You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@spamassassin.apache.org by bu...@bugzilla.spamassassin.org on 2005/05/31 23:17:01 UTC

[Bug 4376] New: SA (v303/snap) 'make' fails (Bus Error) @ "spamc" on OSX 10.4.1 + WORKAROUND

http://bugzilla.spamassassin.org/show_bug.cgi?id=4376

           Summary: SA (v303/snap) 'make' fails (Bus Error) @ "spamc" on OSX
                    10.4.1 + WORKAROUND
           Product: Spamassassin
           Version: 3.0.3
          Platform: Macintosh
        OS/Version: other
            Status: NEW
          Severity: normal
          Priority: P5
         Component: Building & Packaging
        AssignedTo: dev@spamassassin.apache.org
        ReportedBy: openmacnews@speakeasy.net


hi all,

i'm building SpamAssassin (v3.0. & latest snap) on OSX 10.4.1 w/ gcc4.

short story:

	(a) Net::DNS prereq FAILS to build; need DEV version
	(b) config of SA 'complains' about Net::DNS version, and "Net::Ident::_export_hooks()"
	(c) 'make' fails:
		make: *** [spamc/Makefile] Bus error
	(d) workaroud is to manually configure in spamc/

long story -- THE DETAILS ...

1st the prereq 'issue' -- Net::DNS ...

CPAN's Net::DNS fails to build correctly on Tiger:

	cref: http://www.mail-archive.com/macosx@perl.org/msg05790.html

even (4 me) with the proposed fix @:

	http://www.net-dns.org/blog/post/1/15

so i built/installed the latest developer build:

	cref: http://search.cpan.org/%7Eolaf/Net-DNS-0.49_02/
	http://search.cpan.org/CPAN/authors/id/O/OL/OLAF/Net-DNS-0.49_02.tar.gz

with this prereq in place, after successful build/install of razor-agents2.67, i move on to 
SpamAssassin.

with EITHER:

   wget http://apache.oregonstate.edu/spamassassin/Mail-SpamAssassin-3.0.3.tar.gz
   wget http://cvs.apache.org/snapshots/spamassassin/spamassassin_20050531043757.tar.gz

configuring as:

   unsetenv CFLAGS CPPFLAGS CXX CXXFLAGS LDFLAGS LDDLFLAGS LD_PREBIND EXTRA_LDFLAGS 
EXTRA_LIBS LC_ALL LANG LINGUAS
   setenv LANG "en_US"
   setenv LC_ALL "C"
   setenv CFLAGS "-DSPAMC_SSL"

   perl Makefile.PL \
   PREFIX=/usr/local/spamassassin \
   DATADIR=/var/Data/SpamAssassin \
   CONFDIR=/var/Settings/SpamAssassin \
   ENABLE_SSL="yes" \
   LDDLFLAGS="-L/usr/local/ssl/lib -lssl -lcrypto" \
   LDFLAGS="-L/usr/local/ssl/lib -lssl -lcrypto -ldl"

           ########
results in WARNINGS of:
           ########

case (Mail-Spamassin-3.0.3):
	...
	Run Bayes SQL storage tests during 'make test' (additional
	information required)? (y/n) [n]
	
	Argument "0.49_02" isn't numeric in numeric lt (<) at Makefile.PL line 429.
	Checking if your kit is complete...
	Looks good
	Writing Makefile for Mail::SpamAssassin
	Makefile written by ExtUtils::MakeMaker 6.30

case (spamassassin_20050531043757):
	...
	checking module dependencies and their versions...
	Argument "0.49_02" isn't numeric in subroutine entry at (eval 13) line 1.
	Net::Ident::_export_hooks() called too early to check prototype at /usr/local/perl_libs/sitelib/Net/
Ident.pm line 29.
	Checking if your kit is complete...
	Looks good
	Writing Makefile for Mail::SpamAssassin
	Makefile written by ExtUtils::MakeMaker 6.30

#########
NOTE also the "Net::Ident""export_hooks()" for the snapshot case ... it never checks for
#########
the usual:

	"Check network rules during 'make test' (test scripts may fail due to
	network problems)? (y/n) [n]"


a subsequent 'make' FAILS in both cases

		to avoid:
			/usr/local/perl5/bin/perl spamc/configure.pl --prefix="/usr/local/spamassassin" --
sysconfdir="/var/Settings/SpamAssassin" --datadir="/var/Data/SpamAssassin" --enable-ssl="yes"
			cd spamc
			/usr/local/perl5/bin/perl version.h.pl
			version.h.pl: creating version.h
			make: *** [spamc/Makefile] Bus error

after a bit of poking around, it seems there's no "version.h" or "Makefile" getting built in spamc/, so ...
to workaroud this error, b4 invoking 'make' at the top of the tree:

   cd spamc
   perl version.h.pl
   ./configure --enable-ssl

then, simply:

	cd ../
	make

'make' _now_ completes successfully.

