Re: [PATCHES] mingw configure failure detection
OK, fixed. --- Andrew Dunstan wrote: Bruce Momjian wrote: Andrew Dunstan wrote: following up our conversation last night, I have tested the following with autoconf 2.57 and it apparently does the right thing: # Links sometimes fail undetected on Mingw - # so here we detect it and warn the user case $host_os in mingw*) AC_CONFIG_COMMANDS([default],[ for linkspec in : $config_links; do test x$linkspec = x: continue linktry=`echo $linkspec | sed 's,:.*,,'` test -e $linktry || AC_MSG_WARN([ link for $linktry failed - please fix by hand]) done ]) ;; esac OK, got it working nicely: case $host_os in mingw*) AC_CONFIG_COMMANDS([check_win32_symlinks],[ # Links sometimes fail undetected on Mingw - # so here we detect it and warn the user for FILE in $CONFIG_LINKS The quotes in the line above are spurious and need to be removed - they inhibit expansion of the variable into words, which is needed. I have checked that with the quotes removed the macro works correctly. cheers andrew do # test -e works for symlinks in the MinGW console test -e `expr $FILE : '\(^:*\)'` || AC_MSG_WARN([*** link for $FILE - please fix by hand]) done ]) ;; esac Applied. ---(end of broadcast)--- TIP 2: you can get off all lists at once with the unregister command (send unregister YourEmailAddressHere to [EMAIL PROTECTED]) -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 3: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly
Re: [PATCHES] mingw configure failure detection
The patch as applied is totally broken - the tests have to be run by
config.status after it has actually tried to make the links, which is
why I used AC_CONFIG_COMMANDS.
cheers
andrew
Bruce Momjian wrote:
Adjusted attached patch applied. Thanks.
I didn't change 'test -e' because 'test -h' doesn't work (is symlink
test on my platform). test -f actually works on symlinks too in MinGW,
strangely enough. Anyway, the code is only run on MinGW so we don't
have to worry about other platforms. I added a comment explaining it
works.
I got rid of the AC_CONFIG_COMMANDS and just do the for/test loop in the
shell code directly. AC_CONFIG_COMMANDS wouldn't allow me to call
AC_MSG_WARN.
---
Peter Eisentraut wrote:
Andrew Dunstan wrote:
This patch for configure.in detects the link failures in MINGW that I
and others have seen and warns the user they have to fix it up.
I continue to object to these sort of patches if no one wants to bother
to contact the mingw maintainers about this problem. Maybe they have a
better fix available?
Btw. test -e is not portable. And please use AC_MSG_WARN instead of
echo.
---(end of broadcast)---
TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED]
Index: configure
===
RCS file: /cvsroot/pgsql-server/configure,v
retrieving revision 1.358
diff -c -c -r1.358 configure
*** configure 11 May 2004 21:57:13 - 1.358
--- configure 13 May 2004 01:26:20 -
***
*** 17974,17979
--- 17974,17998
ac_config_links=$ac_config_links src/backend/port/dynloader.c:src/backend/port/dynloader/${template}.c src/backend/port/pg_sema.c:${SEMA_IMPLEMENTATION} src/backend/port/pg_shmem.c:${SHMEM_IMPLEMENTATION} src/include/dynloader.h:src/backend/port/dynloader/${template}.h src/include/pg_config_os.h:src/include/port/${template}.h src/Makefile.port:src/makefiles/Makefile.${template}
