I'm running verify on this now. Are you still getting those verify errors in the SSL suit Adam? If so could you check if that is a regression from 7.6.66?
On Fri, 26 May 2006 06:30:01 +0000 (UTC) "Peter Bortas @ Pike developers forum" 10353@lyskom.lysator.liu.se wrote:
I'm running verify on this now. Are you still getting those verify errors in the SSL suit Adam? If so could you check if that is a regression from 7.6.66?
I can't get that far: Making Gmp Compiling /home/adam/Pike-v7.6.78/src/modules/Gmp/mpz_glue.c Compiling /home/adam/Pike-v7.6.78/src/modules/Gmp/next_prime.c Compiling /home/adam/Pike-v7.6.78/src/modules/Gmp/my_mpz_xor.c /home/adam/Pike-v7.6.78/build/openbsd-3.9-i386/precompile.sh --cache precompile.pike >"/home/adam/Pike-v7.6.78/src/modules/Gmp/mpq.c" "/home/adam/Pike-v7.6.78/src/modules/Gmp/mpq.cmod" || { rm "/home/adam/Pike-v7.6.78/src/modules/Gmp/mpq.c"; exit 1; } precompile: /home/adam/Pike-v7.6.78/build/openbsd-3.9-i386/pike -DNOT_INSTALLED -DPRECOMPILED_SEARCH_MORE -m/home/adam/Pike-v7.6.78/build/openbsd-3.9-i386/master.pike /home/adam/Pike-v7.6.78/bin/precompile.pike /home/adam/Pike-v7.6.78/src/modules/Gmp/mpq.cmod (method=Q) /home/adam/Pike-v7.6.78/build/openbsd-3.9-i386/precompile.sh[172]: /home/adam/Pike-v7.6.78/build/openbsd-3.9-i386/pike: not found precompile: /home/adam/Pike-v7.6.78/build/openbsd-3.9-i386/test-pike /home/adam/Pike-v7.6.78/bin/precompile.pike /home/adam/Pike-v7.6.78/src/modules/Gmp/mpq.cmod (method=QQ) /home/adam/Pike-v7.6.78/build/openbsd-3.9-i386/precompile.sh[172]: /home/adam/Pike-v7.6.78/build/openbsd-3.9-i386/test-pike: not found precompile: pike /home/adam/Pike-v7.6.78/bin/precompile.pike /home/adam/Pike-v7.6.78/src/modules/Gmp/mpq.cmod (method=QQQQQ) /home/adam/Pike-v7.6.78/build/openbsd-3.9-i386/precompile.sh[172]: pike: not found precompile: /home/adam/Pike-v7.6.78/build/openbsd-3.9-i386/pike -DNOT_INSTALLED -DPRECOMPILED_SEARCH_MORE -m/home/adam/Pike-v7.6.78/build/openbsd-3.9-i386/master.pike /home/adam/Pike-v7.6.78/bin/precompile.pike /home/adam/Pike-v7.6.78/src/modules/Gmp/mpq.cmod (method=Q) /home/adam/Pike-v7.6.78/build/openbsd-3.9-i386/precompile.sh[172]: /home/adam/Pike-v7.6.78/build/openbsd-3.9-i386/pike: not found precompile: /home/adam/Pike-v7.6.78/build/openbsd-3.9-i386/test-pike /home/adam/Pike-v7.6.78/bin/precompile.pike /home/adam/Pike-v7.6.78/src/modules/Gmp/mpq.cmod (method=QQ) /home/adam/Pike-v7.6.78/build/openbsd-3.9-i386/precompile.sh[172]: /home/adam/Pike-v7.6.78/build/openbsd-3.9-i386/test-pike: not found precompile: pike /home/adam/Pike-v7.6.78/bin/precompile.pike /home/adam/Pike-v7.6.78/src/modules/Gmp/mpq.cmod (method=QQQQQ) /home/adam/Pike-v7.6.78/build/openbsd-3.9-i386/precompile.sh[172]: pike: not found Failed to find a Pike executable that could run /home/adam/Pike-v7.6.78/bin/precompile.pike. *** Error code 1
What are the datestamps on your mpq.c and mpq.cmod? The ones I got from the archive had the following datestamps:
-rw-r--r-- 1 marcus marcus 52586 Sep 10 2004 mpq.c -rw-r--r-- 1 marcus marcus 20790 Mar 23 2004 mpq.cmod
So regeneration of mpq.c should not be performed.
