \s \z
Greg Sabino Mullane
readability. Bizarre '(long *) TRUE' return convention is gone,
in favor of just raising an error internally in dynahash.c when
we detect hashtable corruption. HashTableWalk is gone, in favor
of using hash_seq_search directly, since it had no hope of working
with non-LONGALIGNable datatypes. Simplify some other code that was
made undesirably grotty by promixity to HashTableWalk.
type coercion after failing to find an exact match in pg_proc, but before
considering interpretations that involve a function call with one or
more argument type coercions. This avoids surprises wherein what looks
like a type coercion is interpreted as coercing to some third type and
then to the destination type, as in Dave Blasby's bug report of 3-Oct-01.
See subsequent discussion in pghackers.
which contains stack trace.
One problem, after this patch errors will generate multiline ERROR
messages. Is it acceptable or do I need split it and generate multiple
singleline messages?
Vsevolod Lobko
select 'id' as xxx from table
The issue is:
When the driver gets a data type which does not map into the SQL.Types
it attempts to load the object into a java object. Eventually throwing
an exception indicating that the type "unknown" was not found.
Since the backend defaults "unknown" types to text it was suggested that
the jdbc driver do the same.
This patch does just that.
I have tested it on the above select statement as well as a small
program that serializes, and deserializes a class
Dave Cramer
NLS effort for the 'postgresql' component.
It includes 'ru.po', which should be added
to /src/backend/po/ and a typical patch for
nls.mk with the 'ru' entry added in
the AVAIL_LANGUAGES.
I'll submit a patch with more translations
for pg_dump once CVSweb becomes available.
CVSweb for me right now is the only way
to get to the source code until my system
is completely setup by the end of the week.
--
Serguei A. Mokhov
'aggname (aggtype)'. The old syntax 'aggname aggtype' is still accepted
for backwards compatibility. Fix pg_dump, which was actually broken for
most cases of user-defined aggregates. Clean up error messages associated
with these commands.
-infinity and +infinity. Put TIMESTAMP_NOT_FINITE guard into the routine,
instead of forgetting it at some call sites. Fixes regression test
failures here.
time zones.
SQL99 spec requires a default of zero (round to seconds) which is set
in gram.y as typmod is set in the parse tree. We *could* change to a
default of either 6 (for internal compatibility with previous versions)
or 2 (for external compatibility with previous versions).
Evaluate entries in pg_proc wrt the iscachable attribute for timestamp and
other date/time types. Try to recognize cases where side effects like the
current time zone setting may have an effect on results to decide whether
something is cachable or not.
pg_get_indexdef() function, rather than reaching into the system catalogs
for itself. This eliminates a fair amount of redundant code. Also,
since I just changed pg_get_indexdef() to suppress display of default
index opclasses, this will mean that 7.2 and later dumps will not mention
opclasses unless they are non-default opclasses. Should make life easier
for future index opclass reorganizations.
it to suppress index opclass output for opclasses that are the default
for their datatype; only non-default opclasses are shown explicitly.
This is expected to improve portability of the CREATE INDEX command
across future versions of Postgres --- we've changed index opclasses
too often in the past to think we won't do so again.
Partial support for BEOS (not sure whether second fork of grandchild
process needs these extra calls or not; someone who has BEOS will need
to test it).
> > > > > and --enable-unicode-convertion if it ought to work correctly
> > > > > with Tcl/Tk >= 8.1 (client or server side).
> > > > >
> > > > > - PL/Tcl needs to be changed to use pg_do_encoding_conversion
> > > > > if it runs on a Tcl version >= 8.1 .
> > >
> > > > I'll do pl/tcl part in the next version of patch. Using this approach we
> > > > can eliminate overhead for databases in UNICODE.
> > >
> > > Any progress on this? I'd prefer to get rid of this --enable-pltcl-utf
> > > option before release.
> >
> > Done
> >
> > Next version removes --enable-pltcl-utf switch and enables embedded
> > utf conversion of pgsql if tcl version >=8.1 and --enable-unicode-conversion
portability issues). Caller-visible data structures are now allocated
on MAXALIGN boundaries, allowing safe use of datatypes wider than 'long'.
Rejigger hash_create API so that caller specifies size of key and
total size of entry, not size of key and size of rest of entry.
This simplifies life considerably since each number is just a sizeof(),
and padding issues etc. are taken care of automatically.
from the config file, so that these changes will propagate to backends
started later. Already-started backends continue to ignore changes
in these variables.
upper limit on what we will believe from sysconf(_SC_OPEN_MAX). The
default value is 1000, so that under ordinary conditions it won't
affect the behavior. But on platforms where the kernel promises far
more than it can deliver, this can be used to prevent running out of
file descriptors. See numerous past discussions, eg, pgsql-hackers
around 23-Dec-2000.
rightmost on its tree level, we split 2/3 to the left and 1/3 to the
new right page, rather than the even split we use elsewhere. The idea
is that when faced with a steadily increasing series of inserted keys
(such as sequence or timestamp values), we'll end up with a btree that's
about 2/3ds full not 1/2 full, which is much closer to the desired
steady-state load for a btree. Per suggestion from Ann Harrison of
IBPhoenix.
pointers to data that will be changed by any later call to setlocale.
Must copy what they return to be sure we get the right answer.
Karel Zak, further tweaks by Tom Lane.
existing lock manager and spinlocks: it understands exclusive vs shared
lock but has few other fancy features. Replace most uses of spinlocks
with lightweight locks. All remaining uses of spinlocks have very short
lock hold times (a few dozen instructions), so tweak spinlock backoff
code to work efficiently given this assumption. All per my proposal on
pghackers 26-Sep-01.
two performance improvements. I put an explanation of the
changes at
http://cs1.cs.nyu.edu/been/postgres-rtree.html
The performance improvements are quite significant.
All the changes are in the file src/backend/access/rtree/rtree.c
I was working with the 7.1.3 code.
I'm including the diff output as an attachment.
Kenneth Been
DatabaseMetaData.getColumn(). I proposed a patch that would change the
number of queries to find out all columns in a table from 2 * N + 1 to 1 (N
being the number of columns reported) by using some outer joins. I also
fixed the fact that getColumns() only returned columns that had a default
defined. OTOH, I did not use to change the code required for obtaining a
column's remarks (by using col_description() for 7.2 and requested by Tom
Lane).
Finally, I have found a way to get all the column details in a single query
*and* use col_description() for 7.2 servers. A patch is attached. It
overrules Ren? Pijlman's fix for this that was committed just today, but
still used N + 1 queries (sorry Ren? ;-) )
I also fixed the return values for TABLE_CAT and TABLE_SCHEM from "" to
null, to be more standard compliant (and requested in Ren?'s mail found at
http://fts.postgresql.org/db/mw/msg.html?mid=1034253).
As always, the JDBC1 version has not been tested as I have no JDK 1.1
Jeroen van Vianen
Define a new function, GetCurrentTransactionStartTimeUsec() to get the time
to this precision.