+ # Links sometimes fail undetected on Mingw -
+ # so here we detect it and warn the user
+ case $host_os in mingw*)
+ for linktarget in \
+ src/backend/port/tas.s \
+ src/backend/port/dynloader.c \
+ src/backend/port/pg_sema.c \
+ src/backend/port/pg_shmem.c \
+ src/include/dynloader.h \
+ src/include/pg_config_os.h \
+ src/Makefile.port
+ do
+ # test -e works for symlinks in the MinGW console
+ test -e $linktarget || { echo $as_me:$LINENO: WARNING: *** link for $linktarget failed - please fix by hand 5
+ echo $as_me: WARNING: *** link for $linktarget failed - please fix by hand 2;}
+ done
+ ;;
+ esac
+
ac_config_headers=$ac_config_headers src/include/pg_config.h
Index: configure.in
===
RCS file: /cvsroot/pgsql-server/configure.in,v
retrieving revision 1.346
diff -c -c -r1.346 configure.in
*** configure.in 11 May 2004 21:57:13 - 1.346
--- configure.in 13 May 2004 01:26:22 -
***
*** 1236,1241
--- 1236,1259
src/Makefile.port:src/makefiles/Makefile.${template}
])
+ # Links sometimes fail undetected on Mingw -
+ # so here we detect it and warn the user
+ case $host_os in mingw*)
+ for linktarget in \
+ src/backend/port/tas.s \
+ src/backend/port/dynloader.c \
+ src/backend/port/pg_sema.c \
+ src/backend/port/pg_shmem.c \
+ src/include/dynloader.h \
+ src/include/pg_config_os.h \
+ src/Makefile.port
+ do
+ # test -e works for symlinks in the MinGW console
+ test -e $linktarget || AC_MSG_WARN([*** link for $linktarget failed - please fix by hand])
+ done
+ ;;
+ esac
+
AC_CONFIG_HEADERS([src/include/pg_config.h],
[
# Update timestamp for pg_config.h (see Makefile.global)
---(end of broadcast)---
TIP 6: Have you searched our list archives?
http://archives.postgresql.org
---(end of broadcast)---
TIP 4: Don't 'kill -9' the postmaster
Re: [PATCHES] mingw configure failure detection
Peter Eisentraut wrote: Andrew Dunstan wrote: The patch as applied is totally broken - the tests have to be run by config.status after it has actually tried to make the links, which is why I used AC_CONFIG_COMMANDS. Additional note: Instead of listing the files explicitly, just use $CONFIG_LINKS, which contains the list of files. Well, yes, except that each entry there is in the form linkname:targetname - we'd have to have extra code to split them up. Something like: for ac_file in : $CONFIG_LINKS; do test x$ac_file = x: continue linktry=`echo $ac_file | sed 's,:.*,,'` test .. done cheers andrew ---(end of broadcast)--- TIP 5: Have you checked our extensive FAQ? http://www.postgresql.org/docs/faqs/FAQ.html
Re: [PATCHES] mingw configure failure detection
Andrew Dunstan wrote:
Peter Eisentraut wrote:
Andrew Dunstan wrote:
The patch as applied is totally broken - the tests have to be run by
config.status after it has actually tried to make the links, which is
why I used AC_CONFIG_COMMANDS.
Additional note: Instead of listing the files explicitly, just use
$CONFIG_LINKS, which contains the list of files.
Well, yes, except that each entry there is in the form
linkname:targetname - we'd have to have extra code to split them up.
Something like:
for ac_file in : $CONFIG_LINKS; do test x$ac_file = x: continue
linktry=`echo $ac_file | sed 's,:.*,,'`
test ..
done
OK, attached patch applied.