On Sun, 28 May 2006 20:15:01 +0000 (UTC) "Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) @ Pike (-) developers forum" 10353@lyskom.lysator.liu.se wrote:
What are the datestamps on your mpq.c and mpq.cmod? The ones I got from the archive had the following datestamps:
-rw-r--r-- 1 marcus marcus 52586 Sep 10 2004 mpq.c -rw-r--r-- 1 marcus marcus 20790 Mar 23 2004 mpq.cmod
So regeneration of mpq.c should not be performed.
Yeah, that's what I end up with when I first extract it too. Then for some reason when I run make, it bombs out like that, and if I look now I don't have a mpq.c at all.
Adam
On Sun, 28 May 2006 21:15:02 +0000 (UTC) "Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) @ Pike (-) developers forum" 10353@lyskom.lysator.liu.se wrote:
Sounds like a broken make then. Try GNU make?
Or a broken Makefile? Yeah, it works with gmake. It broke sometime between 7.6.50 which works fine, and 7.6.66 which doesn't.
Adam
As far as I can see, the makefile entry for building .c from .cmod has not changed for three years. But please do investigate; even if it is a bug in make, it would be nice if there is a sane workaround for it.
On Mon, 29 May 2006 09:35:01 +0000 (UTC) "Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) @ Pike (-) developers forum" 10353@lyskom.lysator.liu.se wrote:
As far as I can see, the makefile entry for building .c from .cmod has not changed for three years. But please do investigate; even if it is a bug in make, it would be nice if there is a sane workaround for it.
I'm not sure what is making make decide to rebuild the .c files. If I touch mpf.c and mpq.c before running make, then Gmp builds fine. But then it errors out with the same problem in Kerberos.
Looking at file modification times between 7.6.50 and 7.6.66, I still don't see what change(s) could have made this start happening.
Adam
Maybe there is some trace flag you can turn on to have make tell you its decision tree? If it's BSD make, I believe it has a -d flag.
On Tue, 30 May 2006 09:15:01 +0000 (UTC) "Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) @ Pike (-) developers forum" 10353@lyskom.lysator.liu.se wrote:
Maybe there is some trace flag you can turn on to have make tell you its decision tree? If it's BSD make, I believe it has a -d flag.
Yep, thanks. It seems to be because bin/precompile.pike is newer than mpq.c and mpf.c now.
Adam
Ah, right. Then it's neither a make or a Makefile¹ bug, but a packaging bug.
¹ Well, it might be a bug in "make export".
Ah, right. Then it's neither a make or a Makefile¹ bug, but a packaging bug.
¹ Well, it might be a bug in "make export".
On 5/30/06, Adam Montague amontague@siriushosting.com wrote:
On Tue, 30 May 2006 09:15:01 +0000 (UTC) "Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) @ Pike (-) developers forum" 10353@lyskom.lysator.liu.se wrote:
Maybe there is some trace flag you can turn on to have make tell you its decision tree? If it's BSD make, I believe it has a -d flag.
Yep, thanks. It seems to be because bin/precompile.pike is newer than mpq.c and mpf.c now.
How very odd. I suspect something is funny with the NFS on the machine it was packaged on, but I can't reproduce it sisnce it's down for reinstallation right now.
On Fri, 26 May 2006 06:30:01 +0000 (UTC) "Peter Bortas @ Pike developers forum" 10353@lyskom.lysator.liu.se wrote:
I'm running verify on this now. Are you still getting those verify errors in the SSL suit Adam? If so could you check if that is a regression from 7.6.66?
No, the SSL verify problems are gone.
Adam
pike-devel@lists.lysator.liu.se
-
Adam Montague -
Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) ＠ Pike (-) developers forum -
Peter Bortas -
Peter Bortas ＠ Pike developers forum