'make test' results in:

	5 tests skipped.
	Failed 1/68 test scripts, 98.53% okay. 1/1492 subtests failed, 99.93% okay.
	make: *** [test_dynamic] Error 2
	
	t/razor2....................sh: line 1: razor-report: command not found
	'razor-report < data/spam/001' failed. This may cause this test to fail.
	Argument "0.49_02" isn't numeric in numeric lt (<) at ../blib/lib/Mail/SpamAssassin/Dns.pm line 
1229.
			Not found: spam =  Listed in Razor2
												
						  t/    razor2....................NOK 1# Failed test 1 in t/SATest.pm at line 561
	Argument "0.49_02" isn't numeric in numeric lt (<) at ../blib/lib/Mail/SpamAssassin/Dns.pm line 
1229.

tied, apparently, to the Net::DNS version issue, above; but it doesn't look fatal ...

a subsequent 'make install' completes as well, resulting in:

cd /usr/local/spamassassin/bin
ls -al
	total 188
	drwxr-xr-x  6 root admin   204 May 31 13:40 .
	drwxr-xr-x  3 root admin   102 May 31 12:45 ..
	-r-xr-xr-x  1 root admin 37129 May 31 11:21 sa-learn
	-r-xr-xr-x  1 root admin 29753 May 31 11:22 spamassassin
	-r-xr-xr-x  1 root admin 44800 May 31 13:32 spamc
	-r-xr-xr-x  1 root admin 73346 May 31 11:22 spamd

otool -L spamc
	spamc:
	   /usr/local/ssl/lib/libssl.0.9.7.dylib (compatibility version 0.9.0, current version 0.9.7)
	   /usr/local/ssl/lib/libcrypto.0.9.7.dylib (compatibility version 0.9.0, current version 0.9.7)
	   /usr/lib/libSystem.B.dylib (compatibility version 1.0.0, current version 88.0.0)
	   /usr/lib/libmx.A.dylib (compatibility version 1.0.0, current version 92.0.0)


cheers,

richard



------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug 4376] SA (v303/snap) 'make' fails (Bus Error) @ "spamc" on OSX 10.4.1 + WORKAROUND

Posted by bu...@bugzilla.spamassassin.org.
http://bugzilla.spamassassin.org/show_bug.cgi?id=4376





------- Additional Comments From sidney@sidney.com  2005-05-31 14:47 -------
FWIW, 0.49_02 is an SVN version, the release candidate 2 for 0.50, not a
released version. Do they have to play by the rules for those? :-)

On the other hand, 0.50 hasn't been released yet and I don't think 0.49 runs
under Tiger. Perhaps we do want to tweak the code to accept such version numbers.

The Net::Ident::_export_hooks() warning is not MacOS specific. I rememebr seeing
it but I don't remember the circumstances. My Cygin and Windows machines don't
have Net::Ident, which leaves my Fedora Core 3 as the remaiing possibility. But
I don't recall what made the warning go away, assuming it did.




------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug 4376] SA (v303/snap) 'make' fails (Bus Error) @ "spamc" on OSX 10.4.1 + WORKAROUND

Posted by bu...@bugzilla.spamassassin.org.
http://bugzilla.spamassassin.org/show_bug.cgi?id=4376





------- Additional Comments From jm@jmason.org  2005-06-01 10:07 -------
yeah, I meant SA's trunk -- sounds like it's still cropping up on that platform,
then :(



------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug 4376] SA (v303/snap) 'make' fails (Bus Error) @ "spamc" on OSX 10.4.1 + WORKAROUND

Posted by bu...@bugzilla.spamassassin.org.
http://bugzilla.spamassassin.org/show_bug.cgi?id=4376





------- Additional Comments From openmacnews@speakeasy.net  2005-06-02 21:53 -------
hi,

i built up an OSX 10.3.9 box -- w/ an /usr/local tree identically configured to my 10.4.1 box, except, 
of course, built with gcc33 for OSX 10.3.9, vs. gcc4 for OSX 10.4.1.  in particular, perl is same ver & 
config.

short story -- on OSX 10.3.9, there's no problems with the 'version.h ...' BusError.

build/install/launch are all OK ...

however on 10.4.1 -- actually, now on 2 separate boxes -- the BusError continues to occur.