Allow now() and timestamp 'now' to use this higher precision result so
we now have fractional seconds in this "constant".
Add timestamp without time zone type.
Move previous timestamp type to timestamp with time zone.
Accept another ISO variant for date/time values: yyyy-mm-ddThh:mm:ss
(note the "T" separating the day from hours information).
Remove 'current' from date/time types; convert to 'now' in input.
Separate time and timetz regression tests.
Separate timestamp and timestamptz regression test.
check *can* return HEAPTUPLE_INSERT_IN_PROGRESS or HEAPTUPLE_DELETE_IN_PROGRESS,
even though we have ShareLock on the relation. To wit, this can happen
if the tuple was inserted/deleted by our own transaction. Per report from
Justin Clift 9/23.
summary of changes:
. removal of the tablename property from build.xml
. addition of a dropTable method in JDBC2Tests and cleanups of many
methods in the same
. all tests now use non-deprecated assertXYZ methods instead of the
deprecated assert method
. failure in TimestampTest (testSetTimestamp) fixed. The failure is
because testSetTimestamp was inserting a timestamp with hour 7 but
checkTimeTest was expecting a timestamp with hour 8. AFAICS, there are
no issues wrt daylight savings time and timestamps being pushed in and
pulled out (but more explicit tests should be added in the future)
. failure in TimeTest (testGetTime) fixed. Times to be inserted were
interpreted in the localtime zone but checking was done with the
assumption that the insertion was done in GMT.
. formatting changes in a few of the source files (because I found
it convenient to have consistent formatting while working on them). The
formatting is consistent with the new format for java source files in
PostgreSQL.
Liam Stewart
an already installed iODBC or unixODBC driver manager. In particular,
use the include files provided by the driver manager over our own,
and use the odbcinst library of the driver manager rather than gpps.c.
Migrate portability sections common to several files into psqlodbc.h.
a hung client or lost connection can't indefinitely block a postmaster
child (not to mention the possibility of deliberate DoS attacks).
Timeout is controlled by new authentication_timeout GUC variable,
which I set to 60 seconds by default ... does that seem reasonable?
Frank Miles 7-Sep-01. This is really just sticking a finger in the dike.
Frank's case works now, but we still couldn't support a recursive function
returning a set. Really need to restructure querytrees and execution
state so that the querytree is *read only*. We've run into this over and
over and over again ... it has to happen sometime soon.
the JDBC driver.
This method is currently unimplemented and always returns
ResultSetMetaData.columnNullable. This is obviously incorrect
when a column is defined with NOT NULL or PRIMARY KEY. And we
have to think of check constraints, views, functions etc.
The patch simply changes the return value to
ResultSetMetaData.columnNullableUnknown. This is until someone
comes up with a real implementation of course.
On Fri, 14 Sep 2001 17:53:50 +0200, Tomisaw Kity?ski wrote:
>Hello there,
>
>could someone tell me, please, do I have any chance to get
>proper implementation of above method in JDBC (1.1+) soon?
>
>Current "return 1" works fine on most tables, however it seems
>to be a little bit incorrect with some of them ;)
Ren? Pijlman
by escape processing in the SQL statement. I've tested this for a
while now and it appears to work well. Previously string data
with {d was getting corrupt as the {d was being stripped regardless
of whether it was an escape code or not.
I also added checking for time and timestamp escape processing strings
as per 11.3 in the specification. The patch is against the latest
CVS.
Thomas O'Dowd
under libdir, for a cleaner separation in the installation layout
and compatibility with binary packaging standards. Point backend's
default search location there. The contrib modules are also
installed in the said location, giving them the benefit of the
default search path as well. No changes in user interface
nevertheless.
>
> 1. Now outputs '\\' instead of '\134' when using encode(bytea, 'escape')
> Note that I ended up leaving \0 as \000 so that there are no ambiguities
> when decoding something like, for example, \0123.
>
> 2. Fixed bug in byteain which allowed input values which were not valid
> octals (e.g. \789), to be parsed as if they were octals.
>
> Joe
>
Here's rev 2 of the bytea string support patch. Changes:
1. Added missing declaration for MatchBytea function
2. Added PQescapeBytea to fe-exec.c
3. Applies cleanly on cvs tip from this afternoon
I'm hoping that someone can review/approve/apply this before beta starts, so
I guess I'd vote (not that it counts for much) to delay beta a few days :-)
Joe Conway
> null bytes to be literally '\0', the following can happen:
> 1. User inputs string value as "<null byte>##" where ## are digits in the
> range of 0 to 7.
> 2. PQescapeString converts this to "\0##"
> 3. Escaped string is used in a context that causes "\0##" to be evaluated as
> an octal escape sequence.
I agree that this is a problem, though it is not possible to do
anything harmful with it. In addition, it only occurs if there are
any NUL characters in its input, which is very unlikely if you are
using C strings.
The patch below addresses the issue by removing escaping of \0
characters entirely.
> If the goal is to "safely" encode null bytes, and preserve the rest of the
> string as it was entered, I think the null bytes should be escaped as \\000
> (note that if you simply use \000 the same string truncation problem
> occurs).
We can't do that, this would require 4n + 1 bytes of storage for the
result, breaking the interface.
Florian Weimer
driver's test suite. With previous patches applied, this reduces
the number of failures of the test suite from 6 to 4. The patch
fixes the test case itself, rather than the driver.
Details:
1) The driver correctly provided DatabaseMetaData about the sort
order of NULLs. This was confirmed by Peter Eisentraut on
pgsql-hackers. I fixed the test to accept/require the current
behaviour, and made it dependent on the backend version. See
nullsAreSortedAtStart(), nullsAreSortedAtEnd(),
nullsAreSortedHigh() and nullsAreSortedLow().
2) DatabaseMetaData.supportsOrderByUnrelated() correctly
returned true (an ORDER BY clause can contain columns that are
not in the SELECT clause), but the test case required false.
Fixed that.
3) Replaced deprecated assert() of junit.framework.TestCase by
assertEquals(), assertTrue() and assertNotNull(). This is
because assert will be a new keyword in Java 1.4.
4) Replaced assert(message,false) by the more elegant
fail(message).
Regards,
Ren? Pijlman <rene@lab.applinet.nl>
This patch does the following:
- Adds binary datatype support (bytea)
- Changes getXXXStream()/setXXXStream() methods to be spec compliant
- Adds ability to revert to old behavior
Details:
Adds support for the binary type bytea. The ResultSet.getBytes() and
PreparedStatement.setBytes() methods now work against columns of bytea
type. This is a change in behavior from the previous code which assumed
the column type was OID and thus a LargeObject. The new behavior is
more complient with the JDBC spec as BLOB/CLOB are to be used for
LargeObjects and the getBytes()/setBytes() methods are for the databases
binary datatype (which is bytea in postgres).
