git.postgresql.org Git - postgresql.git/log
Added comment to readline defines.
Moved readline stuff into Makefile.global.
Apply usage patches for European Dates patch from Keith
Quick patch from Andrew to recognize nextstep vs next as a port
Slight change to nabstime.c so that configure is able to handle a system
whereby timezone isn't an int, but tzset() exists...
This isn't a definitive fix, as there is probably an easier way of
fixing the bug...
Convert MISSING_SYSCONF to !HAVE_SYSCONF for autoconf
From: Keith Parks
More removals for next/nextstep. Fix IPC lib for them.
Change next to nextstep where missed.
|From: Keith Parks <emkxp01@mtcc.demon.co.uk>
|Subject: [PATCH] adding SYS_TIME just for fun.
|
|Hi,
|
|Whilst I was playing round with the European dates patch I noticed the sysfunc()
|that allows you to do :-
|
|create table test ( da date);
|insert into test values (SYS_DATE);
|
|and have the current system date inserted.
|
|So I thought it would be nice to have the SYS_TIME facility too.
|
|I've cloned the function and changed a few things and there you have it,
|you can now do:
|
|create table test2 ( ti time);
|insert into test2 values (SYS_TIME);
Remove #include "rlstubs.h", since it doesn't exist anymore
Linux defines MAXINT in values.h, which causes an error when compiling.
Wrap it in an #ifndef to prevent this
Bring in a patch from Keith Parks to move the use of European dates
from a #define to a run-time option '-e'
Man page was updated to reflect new option
Fix typo in struct name.
Remove rlstubs.[ch], since they weren't actually doing anything that couldn't
be #ifdef'd into psql.c itself
From what I can tell, if USE_READLINE is true or false, psql works under
FreeBSD, without configure. Now to test it *again* under sparc_solaris
with configure and see if it works...
Okay, this should pretty much clean up the psql/readline/history mess.
Added a README.readline file until configure is integrated to *try* and
explain the way things stand.
Removed a stray configure .in file
Why can't everyone be standard? :(
Add some code to make psql happy with a system where there is *no*
history code available, but there is readline code...
Modifications to handle the situation where readline.h exists, but
history.h doesn't...previously, it was assumed that both existed, or
didn't exist...but this assumption fails on the one sparc_solaris box
that I have access to, and could exist in other circumstances
Some fixes for using configure under sparc_solaris
would you mind committing the following changes for me? (the first
bug causes compilation to fail on alpha, the second causes a compiler
in this environment
More autosize structure error checks.
The check should be for getrusage(), not rusage() ... change the defines
to reflect what configure is going to define when the time comes
Remove -lsocket -lnsl from X11_LIB, since it should be added on a port-by-port
basis, as its only specific to certain OSs
Disable -Werror by default.
Developers can add it to their Makefile.custom...again, it causes sooooo many
more problems then its worth, from an end-user standpoint.
remove asm/bitops.h, as its causing more problems then it fixes.
Add YACC= bison -y as per Andrew's request...
Change how readline support is included in psql.c ...
See message to hackers@ mailing list concerning this...
Restructure padding to handle structure already 128 bytes(alpha).
Switch over NEED_RUSAGE to HAVE_RUSAGE for configure
I reversed a bunch of the #define's inadvertantly...
Fixed them, and am re-committing
Missed this when I removed the rest of the configure stuff
Change NEED_SIG_JMP to HAVE_SIGSETJMP in preparation for configure
Another switch for configure: NEED_SYS_SELECT_H to HAVE_SYS_SELECT_H
Convert NEED_{RINT,CBRT,ISINF} to HAVE_* in prepration for configure...
Replace occurances of USE_{LIMITS,VALUES}_H with HAVE_* in preparation
for switch over to configure
Remove from include/config.h:
#if defined(aix)
#define TERMIOS_H_LOCATION <termios.h>
#else
#define TERMIOS_H_LOCATION <sys/termios.h>
#endif
libpq/fe-exec.c modified so that location of termios.h is determined
by whether HAVE_TERMIOS_H is defined or not, in preparation for switch
to configure
This should clean everything up back to pre-configure status
Remove all the configure related stuff...
Slight oops...Makefile.global generated by configure *wasn't* supposed
to be committed...
Okay, that does it for tonight...
Cleaned out alot of the #define that can be auto-detect'd by
configure from config.h...
Have configure auto-detect more of the libraries, and shorten out
the ports section of Makefile.global as a result.
Add in the code required to auto-detect the OS and to set PORTNAME...
Right now, it only auto-detects FreeBSD, and the rest still get set
to UNDEFINED, but its a step in the right direction, no?
Auto-sense lib{readline,history,curses,termcap}
Autoconf bin/psql/Makefile
Have Makefile autoconf'd according to whether the system has
readline and associated libraries...
Stray endif prevents compile
Slowly building up config.h so that it is pretty much generated by
configure, thereby reducing the ports dependencies in it :)
Makefile.global.in currently should reflect what Makefile.global
in v6.0 does...
Okay, GNUmakefile.in == GNUmakefile, except for the stuff that
configure is auto-finding
Add config.h to configure
Forgot, we don't use Makefile anymore.
Mv Makefile.in to GNUmakefile.in and reflect the change in the
configure script
Bring in Nat' preliminary work at autoconf'ng PostgreSQL...I imagine
it still needs lots of work, but, hey, at least this way there are more
ppl working off the same "Source"...