<confused> richard </confused>




------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug 4376] SA (v303/snap) 'make' fails (Bus Error) @ "spamc" on OSX 10.4.1 + WORKAROUND

Posted by bu...@bugzilla.spamassassin.org.
http://bugzilla.spamassassin.org/show_bug.cgi?id=4376





------- Additional Comments From schulz@adi.com  2005-06-03 06:36 -------
> built with gcc33 for OSX 10.3.9, vs. gcc4 for OSX 10.4.1.  in particular,
> perl is same ver & config.

Am I correct in understanding that on the OSX 10.3.9 box you have a Perl that
was built with gcc33 and on the OSX 10.4.1 box you have a Perl that was built
with gcc4?  If so, there could be a bug in gcc4 that shows up when building
this version of Perl, or a bug in Perl that does not show up when built with
gcc33.  You might try installing the Perl from the OSX 10.3.9 box on to the
OSX 10.4.1 box and see how that works.  If it works, that would eliminate the
operating system as a source of the problem.  You might then try building with
gcc4, but without optimization. (But note that the gcc33 version with
optimization would be preferable to the gcc4 version without.)



------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug 4376] SA (v303/snap) 'make' fails (Bus Error) @ "spamc" on OSX 10.4.1 + WORKAROUND

Posted by bu...@bugzilla.spamassassin.org.
http://bugzilla.spamassassin.org/show_bug.cgi?id=4376





------- Additional Comments From lwilton@earthlink.net  2005-06-01 23:18 -------
Subject: Re:  SA (v303/snap) 'make' fails (Bus Error) @ "spamc" on OSX 10.4.1 + WORKAROUND

> now, at the moment, i'm (very) hard-pressed to see the issue here ...
>
> anything obvious to y'all?

> Exception:  EXC_BAD_ACCESS (0x0001)
> Codes:      KERN_PROTECTION_FAILURE (0x0002) at 0x00000000

I can't read PPC state dumps without digging out a manual, but this sounds
like a null pointer dereference to me.  The question is, what the heck was
Perl doing when it was dereferencing the pointer, and what pointer was it?





------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug 4376] SA (v303/snap) 'make' fails (Bus Error) @ "spamc" on OSX 10.4.1 + WORKAROUND

Posted by bu...@bugzilla.spamassassin.org.
http://bugzilla.spamassassin.org/show_bug.cgi?id=4376





------- Additional Comments From schulz@adi.com  2005-06-03 12:32 -------
> hmmm .... prereq version nightmares ahead ....

> given that all my builds are (typically) dynamic, lib/version dependencies
> worry me .... must think here.

For most operating systems, programs built on any given version of the operating
system will run on any later version.  The OS shared libraried are versioned in
such a way as to make this work.  As for any libraries built as part of Perl,
you would have to install the whole package, not just the Perl executable
itself.



------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug 4376] SA (v303/snap) 'make' fails (Bus Error) @ "spamc" on OSX 10.4.1 + WORKAROUND

Posted by bu...@bugzilla.spamassassin.org.
http://bugzilla.spamassassin.org/show_bug.cgi?id=4376





------- Additional Comments From openmacnews@speakeasy.net  2005-06-01 11:11 -------
hi,
more detail on the BusError ...

1st, i verified the BusError (as above) on two other boxes, as well as the workaroud.

then, some deeper digging.  upon 'make' w/o the spamc "manual" config, i get:

	...
	cd spamc
	/usr/local/perl5/bin/perl version.h.pl
	version.h.pl: creating version.h
	make: *** [spamc/Makefile] Bus error
	%

console log shows:

	Jun  1 11:02:04 devbox crashdump[19715]: perl crashed
	Jun  1 11:02:05 devbox crashdump[19715]: crash report written to: /Library/Logs/CrashReporter/
perl.crash.log


well, it looks like perl.

and,

	% cat /Library/Logs/CrashReporter/perl.crash.log

shows:

	**********
	
	Host Name:      devbox
	Date/Time:      2005-06-01 11:02:04.717 -0700
	OS Version:     10.4.1 (Build 8B15)
	Report Version: 3
	
	Command: perl
	Path:    /usr/local/perl5/bin/perl
	Parent:  make [19543]
	
	Version: ??? (???)
	