--
Bruce Momjian| http://candle.pha.pa.us
[EMAIL PROTECTED] | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup.| Newtown Square, Pennsylvania 19073
Index: configure
===
RCS file: /cvsroot/pgsql-server/configure,v
retrieving revision 1.359
diff -c -c -r1.359 configure
*** configure 13 May 2004 01:44:59 - 1.359
--- configure 13 May 2004 22:56:38 -
***
*** 17974,17998
ac_config_links=$ac_config_links
src/backend/port/dynloader.c:src/backend/port/dynloader/${template}.c
src/backend/port/pg_sema.c:${SEMA_IMPLEMENTATION}
src/backend/port/pg_shmem.c:${SHMEM_IMPLEMENTATION}
src/include/dynloader.h:src/backend/port/dynloader/${template}.h
src/include/pg_config_os.h:src/include/port/${template}.h
src/Makefile.port:src/makefiles/Makefile.${template}
- # Links sometimes fail undetected on Mingw -
- # so here we detect it and warn the user
- case $host_os in mingw*)
- for linktarget in \
- src/backend/port/tas.s \
- src/backend/port/dynloader.c \
- src/backend/port/pg_sema.c \
- src/backend/port/pg_shmem.c \
- src/include/dynloader.h \
- src/include/pg_config_os.h \
- src/Makefile.port
- do
- # test -e works for symlinks in the MinGW console
- test -e $linktarget || { echo $as_me:$LINENO: WARNING: *** link for
$linktarget failed - please fix by hand 5
- echo $as_me: WARNING: *** link for $linktarget failed - please fix by hand 2;}
- done
- ;;
- esac
-
ac_config_headers=$ac_config_headers src/include/pg_config.h
--- 17974,17979
***
*** 19200,19203
--- 19181,19197
# would make configure fail if this is the last instruction.
$ac_cs_success || { (exit 1); exit 1; }
fi
+
+
+ # Links sometimes fail undetected on Mingw -
+ # so here we detect it and warn the user
+ case $host_os in mingw*)
+ for FILE in $CONFIG_LINKS
+ do
+ # test -e works for symlinks in the MinGW console
+ test -e `expr $FILE : '\(^:*\)'` || { echo $as_me:$LINENO: WARNING: ***
link for $FILE - please fix by hand 5
+ echo $as_me: WARNING: *** link for $FILE - please fix by hand 2;}
+ done
+ ;;
+ esac
Index: configure.in
===
RCS file: /cvsroot/pgsql-server/configure.in,v
retrieving revision 1.347
diff -c -c -r1.347 configure.in
*** configure.in13 May 2004 01:45:02 - 1.347
--- configure.in13 May 2004 22:56:39 -
***
*** 1236,1263
src/Makefile.port:src/makefiles/Makefile.${template}
])
# Links sometimes fail undetected on Mingw -
# so here we detect it and warn the user
case $host_os in mingw*)
! for linktarget in \
! src/backend/port/tas.s \
! src/backend/port/dynloader.c \
! src/backend/port/pg_sema.c \
! src/backend/port/pg_shmem.c \
! src/include/dynloader.h \
! src/include/pg_config_os.h \
! src/Makefile.port
do
# test -e works for symlinks in the MinGW console
! test -e $linktarget || AC_MSG_WARN([*** link for $linktarget failed - please
fix by hand])
done
;;
esac
- AC_CONFIG_HEADERS([src/include/pg_config.h],
- [
- # Update timestamp for pg_config.h (see Makefile.global)
- echo src/include/stamp-h
- ])
-
- AC_OUTPUT
--- 1236,1257
src/Makefile.port:src/makefiles/Makefile.${template}
])
+ AC_CONFIG_HEADERS([src/include/pg_config.h],
+ [
+ # Update timestamp for pg_config.h (see Makefile.global)
+ echo src/include/stamp-h
+ ])
+
+ AC_OUTPUT
+
# Links sometimes fail undetected on Mingw -
# so here we detect it and warn the user
case $host_os in mingw*)
! for FILE in $CONFIG_LINKS
do
# test -e works for symlinks in the MinGW console
! test -e `expr $FILE : '\([^:]*\)'` || AC_MSG_WARN([*** link for $FILE -
please fix by hand])
done
;;
esac
---(end of broadcast)---
TIP 3: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to [EMAIL PROTECTED] so that your
message can get
Re: [PATCHES] mingw configure failure detection
Am Montag, 10. Mai 2004 09:17 schrieb Andrew Dunstan: It appears that Dann Corbit has filed a MinGW bug on this, with accurate problem description, but he received a less than stellar response. See https://sourceforge.net/tracker/index.php? func=detailaid=931263group_id=2435atid=102435 I added a followup to this bug and he requests a minimal reproduceable example. Can someone provide one? ---(end of broadcast)--- TIP 8: explain analyze is your friend
