Add a changelog file for 7.1.3 from 7.1.2
authorMarc G. Fournier <scrappy@hub.org>
Fri, 17 Aug 2001 17:10:16 +0000 (17:10 +0000)
committerMarc G. Fournier <scrappy@hub.org>
Fri, 17 Aug 2001 17:10:16 +0000 (17:10 +0000)
ChangeLogs/ChangeLog-7.1.2-7.1.3 [new file with mode: 0644]

diff --git a/ChangeLogs/ChangeLog-7.1.2-7.1.3 b/ChangeLogs/ChangeLog-7.1.2-7.1.3
new file mode 100644 (file)
index 0000000..5cfb86a
--- /dev/null
@@ -0,0 +1,142 @@
+2001-08-16 11:21  momjian
+
+       * src/interfaces/python/pgmodule.c: This patch fixes the well-known
+       but unfixed bug that fetchone() always returns the first result in
+       the DB-API compliant wrapper. It turned out that the bug was way
+       down in the C code.
+       
+       Gerhard Häring
+
+2001-08-13 16:09  tgl
+
+       * configure: Hmm, looks like Bruce's auto-run-autoconf script
+       doesn't work for old branches...
+
+2001-08-13 15:54  tgl
+
+       * src/backend/executor/nodeHash.c: Make hashjoin give the right
+       answer with toasted input data.
+
+2001-08-09 15:22  inoue
+
+       * src/: backend/access/heap/heapam.c, backend/executor/execMain.c,
+       backend/utils/adt/tid.c, include/access/heapam.h: fix my old fault.
+
+2001-08-09 12:39  momjian
+
+       * configure.in, register.txt, doc/bug.template,
+       doc/src/sgml/version.sgml, src/include/config.h.win32,
+       src/interfaces/libpq/libpq.rc: Update branding for 7.1.3, no
+       HISTORY yet.
+
+2001-08-09 09:52  tgl
+
+       * src/tutorial/Makefile: Fix Cygwin build, per Jason Tishler.
+
+2001-08-08 18:32  tgl
+
+       * src/backend/commands/copy.c: Back-patch fix to disallow COPY
+       TO/FROM a view (or anything else that's not a plain relation).
+
+2001-08-08 18:25  tgl
+
+       * src/backend/utils/init/miscinit.c: Back-patch fix to prevent
+       infinite loop when $PGDATA is not writable.
+
+2001-08-07 14:36  momjian
+
+       * src/: backend/port/beos/support.c, backend/port/dynloader/beos.c,
+       include/port/beos.h: Commit BEOS patch to 7.1.X.
+
+2001-08-03 16:14  tgl
+
+       * src/bin/pg_dump/: pg_dump.c, pg_dump.h: Back-patch fixes for
+       dumping user-defined types and dumping comments on views.
+
+2001-07-31 14:39  tgl
+
+       * src/: backend/optimizer/path/allpaths.c,
+       backend/optimizer/util/clauses.c, backend/utils/adt/ruleutils.c,
+       include/optimizer/clauses.h: Fix optimizer to not try to push WHERE
+       clauses down into a sub-SELECT that 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.  Also, back-patch earlier fix to not push
+       down into sub-SELECT with LIMIT.
+
+2001-07-29 18:12  tgl
+
+       * src/bin/pg_dump/pg_dump.c: Arrange for GRANT/REVOKE on a view to
+       be dumped at the right time, namely after the view definition
+       rather than before it.  Bug introduced in 7.1 by changes to dump
+       stuff in OID ordering.
+
+2001-07-05 22:13  ishii
+
+       * doc/src/sgml/backup.sgml: Fix typo. createdb -t --> createdb -T
+
+2001-07-02 15:34  tgl
+
+       * src/test/regress/expected/: abstime-solaris-1947.out,
+       abstime.out: In any case, it seems the REL7_1 branch needs the
+       update too...
+
+2001-06-29 12:34  tgl
+
+       * src/backend/commands/vacuum.c: Fix longstanding error in VACUUM:
+       sometimes would examine a buffer page after writing/unpinning it. 
+       An actual failure is unlikely, unless the system is tremendously
+       short of buffers ... but a bug is a bug.
+
+2001-06-12 21:02  tgl
+
+       * src/pl/plpgsql/src/pl_exec.c: Back-patch fix for attempt to pfree
+       a value that's not palloc'd (it's a field of a tuple).  I see Jan
+       has already fixed this in current sources, but 7.1.* is pretty
+       badly broken here.
+
+2001-06-12 14:54  tgl
+
+       * src/backend/rewrite/rewriteHandler.c: Repair problem with
+       multi-action rules in combination with any nontrivial manipulation
+       of rtable/jointree by planner.  Rewriter was generating actions
+       that shared rtable/jointree substructure, which caused havoc when
+       planner got to the later actions that it'd already mucked up.
+
+2001-06-06 13:18  tgl
+
+       * src/backend/access/transam/xlog.c: Back-patch change to not keep
+       WAL segments just for UNDO information.
+
+2001-05-31 17:50  momjian
+
+       * doc/src/sgml/release.sgml: Forgot SGML section section id tag for
+       7.1.
+
+2001-05-31 13:33  tgl
+
+       * src/backend/utils/adt/ri_triggers.c: RI triggers would fail for
+       datatypes using old-style equal function, because cached fmgr info
+       contained reference to a shorter-lived data structure.  Also guard
+       against possibility that fmgr_info could fail, leaving an
+       incomplete entry present in the hash table.
+
+2001-05-27 21:01  ishii
+
+       * src/backend/utils/mb/conv.c: Fix a message error in utf_to_local
+
+2001-05-22 10:46  thomas
+
+       * src/interfaces/jdbc/org/postgresql/jdbc2/ResultSet.java: Patch
+       from Barry Lind to correctly decode time zones in timestamp
+       results.   Without patch, the time zone field is ignored and the
+       returned time is   not correct.   Already applied to the
+       development tree...
+
+2001-05-18 12:00  tgl
+
+       * src/include/catalog/pg_aggregate.h: Correct recently-broken
+       avg(interval) definition.  We can't force an initdb to fix this in
+       7.1 installations, but it seems better to be shipping a correct
+       entry than a wrong one.
+