	PID:    19714
	Thread: 0
	
	Exception:  EXC_BAD_ACCESS (0x0001)
	Codes:      KERN_PROTECTION_FAILURE (0x0002) at 0x00000000
	
	Thread 0 Crashed:
	0   libSystem.B.dylib        	0x90124960 bcmp + 224
	1   com.apple.CoreFoundation 	0x9076016c __CFInitialize + 420
	2   dyld                     	0x8fe15afc ImageLoaderMachO::doInitialization(ImageLoader::LinkContext 
const&) + 676
	3   dyld                     	0x8fe0b5b0 ImageLoader::recursiveInitialization(ImageLoader::LinkContext 
const&) + 220
	4   dyld                     	0x8fe0b538 ImageLoader::recursiveInitialization(ImageLoader::LinkContext 
const&) + 100
	5   dyld                     	0x8fe0b538 ImageLoader::recursiveInitialization(ImageLoader::LinkContext 
const&) + 100
	6   dyld                     	0x8fe0b538 ImageLoader::recursiveInitialization(ImageLoader::LinkContext 
const&) + 100
	7   dyld                     	0x8fe0ed38 ImageLoader::link(ImageLoader::LinkContext const&, 
ImageLoader::BindingLaziness, ImageLoader::InitializerRunning, unsigned) + 848
	8   dyld                     	0x8fe0443c dyld::link(ImageLoader*, ImageLoader::BindingLaziness, 
ImageLoader::InitializerRunning) + 396
	9   dyld                     	0x8fe08ba8 NSLinkModule + 232
	10  libSystem.B.dylib        	0x9002dbb8 NSLinkModule + 104
	11  perl                     	0x00003314 XS_DynaLoader_dl_load_file + 472 (crt.c:355)
	12  libperl.dylib            	0x00276fb0 Perl_pp_entersub + 1644 (icplusplus.c:28)
	13  libperl.dylib            	0x0026e780 Perl_runops_standard + 68 (icplusplus.c:28)
	14  libperl.dylib            	0x002076b8 Perl_call_sv + 1348 (icplusplus.c:28)
	15  libperl.dylib            	0x00207ad8 Perl_call_list + 452 (icplusplus.c:28)
	16  libperl.dylib            	0x00240380 Perl_newATTRSUB + 3768 (icplusplus.c:28)
	17  libperl.dylib            	0x00240d18 Perl_utilize + 1660 (icplusplus.c:28)
	18  libperl.dylib            	0x0022ed68 Perl_yyparse + 4284 (icplusplus.c:28)
	19  libperl.dylib            	0x002a97b4 S_doeval + 632 (icplusplus.c:28)
	20  libperl.dylib            	0x002b1810 Perl_pp_require + 4656 (icplusplus.c:28)
	21  libperl.dylib            	0x0026e780 Perl_runops_standard + 68 (icplusplus.c:28)
	22  libperl.dylib            	0x002076b8 Perl_call_sv + 1348 (icplusplus.c:28)
	23  libperl.dylib            	0x00207ad8 Perl_call_list + 452 (icplusplus.c:28)
	24  libperl.dylib            	0x00240380 Perl_newATTRSUB + 3768 (icplusplus.c:28)
	25  libperl.dylib            	0x00240d18 Perl_utilize + 1660 (icplusplus.c:28)
	26  libperl.dylib            	0x0022ed68 Perl_yyparse + 4284 (icplusplus.c:28)
	27  libperl.dylib            	0x002a97b4 S_doeval + 632 (icplusplus.c:28)
	28  libperl.dylib            	0x002b1810 Perl_pp_require + 4656 (icplusplus.c:28)
	29  libperl.dylib            	0x0026e780 Perl_runops_standard + 68 (icplusplus.c:28)
	30  libperl.dylib            	0x002076b8 Perl_call_sv + 1348 (icplusplus.c:28)
	31  libperl.dylib            	0x00207ad8 Perl_call_list + 452 (icplusplus.c:28)
	32  libperl.dylib            	0x00240380 Perl_newATTRSUB + 3768 (icplusplus.c:28)
	33  libperl.dylib            	0x00240d18 Perl_utilize + 1660 (icplusplus.c:28)
	34  libperl.dylib            	0x0022ed68 Perl_yyparse + 4284 (icplusplus.c:28)
	35  libperl.dylib            	0x002a97b4 S_doeval + 632 (icplusplus.c:28)
	36  libperl.dylib            	0x002b1810 Perl_pp_require + 4656 (icplusplus.c:28)
	37  libperl.dylib            	0x0026e780 Perl_runops_standard + 68 (icplusplus.c:28)
	38  libperl.dylib            	0x002091d0 perl_run + 484 (icplusplus.c:28)
	39  perl                     	0x00002810 main + 224 (crt.c:355)
	40  perl                     	0x00001f7c _start + 348 (crt.c:272)
	41  perl                     	0x00001e1c start + 60
	