Changes the behavior of the getBinaryStream(), getAsciiStream(),
getCharacterStream(), getUnicodeStream() and their setXXXStream()
counterparts. These methos now work against either the bytea type
(BinaryStream) or the text types (AsciiStream, CharacterStream,
UnicodeStream). The previous behavior was that these all assumed the
underlying column was of type OID and thus a LargeObject. The
spec/javadoc for these methods indicate that they are for LONGVARCHAR
and LONGVARBINARY datatypes, which are distinct from the BLOB/CLOB
datatypes. Given that the bytea and text types support upto 1G, they
are the LONGVARBINARY and LONGVARCHAR datatypes in postgres.
Added support for turning off the above new functionality. Given that
the changes above are not backwardly compatible (however they are more
spec complient), I added the ability to revert back to the old behavior.
The Connection now takes an optional parameter named 'compatible'. If
the value of '7.1' is passed, the driver reverts to the 7.1 behavior.
If the parameter is not passed or the value '7.2' is passed the behavior
is the new behavior. The mechanism put in place can be used in the
future when/if similar needs arise to change behavior. This is
patterned after how Oracle does this (i.e. Oracle has a 'compatible'
parameter that behaves in a similar manner).
Misc fixes. Cleaned up a few things I encountered along the way.
Note that in testing the patch I needed to ignore whitespace differences
in order to get it to apply cleanly (i.e. patch -l -i byteapatch.diff).
Also this patch introduces a new file
(src/interfaces/jdbc/org/postgresql/util/PGbytea.java).
Barry Lind
>there is still an unpatched reference to pg_description in
>getColumns(), in both jdbc1 and jdbc2.
This was introduced by Jeroen's patch (see
http://fts.postgresql.org/db/mw/msg.html?mid=1032468). Attached
is a patch that returns getColumns() to using "select
obj_description()" instead of direct access to pg_description,
as per the request by Tom.
I've incorporated Jeroen's fix to left outer join with
pg_attrdef instead of inner join, so getColumns() also returns
columns without a default value.
I have, however, not included Jeroen's attempt to combine
multiple queries into one huge multi-join query for better
performance, because:
1) I don't know how to do that using obj_description() instead
of direct access to pg_description
2) I don't think a performance improvement (if any) in this
method is very important
Because of the outer join, getColumns() will only work with a
backend >= 7.1. Since the conditional coding for 7.1/7.2 and
jdbc1/jdbc2 is already giving me headaches I didn't pursue a
pre-7.1 solution.
Regards,
Ren? Pijlman <rene@lab.applinet.nl>
ConnectionTest.testTransactionIsolation() in the JDBC driver's
test suite. This reduces the number of failures of the test
suite from 7 to 6. The patch fixes the test case itself, rather
than the driver.
In addition to the change described in my posting below, I fixed
the part of the test with autocommit enabled. The author of the
test assumed that setting the transaction isolation level would
have no effect, but in fact it does. Perhaps the test case
worked with pre-7.1 behaviour, when the JDBC driver set the
isolation level in every transaction, instead of using "set
session characteristics". Anyway, now it works with a backend
built from current CVS and the behaviour is JDBC compliant.
I also extended the test case by changing the isolation level
before beginning a transaction and verifying it inside the
transaction.
Regards,
Ren? Pijlman
Given the following table:
test=# \d f
Table "f"
Column | Type | Modifiers
--------+---------+-----------
i | integer |
test | text |
If I do the following:
test=# insert into f values(1,'test');
INSERT 139549 1
test=# select i::int8,test from f;
?column? | test
----------+------
1 | test
(1 row)
It doesn't make much sense that the first column should be called
'?column?'.
The patch results in the output appearing like this:
test=# select i::int8,test from f;
i | test
---+------
1 | test
(1 row)
----------
Gavin Sherry
>
> "parse error at [the] end of line"
>
> Attached patch also fixes it. I noticed this while editing the po file.
> If I'm wrong, please ignore the command.c.patch. I will revert my translation
> as well then.
>
> --
> Serguei A. Mokhov
PostgreSQL to support unicode-conversion, but retains binary
compatibility among Tcl versions.
However, it neither checks at compile time not at runtime, if support
for unicode-conversion does really exist and it doesn't prevent the
user from changing the client encoding after initialization. I think
there should be warnings about this somewhere in the documentation.
Reinhard Max
Assign the fixed user id 1 to the user created by initdb.
A stand-alone backend will always set the user id to 1.
(Consequently, the name of that user is no longer important.)
In stand-alone mode, the user id 1 will have implicit superuser
status, to allow repairs even if there are no users defined.
Print a warning message when starting in stand-alone mode when no
users are defined.
Disallow dropping the current user and session user.
Granting/revoking superuser status also grants/revokes usecatupd.
(Previously, it would never grant it back. This could lead to "deadlocks".)
CREATE USER and CREATE GROUP will start allocating user ids at 100
(unless explicitly specified), to prevent accidental creation of a
superuser (plus some room for future extensions).
> > ClientEncoding to SQL_ASCII (like default DatabaseEncoding). Bruce, can
> > you change it? It's one line change. Again thanks.
Forget it! A default client encoding must be set by actual database encoding...
Please apply the small attached patch that solve it better.
Karel Zak
We will no longer try to send elog messages to the client before we have
initialized backend libpq (oops); however, reporting bogus commandline
switches via elog does work now (not irrelevant, because of PGOPTIONS).
Fix problem with inappropriate sending of checkpoint-process messages
to stderr.
suite. This reduces the number of failures from 9 to 7.
Both ConnectionTest and JBuilderTest did not create their own
tables, which caused these test cases to fail with "relation ...
does not exist". It appears these test cases relied on tables
created by the example code elsewhere in the source tree. I've
added the necessary "create table" and "drop table" statements
to the test cases, using the column definitions from the example
code.
While working on that I modified the helper method createTable
in JDBC2Tests.java to take a table parameter, rather than using
table names passed via the properties in build.xml. I'm not sure
what that was good for, and in fact, except for the default
table name "jdbctest", this functionality wasn't used at all.
Ren? Pijlman
discussion on pgsql-hackers (especially the frightening memory dump in
<12273.999562219@sss.pgh.pa.us>), we decided that it is best not to
use identifiers from an untrusted source at all. Therefore, all
claims of the suitability of PQescapeString() for identifiers have
been removed.
Florian Weimer
table creation time. Big deal you say - but this patch is the basis of the
next thing which is adding PRIMARY KEYs after table creation time. (Which
is currently impossible without twiddling catalogs)
Rundown
-------
* I have made the makeObjectName function of analyze.c non-static, and
exported it in analyze.h
* I have included analyze.h and defrem.h into command.c, to support
makingObjectNames and creating indices
* I removed the 'case CONSTR_PRIMARY' clause so that it properly fails and
says you can't add primary keys, rather than just doing nothing and
reporting nothing!!!
* I have modified the docs.
Algorithm
---------
* If name specified is null, search for a new valid constraint name. I'm
not sure if I should "lock" my generated name somehow tho - should I open
the relation before doing this step?
