Bruce Momjian [Fri, 12 May 2000 20:54:22 +0000 (20:54 +0000)]
Fix the off by one errors in ResultSet from 6.5.3, and more.
I'm including a diff of
postgresql-7.0/src/interfaces/jdbc/org/postgresql/jdbc2/ResultSet.java.
I've clearly marked all the fixes I did. Would *someone* who has access
to the cvs please put this in?
Bruce Momjian [Fri, 12 May 2000 18:51:59 +0000 (18:51 +0000)]
This is the second time I've answered this exact same problem in two
days. It seems to be a FAQ, and I think I know why. When creating a 'c'
language function, CREATE FUNCTION is fed the shared object filename,
and seems to succeed. Only when trying to use the function is an error
thrown, by which time the coder thinks something's wrong with executing
the code, not with loading it.
I think I once saw it proposed to load shared objects at function creation
time, but that idea was shot down on the grounds of resident memory bloat,
ISTR. Here's a patch for a compromise: all it does is stat() the file,
just like the loader code does, so that the errors caused by non existent
files, and no directory 'x' permissions (the most common ones, it seems),
get caught while the developer is still thinking about code loading. It
doesn't catch all errors (like the code not being readable by the postgres
user) but seems to catch the most common, without actually opening the file.
Tom Lane [Fri, 12 May 2000 16:10:09 +0000 (16:10 +0000)]
Squash some more CLUSTER bugs. Never has worked on multiple-column
indexes, apparently, nor on functional indexes with more than one input
column (force of natts = 1 was in the wrong branch of IF statement).
Coredumped if source relation contained any uncommitted tuples, due to
failure to test for success return from heap_fetch. Fetched tuple
was passed directly to heap_insert, which clobbers the TID and commit
status in the tuple header it's given, which meant that the source
relation's tuples all got trashed as the copy proceeded. Abort partway
through, and you're left with a lot of missing tuples.
I wonder what else is lurking here ...
Marc G. Fournier [Fri, 12 May 2000 14:33:08 +0000 (14:33 +0000)]
this fixes the bug where setting the entry in he process table no longer works
under FreeBSD ... basically, if setproctitle() exists, use it ...
the draw back right now is the PS_SET_STATUS stuff doesn't work, but am looking
into that one right now ... at lesat now you can see who is connecting where
and from where ...
Tom Lane [Fri, 12 May 2000 01:33:56 +0000 (01:33 +0000)]
Repair list-vs-node confusion that resulted in failure for INNER JOIN ON.
Make it behave correctly when there are more than two tables being
joined, also. Update regression test expected outputs.
Bruce Momjian [Thu, 11 May 2000 04:00:00 +0000 (04:00 +0000)]
Oops, plpgsql didn't have the datetime->timestamp and timespan->interval
mappings. In fact, it had them backward because it was using the 6.5.*
code. Copied them from parser/gram.y, so it is fixed now. Looks like
our first 7.0.1 fix. Oops, seems Tom has beat me to it as I was typing
this.
Tom Lane [Thu, 11 May 2000 03:54:18 +0000 (03:54 +0000)]
Fix CLUSTER ... or at least undo the bit-rot it's suffered since 6.5.
It's still pretty fundamentally bogus :-(.
Freebie side benefit: ALTER TABLE RENAME works on indexes now.
Tom Lane [Fri, 5 May 2000 03:08:20 +0000 (03:08 +0000)]
Accept pg_group as well as pg_shadow data from dumpall script.
Rearrange handling of VACUUMs so that they are certain to be executed
as superuser not some random user; also, do not forget to vacuum
template1 itself.
Tatsuo Ishii [Tue, 2 May 2000 08:13:08 +0000 (08:13 +0000)]
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.
Tom Lane [Sun, 30 Apr 2000 21:29:23 +0000 (21:29 +0000)]
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. I was able to provoke an Assert failure when compiled with
asserts + -DCLOBBER_FREED_MEMORY, if I did something that would cause
an error to be reported by the backend large-object code, because indeed
that code operates in a context that gets deleted partway through xact
abort --- and CurrentMemoryContext was still pointing at it! Boo hiss.
Tom Lane [Fri, 28 Apr 2000 05:07:34 +0000 (05:07 +0000)]
Setting statistic options from SET PG_OPTIONS caused a backend crash
because StatFp never got set in that case. Set it immediately before
use to eliminate such problems.
Tom Lane [Fri, 28 Apr 2000 00:12:44 +0000 (00:12 +0000)]
plpgsql RAISE statement was careless about the possibility of a NULL
field value being displayed; produced coredump instead of the expected
<NULL> display.
Tom Lane [Thu, 27 Apr 2000 20:32:41 +0000 (20:32 +0000)]
Except_Intersect_Rewrite() failed to ignore resjunk targetlist entries,
thus causing failure if one sub-select had resjunk entries that the other
did not (cf. bug report from Espinosa 4/27/00).
Tom Lane [Thu, 27 Apr 2000 18:35:04 +0000 (18:35 +0000)]
Repair problem noted by Elphick: make_rels_by_joins failed to handle
cases where joinclauses were present but some joins have to be made
by cartesian-product join anyway. An example is
SELECT * FROM a,b,c WHERE (a.f1 + b.f2 + c.f3) = 0;
Even though all the rels have joinclauses, we must join two of them
in cartesian style before we can use the join clause...
Tom Lane [Thu, 27 Apr 2000 02:59:17 +0000 (02:59 +0000)]
-D switch to postmaster should override any PGDATA environment variable
that might be hanging about. Now it does ... amazing nobody noticed
this before ...
Tom Lane [Wed, 26 Apr 2000 23:39:10 +0000 (23:39 +0000)]
Tweak outUnique to include uniqColIdx[] field in the printout. This does
not cause any compatibility problems because stored rules don't contain
plan nodes --- in fact, we don't even have a readfunc for Unique nodes.
Tom Lane [Wed, 26 Apr 2000 23:35:34 +0000 (23:35 +0000)]
On HPUX, shl_load should be called with options BIND_IMMEDIATE rather
than BIND_DEFERRED. That way, if the loaded library has unresolved
references, shl_load fails cleanly. As we had it, shl_load would
succeed and then the dynlinker would call abort() when we try to call
into the loaded library. abort()ing a backend is uncool.
Tom Lane [Sun, 23 Apr 2000 04:26:32 +0000 (04:26 +0000)]
Our test to see if we had permission to install into Perl5 install area
always failed if Perl makefile's INSTALLSITELIB variable was specified
in terms of another variable. Fix by adding an echo-installdir target
to the Perl makefile, which the upper-level Makefile can invoke.
Tom Lane [Sun, 23 Apr 2000 02:08:33 +0000 (02:08 +0000)]
Update obsolete info in CREATE INDEX ref page. (I had fixed the text
description in indices.sgml, but missed the near-duplicate prose in
the reference page...)