	Thread 0 crashed with PPC Thread State:
	  srr0: 0x90124960 srr1: 0x0200f030                vrsave: 0x00000000
		cr: 0x44424248  xer: 0x00000003   lr: 0x9076016c  ctr: 0x00000004
		r0: 0x000009a4   r1: 0xbfffabf0   r2: 0x00000006   r3: 0x00000000
		r4: 0x907dc65c   r5: 0x00000000   r6: 0x00000000   r7: 0xfffff66c
		r8: 0x00000004   r9: 0x00000010  r10: 0x90001490  r11: 0xa07282ec
	   r12: 0x90124880  r13: 0x000001e6  r14: 0xc8db7340  r15: 0x000001e6
	   r16: 0xc8dc7891  r17: 0x01106e60  r18: 0x00000000  r19: 0x00000001
	   r20: 0x000001e6  r21: 0xc8dce0fd  r22: 0x00000001  r23: 0x018c2cd0
	   r24: 0x8fe52000  r25: 0x00000001  r26: 0xa071ffd0  r27: 0xbfffd9b0
	   r28: 0xbfffd9b4  r29: 0x00000001  r30: 0x00000005  r31: 0x9075ffd0
	
	Binary Images Description:
		0x1000 -     0x3fff perl 	/usr/local/perl5/bin/perl
		0xf000 -    0x10fff Cwd.bundle 	/usr/local/perl_libs/privlib/darwin-thread-multi-2level/
auto/Cwd/Cwd.bundle
	   0x18000 -    0x18fff Hostname.bundle 	/usr/local/perl_libs/privlib/darwin-thread-
multi-2level/auto/Sys/Hostname/Hostname.bundle
	   0x1b000 -    0x1cfff Base64.bundle 	/usr/local/perl_libs/privlib/darwin-thread-
multi-2level/auto/MIME/Base64/Base64.bundle
	   0x47000 -    0x49fff Fcntl.bundle 	/usr/local/perl_libs/privlib/darwin-thread-multi-2level/
auto/Fcntl/Fcntl.bundle
	   0x4c000 -    0x64fff POSIX.bundle 	/usr/local/perl_libs/privlib/darwin-thread-multi-2level/
auto/POSIX/POSIX.bundle
	   0x71000 -    0x73fff IO.bundle 	/usr/local/perl_libs/privlib/darwin-thread-multi-2level/
auto/IO/IO.bundle
	   0x76000 -    0x7afff Socket.bundle 	/usr/local/perl_libs/privlib/darwin-thread-multi-2level/
auto/Socket/Socket.bundle
	   0x7d000 -    0x7efff DNS.bundle 	/usr/local/perl_libs/sitelib/darwin-thread-multi-2level/
auto/Net/DNS/DNS.bundle
	   0x81000 -    0x86fff Socket6.bundle 	/usr/local/perl_libs/sitelib/darwin-thread-
multi-2level/auto/Socket6/Socket6.bundle
	   0x89000 -    0x8cfff SHA1.bundle 	/usr/local/perl_libs/sitelib/darwin-thread-multi-2level/
auto/Digest/SHA1/SHA1.bundle
	   0x8f000 -    0x9afff DB_File.bundle 	/usr/local/perl_libs/privlib/darwin-thread-multi-2level/
auto/DB_File/DB_File.bundle
	  0x205000 -   0x30dfff libperl.dylib 	/usr/local/perl_libs/privlib/darwin-thread-multi-2level/
CORE/libperl.dylib
	  0x505000 -   0x5b3fff libdb-4.3.dylib 	/usr/local/berkeley-db/lib/libdb-4.3.dylib
	 0xa000000 -  0xa09bfff com.tcltk.tcllibrary 8.5a3	/Library/Frameworks/Tcl.framework/
Versions/8.5/Tcl
	0x8fe00000 - 0x8fe50fff dyld 43	/usr/lib/dyld
	0x90000000 - 0x901a6fff libSystem.B.dylib 	/usr/lib/libSystem.B.dylib
	0x901fe000 - 0x90202fff libmathCommon.A.dylib 	/usr/lib/system/libmathCommon.A.dylib
	0x90727000 - 0x90800fff com.apple.CoreFoundation 6.4.1 (368.1)
	/System/Library/Frameworks/CoreFoundation.framework/Versions/A/CoreFoundation
	0x9084b000 - 0x9094dfff libicucore.A.dylib 	/usr/lib/libicucore.A.dylib
	0x909a7000 - 0x90a2bfff libobjc.A.dylib 	/usr/lib/libobjc.A.dylib
	0x90ae3000 - 0x90af5fff libauto.dylib 	/usr/lib/libauto.dylib
	0x91d33000 - 0x91d53fff libmx.A.dylib 	/usr/lib/libmx.A.dylib