* Open relation in access exclusive mode
* Check that the constraint does not already exist
* Define the new index
* Warn if they're doubling up on an existing index
Christopher Kings-Lynne
>tcl-extension for postgreSQL.
>I'm currently using 7.0 and always getting a seg fault when I try to
>read from the database connection after issueing a "COPY table TO
>stdout;" (I'm using the connection handle, *not* the result handle).
>Maybe this is fixed in a later release.
>The README file in src/interfaces/libpgtcl tells me, that this should
>work, but unforunately it doesn't.
Yes, it seems broken. It is a bug in libpgtcl. Are you running Tcl >= 8.3.2?
That's when the Tcl team changed the data structure for channel
callbacks. The change itself was designed to be backward compatible, but I
suspect a related change made the code more sensitive to errors in the
structure (NULL pointers where functions are required). Either that, or
nobody has tried to use libpgtcl with COPY in a long time.
First, I have to say I can't think of a good reason to use PostgreSQL's
COPY command from a Tcl application. I think it should only be used with
psql for importing data from another source into PostgreSQL, or for
exporting PostgreSQL data into another database (but why would anyone do
that?) If it was me, I would stick with SELECT and INSERT and be "SQL
Compliant".
OK, editorial is over. Try applying the patch below to fix
src/interfaces/libpgtcl/pgtclId.c
and let us know if it works. I did little testing on it, but my test did
segfault before and ran fine (copy in and copy out) after the patch. This
is for PostgreSQL-7.1.2 - since you are running older 7.0, I don't know if
this will work, but I suspect it will.
PS It's the absence of PgWatchProc which kills it. I didn't upgrade it
to the "V2" channel type structure, so it should be compatible with older
Tcl's. But aside from gets and puts, I doubt any other file operations
would work on the handle during a copy.
ljb
occur unconditionally, even if the rule should otherwise execute
conditionally. This is more useful than giving an error, even though it's
not truly the correct behavior. Per today's pghackers discussion.
for them, and making them just wastes time during backend startup/shutdown.
Also, remove compile-time MAXBACKENDS limit per long-ago proposal.
You can now set MaxBackends as high as your kernel can stand without
any reconfiguration/recompilation.
>
>> On Mon, 3 Sep 2001 22:01:17 -0500, you wrote:
>> public boolean isWritable(int column) throws SQLException
>> {
>> return !isReadOnly(column);
>> }
Actually, I think this change has a consequence for this method
in the same class:
public boolean isDefinitelyWritable(int column)
throws SQLException
{
return isWritable(column);
}
This is from the JDBC spec
(http://java.sun.com/j2se/1.3/docs/api/java/sql/ResultSetMetaData.html):
isReadOnly() - Indicates whether the designated column is
definitely not writable.
isWritable() - Indicates whether it is possible for a write on
the designated column to succeed.
isDefinitelyWritable() - Indicates whether a write on the
designated column will definitely succeed.
At this time we don't really implement the fine semantics of
these methods. I would suggest the following defaults:
isReadOnly() false
isWritable() true
isDefinitelyWritable() false
And that would mean that your patch is correct, but
isDefinitelyWritable() would need to be patched accordingly:
public boolean isDefinitelyWritable(int column)
throws SQLException
{
return false;
}
Again, both in jdbc1 and jdbc2.
Regards,
Ren? Pijlman <rene@lab.applinet.nl>
>public boolean isWritable(int column) throws SQLException
>{
> if (isReadOnly(column))
> return true;
> else
> return false;
>}
The author probably intended:
public boolean isWritable(int column) throws SQLException
{
return !isReadOnly(column);
}
And if he would have coded it this way he wouldn't have made
this mistake :-)
>hence, isWritable() will always return false. this is something
>of a problem :)
Why exactly? In a way, true is just as incorrect as false, and
perhaps it should throw "not implemented". But I guess that
would be too non-backwardly-compatible.
>let me know if i can provide further information.
Will you submit a patch?
Regards,
Ren? Pijlman <rene@lab.applinet.nl>
-------------------------------------------------------------------
Subject: Re: [PATCHES] encoding names
From: Karel Zak <zakkr@zf.jcu.cz>
To: Peter Eisentraut <peter_e@gmx.net>
Cc: pgsql-patches <pgsql-patches@postgresql.org>
Date: Fri, 31 Aug 2001 17:24:38 +0200
On Thu, Aug 30, 2001 at 01:30:40AM +0200, Peter Eisentraut wrote:
> > - convert encoding 'name' to 'id'
>
> I thought we decided not to add functions returning "new" names until we
> know exactly what the new names should be, and pending schema
Ok, the patch not to add functions.
> better
>
> ...(): encoding name too long
Fixed.
I found new bug in command/variable.c in parse_client_encoding(), nobody
probably never see this error:
if (pg_set_client_encoding(encoding))
{
elog(ERROR, "Conversion between %s and %s is not supported",
value, GetDatabaseEncodingName());
}
because pg_set_client_encoding() returns -1 for error and 0 as true.
It's fixed too.
IMHO it can be apply.
Karel
PS:
* following files are renamed:
src/utils/mb/Unicode/KOI8_to_utf8.map -->
src/utils/mb/Unicode/koi8r_to_utf8.map
src/utils/mb/Unicode/WIN_to_utf8.map -->
src/utils/mb/Unicode/win1251_to_utf8.map
src/utils/mb/Unicode/utf8_to_KOI8.map -->
src/utils/mb/Unicode/utf8_to_koi8r.map
src/utils/mb/Unicode/utf8_to_WIN.map -->
src/utils/mb/Unicode/utf8_to_win1251.map
* new file:
src/utils/mb/encname.c
* removed file:
src/utils/mb/common.c
--
Karel Zak <zakkr@zf.jcu.cz>
http://home.zf.jcu.cz/~zakkr/
C, PostgreSQL, PHP, WWW, http://docs.linux.cz, http://mape.jcu.cz
> pam_strerror() should be used a few more times, rather than just saying
> "Error!". Also, the configure.in snippet seems wrong. You add
> -I$pam_prefix/include/security to $INCLUDES and then you #include
> <security/pam_appl.h>. This whole thing is probably unnecessary, since
> PAM is a system library on the systems where it exists, so the headers
> and libraries are found automatically, unlike OpenSSL and
> Kerberos.
See attached revised patch. (I'm sure the configure.in stuff can be done
right/better, I'm just not enough of a autoconf guru to know what to
change it to.)
Dominic J. Eidson
- new millisecond (ms) and microsecond (us) support
- more robus parsing from string - used is separator checking for
non-exact formats like to_date('2001-9-1', 'YYYY-MM-DD')
- SGML docs are included
Karel Zak
flawed in the following ways:
1. Only returned columns that had a default value defined, rather than all
columns in a table
2. Used 2 * N + 1 queries to find out attributes, comments and typenames
for N columns.