Re: [PATCHES] mingw configure failure detection
Peter Eisentraut [EMAIL PROTECTED] writes: Am Montag, 10. Mai 2004 09:17 schrieb Andrew Dunstan: It appears that Dann Corbit has filed a MinGW bug on this, with accurate problem description, but he received a less than stellar response. See https://sourceforge.net/tracker/index.php? func=detailaid=931263group_id=2435atid=102435 I added a followup to this bug and he requests a minimal reproduceable example. Can someone provide one? From what we've been hearing, all you have to do is download current Postgres (snapshot tarball or CVS pull) and try configure a few times (with make distclean to clean up). That's perhaps not minimal but it should at least convince him the bug exists ... regards, tom lane ---(end of broadcast)--- TIP 9: the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match
Re: [PATCHES] mingw configure failure detection
Tom Lane wrote: Peter Eisentraut [EMAIL PROTECTED] writes: Am Montag, 10. Mai 2004 09:17 schrieb Andrew Dunstan: It appears that Dann Corbit has filed a MinGW bug on this, with accurate problem description, but he received a less than stellar response. See https://sourceforge.net/tracker/index.php? func=detailaid=931263group_id=2435atid=102435 I added a followup to this bug and he requests a minimal reproduceable example. Can someone provide one? From what we've been hearing, all you have to do is download current Postgres (snapshot tarball or CVS pull) and try configure a few times (with make distclean to clean up). That's perhaps not minimal but it should at least convince him the bug exists ... I just installed the configure symlink test code that throws a warning if the symlinks don't work, so it is even easier to test now. Just run configure until it fails --- nothing to compile. -- Bruce Momjian| http://candle.pha.pa.us [EMAIL PROTECTED] | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup.| Newtown Square, Pennsylvania 19073 ---(end of broadcast)--- TIP 8: explain analyze is your friend
Re: [PATCHES] mingw configure failure detection
I said: Peter Eisentraut said: Andrew Dunstan wrote: This patch for configure.in detects the link failures in MINGW that I and others have seen and warns the user they have to fix it up. I continue to object to these sort of patches if no one wants to bother to contact the mingw maintainers about this problem. Maybe they have a better fix available? I have already said I would follow it up. But the problem exists today and shouldn't be left until we get an answer. Maybe it will take a week or a month, or there will be no good answer at all. It appears that Dann Corbit has filed a MinGW bug on this, with accurate problem description, but he received a less than stellar response. See https://sourceforge.net/tracker/index.php? func=detailaid=931263group_id=2435atid=102435 (Despite previous speculation, he did indicate that the problem is intermittent, not constant). Cheers andrew ---(end of broadcast)--- TIP 8: explain analyze is your friend
Re: [PATCHES] mingw configure failure detection
Peter Eisentraut said: Andrew Dunstan wrote: This patch for configure.in detects the link failures in MINGW that I and others have seen and warns the user they have to fix it up. I continue to object to these sort of patches if no one wants to bother to contact the mingw maintainers about this problem. Maybe they have a better fix available? I have already said I would follow it up. But the problem exists today and shouldn't be left until we get an answer. Maybe it will take a week or a month, or there will be no good answer at all. All this does is issue a warning. This problem has happened to a number of users. So far the reponse has largely been 'fix it by hand and move on'. I think our reputation is and ought to be better than that. Btw. test -e is not portable. Since this would only be run under Mingw, where it does work, why does that matter? And please use AC_MSG_WARN instead of echo. I guess Bruce can make this change. cheers andrew ---(end of broadcast)--- TIP 7: don't forget to increase your free space map settings