now, at the moment, i'm (very) hard-pressed to see the issue here ...

anything obvious to y'all?

cheers,

richard



------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug 4376] SA (v303/snap) 'make' fails (Bus Error) @ "spamc" on OSX 10.4.1 + WORKAROUND

Posted by bu...@bugzilla.spamassassin.org.
http://bugzilla.spamassassin.org/show_bug.cgi?id=4376





------- Additional Comments From jm@jmason.org  2005-05-31 15:05 -------
'The Net::Ident::_export_hooks() warning is not MacOS specific. I rememebr seeing
it but I don't remember the circumstances. My Cygin and Windows machines don't
have Net::Ident, which leaves my Fedora Core 3 as the remaiing possibility. But
I don't recall what made the warning go away, assuming it did.'

I checked in a fix for this on trunk -- it's a one-liner hack in our code ;)



------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug 4376] SA (v303/snap) 'make' fails (Bus Error) @ "spamc" on OSX 10.4.1 + WORKAROUND

Posted by bu...@bugzilla.spamassassin.org.
http://bugzilla.spamassassin.org/show_bug.cgi?id=4376





------- Additional Comments From openmacnews@speakeasy.net  2005-06-03 07:41 -------
>Am I correct in understanding that on the OSX 10.3.9 box you have a Perl that
>was built with gcc33 and on the OSX 10.4.1 box you have a Perl that was built
>with gcc4?

yes, you are correct.

>If so, there could be a bug in gcc4 that shows up when building
>this version of Perl, or a bug in Perl that does not show up when built with
>gcc33.  You might try installing the Perl from the OSX 10.3.9 box on to the
>OSX 10.4.1 box and see how that works.  If it works, that would eliminate the
>operating system as a source of the problem.  You might then try building with
>gcc4, but without optimization. (But note that the gcc33 version with
>optimization would be preferable to the gcc4 version without.)

hmmm .... prereq version nightmares ahead ....

given that all my builds are (typically) dynamic, lib/version dependencies worry me .... must think here.





------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug 4376] SA (v303/snap) 'make' fails (Bus Error) @ "spamc" on OSX 10.4.1 + WORKAROUND

Posted by bu...@bugzilla.spamassassin.org.
http://bugzilla.spamassassin.org/show_bug.cgi?id=4376





------- Additional Comments From lwilton@earthlink.net  2005-06-03 00:18 -------
Subject: Re:  SA (v303/snap) 'make' fails (Bus Error) @ "spamc" on OSX 10.4.1 + WORKAROUND

> however on 10.4.1 -- actually, now on 2 separate boxes -- the BusError
continues to occur.
>
> <confused> richard </confused>

If Perl and the hardware is the same but the os is different, and this
results in a crash, I can say (as an ex OS designer) that this means that
one OS is giving some different response to Perl on some sort of query, and
Perl isn't expecing the response it is getting.  Who'se problem this is
isn't real important yet, the important thing is to locate the difference
that seems to be resulting in perl doing a null pointer dereference.  Once
the situation is understood a fix can be developed.

Is there a way to run perl in some internal debug mode where it will spit
out a lot of information on what it is doing?  It probably isn't running
very long before it crashes, so the debug result shouldn't be absolutely
huge.  Setting up to attach gdb to watch the crash would also be very
informative; but only if you know a lot about the insides of Perl itself.