By using some outer join syntax it is possible to retrieve all necessary
information in just one SQL statement. This means this version is only
suitable for PostgreSQL >= 7.1. Don't know whether that's a problem.
I've tested this function with current sources and 7.1.3 and patched both
jdbc1 and jdbc2. I haven't compiled nor tested the jdbc1 version though, as
I have no JDK 1.1 available.
Note the discussion in http://fts.postgresql.org/db/mw/msg.html?mid=1029626
regarding differences in obtaining comments on database object in 7.1 and
7.2. I was unable to use the following syntax (or similar ones):
select
...,
description
from
...
left outer join col_description(a.attrelid, a.attnum) description
order by
c.relname, a.attnum;
(the error was parse error at or near '(') so I had to paste the actual
code for the col_description function into the left outer join. Maybe
someone who is more knowledgable about outer joins might provide me with a
better SQL statement.
Jeroen van Vianen
includes windows.h, which #defines ERROR to 0. PostgreSQL's logging functions
define ERROR to -1. This patch redefines ERROR to -1 to avoid current or
future breakage of the logging functions.
Gerhard H?ring
> > -patches as well.
>
> The -Bdynamic probably ought to disappear.
That was there already, but I have no objections. I was trying to
make minimal changes.
Larry Rosenman
the JDBC driver.
I've done this by extracting it into a new method object called
QueryExecutor (should go into org/postgresql/core/) and then taking it
apart into different methods in that class.
A short summary:
* Extracted ExecSQL() from Connection into a method object called
QueryExecutor.
* Moved ReceiveFields() from Connection to QueryExecutor.
* Extracted parts of the original ExecSQL() method body into smaller
methods on QueryExecutor.
* Bug fix: The instance variable "pid" in Connection was used in two
places with different meaning. Both were probably in dead code, but it's
fixed anyway.
Anders Bengtsson
for the changed files and a few new files:
- test/jdbc2/BatchExecuteTest.java
- util/MessageTranslator.java
- jdbc2/PBatchUpdateException.java
As an aside, is this the best way to submit a patch consisting
of both changed and new files? Or is there a smarter cvs command
which gets them all in one patch file?
This patch fixes batch processing in the JDBC driver to be
JDBC-2 compliant. Specifically, the changes introduced by this
patch are:
1) Statement.executeBatch() no longer commits or rolls back a
transaction, as this is not prescribed by the JDBC spec. Its up
to the application to disable autocommit and to commit or
rollback the transaction. Where JDBC talks about "executing the
statements as a unit", it means executing the statements in one
round trip to the backend for better performance, it does not
mean executing the statements in a transaction.
2) Statement.executeBatch() now throws a BatchUpdateException()
as required by the JDBC spec. The significance of this is that
the receiver of the exception gets the updateCounts of the
commands that succeeded before the error occurred. In order for
the messages to be translatable, java.sql.BatchUpdateException
is extended by org.postgresql.jdbc2.PBatchUpdateException() and
the localization code is factored out from
org.postgresql.util.PSQLException to a separate singleton class
org.postgresql.util.MessageTranslator.
3) When there is no batch or there are 0 statements in the batch
when Statement.executeBatch() is called, do not throw an
SQLException, but silently do nothing and return an update count
array of length 0. The JDBC spec says "Throws an SQLException if
the driver does not support batch statements", which is clearly
not the case. See testExecuteEmptyBatch() in
BatchExecuteTest.java for an example. The message
postgresql.stat.batch.empty is removed from the language
specific properties files.
4) When Statement.executeBatch() is performed, reset the
statement's list of batch commands to empty. The JDBC spec isn't
100% clear about this. This behaviour is only documented in the
Java tutorial
(http://java.sun.com/docs/books/tutorial/jdbc/jdbc2dot0/batchupdates.html).
Note that the Oracle JDBC driver also resets the statement's
list in executeBatch(), and this seems the most reasonable
interpretation.
5) A new test case is added to the JDBC test suite which tests
various aspects of batch processing. See the new file
BatchExecuteTest.java.
Regards,
Ren? Pijlman
> Can someone research this and figure out what the proper solution for
> this is? Seems we are going around in circles if we keep
> adding/removing DLLIMPORT.
I believe that the attached patch is the correct solution -- I apologize
for the gyrations. With the attached patch, Cygwin libpq++ builds
cleanly again. The root cause was that DLLIMPORT was defaulting to
__declspec(dllimport) since BUILDING_DLL was *not* defined when building
the libpq++ DLL.
Unfortunately, to test my patch requires changing the following makefile:
src/interfaces/libpq++/examples/Makefile
and the #includes in all of the *.cc to build against the source tree
instead of the following hardcoded installation directory structure:
/usr/local/pgsql
I was able to manually build
src/interfaces/libpq++/examples/testlibpq0.exe
against my Cygwin libpq++ without errors. However, I have not tried to
actually test testlibpq0.exe.
Is this sufficient? Or, do you want me to clean up libpq++/examples too?
(Or, is it silly to even ask? :,)) Let me know how you want to proceed and
I will submit a patch to pgsql-patches.
Jason Tishler
Now with documentation update and disabling of UTF conversion for Tcl <=8.0
On Fri, 24 Aug 2001, Vsevolod Lobko wrote:
> On Thu, 23 Aug 2001, Tom Lane wrote:
>
> > > Is this looks better?
> >
> > It does, but one small gripe: the lack of semicolons will probably cause
> > pg_indent to mess up the indentation. (I know emacs' autoindent mode
> > will not work nicely with it, either.) Please set up the macros so that
> > you write
> >
> > UTF_BEGIN;
> > Tcl_DStringAppend(&unknown_src, UTF_E2U(part), -1);
> > UTF_END;
> >
> > and then I'll be happy.
>
> Attached revised patch
>
> > Your point about overhead is a good one, so I retract the gripe about
> > using a configure switch. But please include documentation patches to
> > describe the configure option in the administrator's guide (installation
> > section).
>
> This patch still uses configure switch for enabling feature.
>
> For enabling based on tcl version we have 2 posibilites:
> 1) having feature enabled by default, but in pltcl.c check for tcl
> version and disable it for old versions
> 2) enable or disable at configure time based on tcl version, but there
> are problem - current configure don't checks for tcl version at all
> and my configure skills not enought for adding this
>
Vsevolod Lobko
two additional files win32.mak and libpgtcl.def.
This patch allows to compile libpgtcl.dll on Windows
with tcl > 8.0. I've tested it on WinNT (VC6.0), SUSE Linux (7.0)
and Solaris 2.6 with tcl 8.3.3.
Mikhail Terekhov
really played it totally safe in my last suggestion, the system table might
pick up the msg but not the netmsg.dll, so better try both.
I also added a hex printout of the "errno" appended to all messages, that's
nicer.
If anyone hate my coding style, or that i'm using goto constructs, just tell
me, and i'll rework it into a nested if () thing.
