TablespaceCreateDbspace should function normally even on platforms that do not

have symlinks (ie, Windows).  Although it'll never be called on to do anything
useful during normal operation on such a platform, it's still needed to
re-create dropped directories during WAL replay.
This commit is contained in:
Tom Lane 2006-03-29 15:15:43 +00:00
parent 0a971e2f20
commit 2154e1c11e
1 changed files with 5 additions and 6 deletions

View File

@ -37,7 +37,7 @@
*
*
* IDENTIFICATION
* $PostgreSQL: pgsql/src/backend/commands/tablespace.c,v 1.32 2006/03/24 04:32:13 tgl Exp $
* $PostgreSQL: pgsql/src/backend/commands/tablespace.c,v 1.33 2006/03/29 15:15:43 tgl Exp $
*
*-------------------------------------------------------------------------
*/
@ -81,9 +81,6 @@ static void set_short_version(const char *path);
* object in the tablespace, create the subdirectory. If the subdirectory
* already exists, just fall through quietly.
*
* If tablespaces are not supported, this is just a no-op; CREATE DATABASE
* is expected to create the default subdirectory for the database.
*
* isRedo indicates that we are creating an object during WAL replay.
* In this case we will cope with the possibility of the tablespace
* directory not being there either --- this could happen if we are
@ -91,11 +88,14 @@ static void set_short_version(const char *path);
* We handle this by making a directory in the place where the tablespace
* symlink would normally be. This isn't an exact replay of course, but
* it's the best we can do given the available information.
*
* If tablespaces are not supported, you might think this could be a no-op,
* but you'd be wrong: we still need it in case we have to re-create a
* database subdirectory (of $PGDATA/base) during WAL replay.
*/
void
TablespaceCreateDbspace(Oid spcNode, Oid dbNode, bool isRedo)
{
#ifdef HAVE_SYMLINK
struct stat st;
char *dir;
@ -178,7 +178,6 @@ TablespaceCreateDbspace(Oid spcNode, Oid dbNode, bool isRedo)
}
pfree(dir);
#endif /* HAVE_SYMLINK */
}
/*