BTW, it might be worth checking on a Perl site or list and see if there is
any known similar problem with the OS that is causing you grief.  This might
be a known (although possibly not solved) Perl bug.





------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug 4376] SA (v303/snap) 'make' fails (Bus Error) @ "spamc" on OSX 10.4.1 + WORKAROUND

Posted by bu...@bugzilla.spamassassin.org.
http://bugzilla.spamassassin.org/show_bug.cgi?id=4376





------- Additional Comments From openmacnews@speakeasy.net  2005-06-03 07:49 -------
> Who'se problem this is
> isn't real important yet, the important thing is to locate the difference
> that seems to be resulting in perl doing a null pointer dereference.  Once
> the situation is understood a fix can be developed.

agreed

> Is there a way to run perl in some internal debug mode where it will spit
> out a lot of information on what it is doing?  It probably isn't running
> very long before it crashes, so the debug result shouldn't be absolutely
> huge.

"perl -d", i suppose ... but i've ever used it to date, so not sure

> Setting up to attach gdb to watch the crash would also be very
> informative; but only if you know a lot about the insides of Perl itself.

re perl debugging -- i'm in over my head here ...

an occassional script is 1 thig, perl itself is another

happy to help, i'll need some newbie style guidance -- here or offlie, doesn't matter

> BTW, it might be worth checking on a Perl site or list and see if there is
> any known similar problem with the OS that is causing you grief.  This might
> be a known (although possibly not solved) Perl bug.

osx 10.4.1 bigs are a-plenty -- nothing specific, as yet tho ...




------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug 4376] SA (v303/snap) 'make' fails (Bus Error) @ "spamc" on OSX 10.4.1 + WORKAROUND

Posted by bu...@bugzilla.spamassassin.org.
http://bugzilla.spamassassin.org/show_bug.cgi?id=4376





------- Additional Comments From openmacnews@speakeasy.net  2005-06-01 09:47 -------
just fyi:

after making the changes/mods as i'd reported, 'make test' for a svn co'd spamassassin r179341 
results in:

   All tests successful, 10 tests skipped.
   Files=89, Tests=2400, 9527 wallclock secs (2697.29 cusr + 707.62 csys = 3404.91 CPU)

and (so far) is working as expected with my MTA ...

Justin -- a question:

   re: "I checked in a fix for this on trunk" ...

i presume you mean SA's trunk? fwiw, i'm still seeing the Net::Ident::_export_hooks() warning as of this 
rev (r179341).  did i miss/misunderstand something, or is it in a later rev?

cheers,

richard



------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug 4376] SA (v303/snap) 'make' fails (Bus Error) @ "spamc" on OSX 10.4.1 + WORKAROUND

Posted by bu...@bugzilla.spamassassin.org.
http://bugzilla.spamassassin.org/show_bug.cgi?id=4376





------- Additional Comments From openmacnews@speakeasy.net  2005-05-31 14:53 -------
> FWIW: Net::DNS isn't a prereq.  It's an optional module we'll use if available.

fair nuf. =)

it is, though, for razor-agents ... 

...

>>       "Check network rules during 'make test' (test scripts may fail due to
>>       network problems)? (y/n) [n]"

> This is different in 3.1.  No more questions. :)

ok.

>> make: *** [spamc/Makefile] Bus error

> Interesting.  I'm not sure why you'd get a "Bus error" from that.

no clue here either, as yet ...

thx,

richard



------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug 4376] SA (v303/snap) 'make' fails (Bus Error) @ "spamc" on OSX 10.4.1 + WORKAROUND

Posted by bu...@bugzilla.spamassassin.org.
http://bugzilla.spamassassin.org/show_bug.cgi?id=4376





------- Additional Comments From sidney@sidney.com  2005-05-31 15:03 -------
Correction: Net::DNS 0.49_02 has been released on CPAN

http://www.net-dns.org/blog/post/1/18




------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug 4376] SA (v303/snap) 'make' fails (Bus Error) @ "spamc" on OSX 10.4.1 + WORKAROUND

Posted by bu...@bugzilla.spamassassin.org.
http://bugzilla.spamassassin.org/show_bug.cgi?id=4376





------- Additional Comments From felicity@kluge.net  2005-05-31 15:06 -------
Subject: Re:  SA (v303/snap) 'make' fails (Bus Error) @ "spamc" on OSX 10.4.1 + WORKAROUND