Magnus Naeslund(f)
If there's anyone out there who's actually using datatype-defined
default values, this will be an incompatible change in behavior ...
but the old behavior was so broken that I doubt anyone was using it.
Standardize on %X/%X as the formatting for XLOG position display --- we
had a couple of different formats before, and none of 'em were as useful
as hex offsets IMHO.
available in freeSemMap. As noted by Tatsuo, this is now a likely
scenario for detecting MaxBackends-exceeded; if MaxBackends is a multiple
of PROC_NSEMS_PER_SET then we will fail here and not in sinval.c. The
cleanup path did not work correctly before, anyway.
system. Some systems did not understand the 'l' section, and in general
it wasn't entirely appropriate.
On SCO OpenServer, the man pages won't be installed at all until someone
figures out their man system.
Client headers are no longer in a subdirectory, since they have been made
namespace-clean.
Internal libpq headers are in a private subdirectory.
Server headers are in a private subdirectory. pg_config has a new option
to point there.
Since we're assuming a C++ compiler knows what 'bool' is, seems we
should assume it knows 'true' and 'false' too. This prevents problems
on some systems, per report from Leandro Fanzone.
for some reason displays a zero oid differently. Possibly we should
revert that schema change, but it's easy to make pg_dump accept both
spellings so I'll do that for now.
in interfaces/perl5 a brief while ago.
Also, since building PL/Perl without a shared libperl actually works on
some platforms we can enable it there to get some development happening.
I've only checked off linux right now, but others should be added in the
future.
longer compiles, due to objects being referenced in this patch that do
not exist in JDK1.1.
Barry Lind
---------------------------------------------------------------------------
The JDBC driver requires
permission java.net.SocketPermission "host:port", "connect";
in the policy file of the application using the JDBC driver
in the postgresql.jar file. Since the Socket() call in the
driver is not protected by AccessController.doPrivileged() this
permission must also be granted to the entire application.
>>>>
>>>> permission java.net.SocketPermission "host:port", "connect";
>>>>
>>>>in the policy file of the application using the JDBC driver
>>>>in the postgresql.jar file. Since the Socket() call in the
>>>>driver is not protected by AccessController.doPrivileged() this
>>>>permission must also be granted to the entire application.
>>>>
>>>>The attached diff fixes it so that the connect permission can be
>>>>restricted just the the postgresql.jar codeBase if desired.
David Daney
org.postgresql.util.Serialize and org.postgresql.jdbc2.PreparedStatement
that fixes the ability to "serialize" a simple java class into a
postgres table.
The current cvs seems completely broken in this support, so the patch
puts it into working condition, granted that there are many limitations
with serializing java classes into Postgres.
The code to do serialize appears to have been in the driver since
Postgres 6.4, according to some comments in the source. My code is not
adding any totally new ability to the driver, rather just fixing what
is there so that it actually is usable. I do not think that it should
affect any existing functions of the driver that people regularly
depend on.
The code is activated if you use jdbc2.PreparedStatement and try to
setObject some java class type that is unrecognized, like not String or
not some other primitive type. This will cause a sequence of function
calls that results in an instance of Serialize being instantiated for
the class type passed. The Serialize constructor will query pg_class
to see if it can find an existing table that matches the name of the
java class. If found, it will continue and try to use the table to
store the object, otherwise an SQL exception is thrown and no harm is
done. Serialize.create() has to be used to setup the table for a java
class before anything can really happen with this code other than an
SQLException (unless by some freak chance a table exists that it thinks
it can use).
I saw a difference in Serialize.java between 7.1.3 and 7.2devel that I
didn't notice before, so I had to redo my changes from the 7.2devel
version (why I had to resend this patch now). I was missing the
fixString stuff, which is nice and is imporant to ensure the inserts
will not fail due to embedded single quote or unescaped backslashes. I
changed that fixString function in Serialize just a little since there
is no need to muddle with escaping newlines: only escaping single quote
and literal backslashes is needed. Postgres appears to insert newlines
within strings without trouble.
buffer manager with 'pg_clog', a specialized access method modeled
on pg_xlog. This simplifies startup (don't need to play games to
open pg_log; among other things, OverrideTransactionSystem goes away),
should improve performance a little, and opens the door to recycling
commit log space by removing no-longer-needed segments of the commit
log. Actual recycling is not there yet, but I felt I should commit
this part separately since it'd still be useful if we chose not to
do transaction ID wraparound.
This patch moves the logic that looks up TypeOid, PGTypeName, and
SQLTypeName from Field to Connection. It is moved to connection since
it needs to differ from the jdbc1 to jdbc2 versions and Connection
already has different subclasses for the two driver versions. It also
made sense to move the logic to Connection as some of the logic was
already there anyway.
Barry Lind
following email.
> > The problem: When I call getBigDecimal() on a ResultSet, it
> > sometimes throws an exception:
> >
> > Bad BigDecimal 174.50
> > at org.postgresql.jdbc2.ResultSet.getBigDecimal(ResultSet.java:373)
> > at org.postgresql.jdbc2.ResultSet.getBigDecimal(ResultSet.java:984)
> > ...blah blah blah...
> > org.postgresql.util.PSQLException: Bad BigDecimal 174.50
Barry Lind
so it may be a transit problem. Also removed the 'txt' suffix
in case that was confusing some transport layer trying to be
too inteligent for our own good.
This may have been because the Array.java class from the
previous patch didn't seem to have made it into the snapshot
build for some reason. This patch should at least fix that issue.
Greg Zoller
> It seems that win9x doesn't have the "netmsg.dll" so it defaults to "normal"
> FormatMessage.
> I wonder if one could load wsock32.dll or winsock.dll on those systems
> instead of netmsg.dll.
>
> Mikhail, could you please test this code on your nt4 system?
> Could someone else test this code on a win98/95 system?
>
> It works on win2k over here.
It works on win2k here too but not on win98/95 or winNT.
Anyway, attached is the patch which uses Magnus's my_sock_strerror
function (renamed to winsock_strerror). The only difference is that
I put the code to load and unload netmsg.dll in the libpqdll.c
(is this OK Magnus?).
Mikhail Terekhov
pgsql-hackers. pg_opclass now has a row for each opclass supported by each
index AM, not a row for each opclass name. This allows pg_opclass to show
directly whether an AM supports an opclass, and furthermore makes it possible
to store additional information about an opclass that might be AM-dependent.
pg_opclass and pg_amop now store "lossy" and "haskeytype" information that we
previously expected the user to remember to provide in CREATE INDEX commands.
Lossiness is no longer an index-level property, but is associated with the
use of a particular operator in a particular index opclass.
Along the way, IndexSupportInitialize now uses the syscaches to retrieve
pg_amop and pg_amproc entries. I find this reduces backend launch time by
about ten percent, at the cost of a couple more special cases in catcache.c's
IndexScanOK.
Initial work by Oleg Bartunov and Teodor Sigaev, further hacking by Tom Lane.
initdb forced.