The last patch fixes some incongruences in the #define used to compile the
Tcl arrays support. Here are the correct values to be defined in config.h
and pgtclCmds.c.
Submitted by: Massimo Dal Zotto <dz@cs.unitn.it>
The second patch adds a more explicative error message to BufferPoolCheckLeak.
It should be completely harmless.
Submitted by: Massimo Dal Zotto <dz@cs.unitn.it>
Document and enable ACLGROUP_PATCH submitted by Massimo
Patch from Massimo Dal Zotto <dz@cs.unitn.it>
The first patch changes the behavior of aclcheck for groups. Currently an user
can access a table only if he has the required permission for ALL the groups
defined for that table. With my patch he can access a table if he has the
permission for ONE of the groups, which seems to me a more useful thing.
If you think this should be the correct behavior of the acl group check feel
free to remove the #ifdef, if not please add a commented line to config.h.
Remove SB_PAD. Compute padding at compile time.
Fix from Andrew for compiling libpgtcl...basically, switch off -Werror,
which probably isn't needed, since we are going to take off -Werror
anyway, right?
Small fixes for SVR4 port by Frank Ridderbusch
Added EXAMPLES section.
Removed unlinked patch that is not neede now that Vadim has properly fixed it.
Add "else true" to make it magically work on Ultrix.
INDEXSCAN_PATCH changes: should work for (Param OP VAR) too
(but I didn't test FUNC OP PARAM and PARAM OP FUNC - no time)
1. INDEXSCAN_PATCH changes: (op Param Var) should work too
2. IndexScanableOperand now uses match_indexkey_operand
instead of equal_indexkey_var (if we have some index on attribute X
then we shouldn't use it for 'where some_func(X) OP CONST').
Fixing bug in INDEXSCAN_PATCH:
ExecInitIndexScan now works with operands of Param type and
(!!!) postquel_execute() now substitutes param values
before calling postquel_start().
Another change from /usr/include/machine/limits.h to just limits.h
Another one that should be including limits.h vs machine/limits.h
Taking a chance here. Under both Solaris and FreeBSD, there is a
/usr/include/limits.h (which quiets the costsize.c warnings)...under
FreeBSD, /usr/include/limits.h *includes* machine/limits.h, while under
Solaris, there is no such things as /usr/include/machine...
Problem with Solaris pointed out by Mark Wahl
the i386_solaris port requires inet_aton.o
Pointed out by Mark Wahl
Cast constants to the type of the other binary operand.
Invalidate vacuum relation cache to use new row counts from vacuum.
Add link options for libpq library search.
Rename GNUMakefile to GNUmakefile so gmake finds it.
Made the patch, but forgot to commit it...
Fixing possible losing data changes:
1. New flag - BM_JUST_DIRTIED - added for BufferDesc;
2. All data "dirtiers" (WriteBuffer and WriteNoReleaseBuffer)
set this flag (and BM_DIRTY too);
3. All data "flushers" (FlushBuffer, BufferSync and BufferReplace)
turn this flag off just before calling smgr[blind]write/smgrflush
and check this flag after flushing buffer: if it turned ON then
BM_DIRTY will stay ON.
New flag for BufferDesc - BM_JUST_DIRTIED, - to prevent
losing data changes.
1. Setting rdesc->rd_tmpunlinked to FALSE in heap_creatr () just after
smgrcreate ().
2. Checking rdesc->rd_tmpunlinked in heap_destroy () & heap_destroyr ()
before calling smgrunlink ().
Use gcc to make Linux ELF shared library, instead of ld.
Change "ld" program back to ld from gcc for Linux. Backend make files need ld.
Must find alternate solution to libpq shared library build problem.
Update handling of backslashes, and pg_user dump.
Add missing #.
Add include files needed for stat().
Rename Makefile to GNUMakefile and add a "use GNU Make" Makefile so people
don't accidentally use non-GNU make.
Set LD = gcc for Linux so shared libpq build works.
Prevent UNIQUE indexes for non-btree access methods.
Remove blank lines.
Removed unneeded custom shell script call.
Add pg_dump's option meanings to the manual page.
Unapply a patch that wasn't a patch...
Update README.support to match what is on the WWW Site
Remove RELEASE.patchlevel...it hasn't been modified since 1.02 *sigh*
Pointed out by Andrew
point the installer at src/test/regress for testing
pointed out by Andrew...
Minor change to file point out by Andrew
This is one of those patches that obviously fixes something, just not
sure exactly what it is it fixes...
Various updates to install, including redirecting installers to
Makefile.custom and pointers at the customize script...
Added pg_dumpall to source tree.
Change EXPLAIN options to just use VERBOSE.
FlushLocalBuffer () releases buffer only if required
by caller.
No more LateWrite, but there is WriteMode;
SetBufferWriteMode () added;
FlushBuffer () fixed: now directly calls smgrflush () and
releases buffer only if required by caller.
No more LateWrite.
VariableRelationPutNextXid () now flushes variable relation
after writing next free XID.
SetBufferWriteMode () added; FlushLocalBuffer () fixed (shouldn't
release buffer if called from WriteNoReleaseBuffer ())
Add inet_aton() prototype.
textin fixed: no more zero-byte (thanks, Erich)
This is the main PostgreSQL git repository.
RSS
Atom