Well folks, we're in the last stretch here. I've just created a Release
Candidate 3, primarily to incorporate the doc changes recently commit'd by
Thomas Lockhart ...
The Release Candidate is meant to represent what the Final Release will
look like, pending any minor changes between now and release. Since I'd
prefer not to do a release on Friday, just before a weekend, unless there
are any objections, the Full Release is schedualed to be on Monday, May
8th, giving a few more days of testing by the developer community ...
As usual, Release Candidates are available at:
ftp://ftp.postgresql.org/pub
in both single and multi tar formats ...
--------
- Remove derived files from CVS.
- Remove bogus 'xid loop detected' check, which actually wasn't detecting
loops, but just arbitrarily failing at 1000 locks.
- Reset CurrentMemoryContext to TopMemoryContext at the beginning of error
cleanup, ie, as soon as we have caught the longjmp. This ensures that
current context will be a valid context throughout error cleanup. Before
it was possible that current context was pointing at a context that would
get deleted during cleanup, leaving any subsequent pallocs in deep
trouble.
- Modify getdatabaseencoding(), pg_encoding_to_char()
pg_char_to_encoding() in multibyte disbaled case so that it does not
throw an error, rather return HARD CODED default value (currently SQL_ASCII).
This would solve the "non-mb backend vs. mb-enabled frontend" problem.