> Shouldn't
>
> throw new PSQLException("metadata unavailable");
>
> in getTypeInfo() be something like:
>
> throw new PSQLException("postgresql.meta.unavailable");
>
> to allow translation of the error message in the
> errors*.properties files?
You're right. Attached is an updated patch that also includes a message
in error.properties. I've attempted a French message in
errors_fr.properties but beware that I haven't written French in quite a
few years. Don't know Italian, German, or Dutch so I can't do those.
Liam Stewart
SQLxxxx() to PGAPI_xxxx().
2) Handle an escaped date/time format as a parameter.
3) Improve the tuple allocation a little.
4) The preparation of ODBC 3.0 a little.
5) Updatable cursors(may be deprecated before long).
attempt at a patch to 7.1.2 to support Array.
[I think I've solved the mangled patch problem. Hotmail seems to
try to format the text file, so gzipping it should solve this
problem.]
In this patch I've incorporated Barry's feedback. Specifically:
1) OIDs are no longer hard-coded into Array.java. In order to
support this change I added a getOID(String) method to Field.java
which receives a PostgreSQL field type and returns a value from
java.sql.Types. I couldn't get away from using OIDs altogether
because the JDBC spec for Array specifies that some methods return
a ResultSet. This requires I construct Field objects,
which means I need OIDs. At least this approach doesn't hard
code these values. A Hashtable cache has been added to Field
so that an SQL lookup isn't necessary (following the model already
in Field.java).
2) Rewired the base formatting code in ResultSet.java to use 'to'
methods, which are then exposed as static methods in ResultSet.
These methods are used in Array to format the data without
duplications in the code.
3) Artifact call to first() in ResultSet.getArray() removed.
Greg Zoller
includes two changes in the JDBC driver:
1) When connected to a backend >= 7.2: use obj_description() and
col_description() instead of direct access to pg_description.
2) In DatabaseMetaData.getTables()/getColumns()/getProcedures():
when there is no comment on the object, return null in the
REMARKS column of the ResultSet, instead of the default string
"no remarks".
Change 2 first appeared as a side-effect of change 1, but it is
actually more compliant with the JDBC spec: "String object
containing an explanatory comment on the table/column/procedure,
which may be null". The default string "no remarks" was strictly
speaking incorrect, as it could not be distinguished from a real
user comment "no remarks". So I removed the default string
completely.
Change 2 might break existing code that doesn't follow the JDBC
spec and isn't prepared to handle a null in the REMARKS column
of getTables()/getColumns()/getProcedures.
Patch tested with jdbc2 against both a 7.1 and a CVS tip
backend. I did not have a jdbc1 environment to build and test
with, but since the touched code is identical in jdbc1 and jdbc2
I don't foresee any problems.
Regards,
Ren? Pijlman
has an alias SERIAL4 and a sister SERIAL8. SERIAL8 is just the same
except the created column is type int8 not int4.
initdb forced. Note this also breaks any chance of pg_upgrade from 7.1,
unless we hack up pg_upgrade to drop and recreate sequences. (Which is
not out of the question, but I don't wanna do it.)
Allow pg_shadow to be MD5 encrypted.
Add ENCRYPTED/UNENCRYPTED option to CREATE/ALTER user.
Add password_encryption postgresql.conf option.
Update wire protocol version to 2.1.
for speed reasons; its result type also changes to int8. avg() on these
datatypes now accumulates the running sum in int8 for speed; but we still
deliver the final result as numeric, so that fractional accuracy is
preserved.
count() now counts and returns in int8, not int4. I am a little nervous
about this possibly breaking users' code, but there didn't seem to be
a strong sentiment for avoiding the problem. If we get complaints during
beta, we can change count back to int4 and add a "count8" aggregate.
For that matter, users can do it for themselves with a simple CREATE
AGGREGATE command; the int4inc function is still present, so no C hacking
is needed.
Also added max() and min() aggregates for OID that do proper unsigned
comparison, instead of piggybacking on int4 aggregates.
initdb forced.
a tad sloppy about generating the targetlist for some nodes, by generating
a tlist entry that claimed to be a constant when the value wasn't actually
constant. This caused setrefs.c to do the wrong thing later on.
syntax for language names (instead of 'string').
createlang now handles the case where a second language uses the same call
handler as an already installed language (e.g., plperl/plperlu).
droplang now handles the reverse case, i.e., dropping a language where
the call handler is still used by another language. Moreover, droplang
can now be used to drop any user-defined language, not just the supplied
ones.
Don't hardcode the maximum accepted server version, use PG_VERSION instead.
Install a notice processor so notices are handled like error messages.
Word smithing.
default, but OIDS are removed from many system catalogs that don't need them.
Some interesting side effects: TOAST pointers are 20 bytes not 32 now;
pg_description has a three-column key instead of one.
Bugs fixed in passing: BINARY cursors work again; pg_class.relhaspkey
has some usefulness; pg_dump dumps comments on indexes, rules, and
triggers in a valid order.
initdb forced.
* Merges identical code from org.postgresql.jdbc[1|2].Statement into
org.postgresql.Statement.
* Moves escapeSQL() method from Connection to Statement (the only place
it's used)
* Minor cleanup of the new isolation level stuff.
* Minor cleanup of version string handling.
Anders Bengtsson
(as proposed in http://fts.postgresql.org/db/mw/msg.html?mid=1028327)
2. support for 'pass-by-value' arguments - to test this
we used special opclass for int4 with values in range [0-2^15]
More testing will be done after resolving problem with
index_formtuple and implementation of B-tree using GiST
3. small patch to contrib modules (seg,cube,rtree_gist,intarray) -
mark functions as 'isstrict' where needed.
Oleg Bartunov
Here is a context diff from latest cvs
And I see why you couldn't apply the last diff, the setCatalog diff has
been backed out, that was causing the compile problem in the first
place.
This following one needs to be applied to allow the current cvs to
compile
Dave Cramer
clauses are equal(), before trying to match them up using btree opclass
inference rules. This allows it to recognize many simple cases involving
non-btree operations, for example 'x IS NULL'. Clean up code a little.
Make sure it exits immediately when collector process dies --- in old code,
buffer process would hang around and compete with the new buffer process
for packets. Make sure it doesn't block on writing the pipe when the
collector falls more than a pipeload behind. Avoid leaking pgstats FDs
into every backend.
cvs.
The Debian bug report says, "The upstream source makes use of NOFILE
unconditionalized. As the Hurd doesn't have an arbitrary limit on the
number of open files, this is not defined. But _SC_OPEN_MAX works fine
and returns 1024 (applications can increase this as they want), so I
suggest the below diff. Please forward this upstream, too."
Oliver Elphick
check
> in convert.c
> does not consider the fact that the value in the field has been altered to
> be a '1' if the
> backend handed it a 't'. The net result being that the first row on any
> subsequent queries
> has all it's boolean set to 0.