On Tue, May 31, 2005 at 03:03:48PM -0700, bugzilla-daemon@bugzilla.spamassassin.org wrote:
> Correction: Net::DNS 0.49_02 has been released on CPAN
> http://www.net-dns.org/blog/post/1/18

"Thanks to Sidney Markovitz for the assessment of the problem and suggestions
for a fix"

You should have also suggested that they not break the convention of
all-numeric version values. ;)





------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug 4376] SA (v303/snap) 'make' fails (Bus Error) @ "spamc" on OSX 10.4.1 + WORKAROUND

Posted by bu...@bugzilla.spamassassin.org.
http://bugzilla.spamassassin.org/show_bug.cgi?id=4376


openmacnews@speakeasy.net changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WONTFIX




------- Additional Comments From openmacnews@speakeasy.net  2005-09-07 19:30 -------
soon to be irrelevant ....

not a prob w/ svn trunk.



------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug 4376] SA (v303/snap) 'make' fails (Bus Error) @ "spamc" on OSX 10.4.1 + WORKAROUND

Posted by bu...@bugzilla.spamassassin.org.
http://bugzilla.spamassassin.org/show_bug.cgi?id=4376





------- Additional Comments From mrw7mrw@yahoo.ca  2005-09-27 13:27 -------

> 	(b) config of SA 'complains' about Net::DNS version, and
"Net::Ident::_export_hooks()"

> NOTE also the "Net::Ident""export_hooks()" for the snapshot case ... it never 


I have a solution for the Net::Ident problem.  I know that this isn't the best
place to put it, but a google search for Net::Ident and export_hooks led me here
and only here.  Hopefully this info will help others as well.

The problem is that the prototype for _export_hooks isn't declared before it is
called.  You simply need to insert "sub _export_hooks();" at line 29 in
Ident.pm, right before _export_hooks is called.   You should do this before
running perl Makefile.PL.

A second problem that you may run into, though, is that the tests for Ident.pm
try to connect to ident servers on a few hosts, some of which aren't running the
ident daemon, or at least won't accept connections from my machine!  :-)  Before
running make test, edit t/hosts to include a list of hosts that you KNOW are
running the ident daemon.  Me, I simply put 127.0.0.1 in there.



------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug 4376] SA (v303/snap) 'make' fails (Bus Error) @ "spamc" on OSX 10.4.1 + WORKAROUND

Posted by bu...@bugzilla.spamassassin.org.
http://bugzilla.spamassassin.org/show_bug.cgi?id=4376





------- Additional Comments From felicity@kluge.net  2005-05-31 14:35 -------
Subject: Re:   New: SA (v303/snap) 'make' fails (Bus Error) @ "spamc" on OSX 10.4.1 + WORKAROUND

On Tue, May 31, 2005 at 02:17:01PM -0700, bugzilla-daemon@bugzilla.spamassassin.org wrote:
> hi all,

Thanks for opening the ticket. :)

> 	(a) Net::DNS prereq FAILS to build; need DEV version

FWIW: Net::DNS isn't a prereq.  It's an optional module we'll use if available.

> 	Argument "0.49_02" isn't numeric in numeric lt (<) at Makefile.PL line 429.

Yeah, it's not numeric...  Perhaps when we're checking for version numbers we
ought to do something like:

  $ver =~ s/^(\d+\.\d+).*$/$1/;

or something?  This is the same issue as if "0.49.1", etc.  The problem really
is with Net::DNS using a non-standard version number, but if it's simple
enough to avoid we may want to add in a kluge.  (if anyone asks, this is
why internally our versions are x.yyyzzz, not x.y.z ...)

> 	Net::Ident::_export_hooks() called too early to check prototype at /usr/local/perl_libs/sitelib/Net/
> Ident.pm line 29.

No idea what that's about, but I don't use Net::Ident...

> NOTE also the "Net::Ident""export_hooks()" for the snapshot case ... it never checks for
> the usual:
> 
> 	"Check network rules during 'make test' (test scripts may fail due to
> 	network problems)? (y/n) [n]"

This is different in 3.1.  No more questions. :)

> 			/usr/local/perl5/bin/perl version.h.pl
> 			version.h.pl: creating version.h
> 			make: *** [spamc/Makefile] Bus error

Interesting.  I'm not sure why you'd get a "Bus error" from that.





------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.