Aidan Mountford
1) improves performance of commit/rollback by reducing number of round
trips to the server
2) uses 7.1 functionality for setting the transaction isolation level
3) backs out a patch from 11 days ago because that code failed to
compile under jdk1.1
Details:
1) The old code was doing the following for each commit:
commit
begin
set transaction isolation level xxx
thus a call to commit was performing three round trips to the database.
The new code does this in one round trip as:
commit; begin; set transaction isolation level xxx
In a simple test program that performs 1000 transactions (where each
transaction does one simple select inside that transaction) has the
following before and after timings:
Client and Server on same machine
old new
--- ---
1.877sec 1.405sec 25.1% improvement
Client and Server on different machines
old new
--- ---
4.184sec 2.927sec 34.3% improvement
(all timings are an average of four different runs)
2) The driver was using 'set transaction isolation level xxx' at the
begining of each transaction, instead of using the new 7.1 syntax of
'set session characteristics as transaction isolation level xxx' which
only needs to be done once instead of for each transaction. This is
done conditionally (i.e. if server is 7.0 or older do the old behaviour,
else do the new behaviour) to not break backward compatibility. This
also required the movement of some code to check/test database version
numbers from the DatabaseMetaData object to the Connection object.
3) Finally while testing, I discovered that the code that was checked in
11 days ago actually didn't compile. The code in the patch for
Connection.setCatalog() used Properties.setProperty() which only exists
in JDK1.2 or higher. Thus compiling the JDBC1 driver failed as this
method doesn't exist. Thus I backed out that patch.
Barry Lind
platforms system(2) gets confused unless the signal handler is set to
SIG_DFL, not SIG_IGN. pgstats.c now uses pqsignal() as it should,
not signal(). Also, arrange for the stats collector process to show
a reasonable ID in 'ps', rather than looking like a postmaster.
Dump the alignment and storage information for user-defined types (how'd
that manage to slip through the cracks?), and don't dump 'shell' types
that don't have typisdefined set. Fix badly broken logic for dependencies
of type definitions (did not work for more than one user-defined type...).
Avoid memory leakage within pg_dump by being more careful to release
storage used by PQExpBuffer objects.
exec_eval_simple_expr shortcut, which was diked out in 7.1 because it
leaked too much space. CVS tip now leaks no memory in Chris Ruprecht's
example, which formerly leaked to the tune of 500 MB. (Much of this
is work that Jan already did; this commit just cleans up around the
edges.)
Sorry I don't have the original around to make a quick diff, but its a very small change... I think this should be in the next release, there's no reason not to have it.
its a function with no expected arguments, so you can use it like:
spi_exec "INSERT INTO mytable(columns...) VALUES(values..)"
set oid [spi_lastoid]
spi_exec "SELECT mytable_id from mytable WHERE oid=$oid"
It just didn't make sense for me to use plpgsql and pltcl, or just screw
them both and use SPI from C.
bob@redivi.com
that call is not needed to prepare for SO_PEERCRED. Also, simplify code
so that #ifdef SO_PEERCRED appears in only one place, to make it easier
to support other platforms with variants of this capability.
points out how silly it is to use Autoconf to test for a preprocessor
symbol, when one can equally easily #ifdef on the symbol itself.
Accordingly, revert configure to prior state and do it that way.
system supports SO_PEERCRED requests for Unix sockets. This is an
amalgamation of patches submitted by Helge Bahmann and Oliver Elphick,
with some editorializing by yours truly.
platform.
TIOCGWINSZ is defined as follows:
Linux asm/ioctls.h
FreeBSD sys/ttycom.h. This file is included by sys/ioctl.h.
Solaris sys/termios.h
This patch tells print.c to know TIOCGWINSZ on Solaris platform. Same code is
founded in src/bin/psal/common.c.
Kenji Sugita
number in the data structure so that we can give at least a minimally
useful idea of where the mistake is when we issue syntax error messages.
Move the ClientAuthentication() call to where it should have been in
the first place, so that postmaster memory releasing can happen in a
reasonable place also. Update obsolete comments, correct one real bug
(auth_argument was not picked up correctly).
more care with resjunk tlist entries than it was doing. The original
coding ignored resjunk entries entirely, but a resjunk entry that is
in either the distinctClause or sortClause lists indicates that DISTINCT
ON was used. It's important for ruleutils.c to get this right, else we
may dump views using DISTINCT ON incorrectly.
has a DISTINCT ON clause, per bug report from Anthony Wood. While at it,
improve the DISTINCT-ON-clause recognizer routine to not be fooled by out-
of-order DISTINCT lists.
echo "command" | postgres
to the style
postgres <<EOF
command
EOF
This makes the script more legible (IMHO anyway) by reducing the need
to escape quotes, and allows us to execute successive SQL commands in
a single standalone-backend run, rather than needing to start a new
standalone backend for each command. With all the CREATE VIEWs that
are getting done now, this makes for a rather substantial reduction
in the runtime of initdb. (Some of us do initdb often enough to care
how long it runs ;-).)
connection implementations (org.postgresql.jdbc[1|2].Connection) into
their superclass (org.postgresql.Connection).
It also changes the close() methods of Connection and PG_Stream, so that
PG_Stream no longer is responsible for sending the termination packet 'X'
to the backend. I figured that protocol-level stuff like that belonged in
Connection more than in PG_Stream.
Anders Bengtsson
in Connection - note: I've updated setCatalog(String catalog) from my previous
diff so it checks whether it is already connected to the specified catalog.
Jason Davies
Here's a patch against the current CVS. The changes from the previous
patch are mostly related to the changed interface for PG_Stream.
Anders Bengtsson
changes on this new source to make non-blocking connection work. I
tested it, and PQSendQuery and PQGetResult are working fine.
In win32.h I added one line:
#define snprintf _snprintf
Darko Prenosil
functions do not set errno, so some normal conditions are treated as
fatal errors. e.g. fetching large tuples fails, as at some point recv()
returns EWOULDBLOCK. here's a patch, which replaces errno with
WSAGetLastError(). i've tried to to affect non-win32 code.
Dmitry Yurtaev
rather than deleting them only to have to create more. Steady state
is 2*CHECKPOINT_SEGMENTS + WAL_FILES + 1 segment files, which will
simply be renamed rather than constantly deleted and recreated.
To make this safe, added current XLOG file/offset number to page
header of XLOG pages, so that an un-overwritten page from an old
incarnation of a logfile can be reliably told from a valid page.
This change means that if you try to restart postmaster in a CVS-tip
database after installing the change, you'll get a complaint about
bad XLOG page magic number. If you don't want to initdb, run
contrib/pg_resetxlog (and be sure you shut down the old postmaster
cleanly).
in GetSnapshotData, GetNewTransactionId, CommitTransaction, AbortTransaction,
etc. Correct race condition in transaction status testing in
HeapTupleSatisfiesVacuum --- this wasn't important for old VACUUM with
exclusive lock on its table, but it sure is important now. All per
pghackers discussion 7/11/01 and 7/12/01.