It has been spotem some time ago here: http://lists.freebsd.org/pipermail/freebsd-questions/2009-April/196808.html isoqlog just drops core after reading config file under amd64 with segmentation fault How-To-Repeat: # isoqlog Year: 2009 Month: 10 outputdir:/usr/local/share/sqstat/exim htmldir:/usr/local/share/isoqlog/htmltemp logtype:exim logstore:/var/log/exim/mainlog langfile:/usr/local/share/isoqlog/lang/polish maxsender:1000 maxreceiver:1000 maxtotal:1000 maxbyte:1000 hostname: mainframe.kkip.pl Domains <hidden> Domains <hidden> Domains <hidden> Segmentation fault (core dumped)
Maintainer of mail/isoqlog, Please note that PR ports/139867 has just been submitted. If it contains a patch for an upgrade, an enhancement or a bug fix you agree on, reply to this email stating that you approve the patch and a committer will take care of it. The full text of the PR can be found at: http://www.freebsd.org/cgi/query-pr.cgi?pr=ports/139867 -- Edwin Groothuis via the GNATS Auto Assign Tool edwin@FreeBSD.org
State Changed From-To: open->feedback Awaiting maintainers feedback (via the GNATS Auto Assign Tool)
ping....
Hi Bartosz, hi Omer! Omer: Could you please have a look at Bartosz's problem report for mail/isoqlog at [1]? Bartosz: Are you still seeing the problem you reported in ports/139867 [1], or has it been resolved in the meantime? I do not have a fix for the segmentation fault and I don't use isoqlog, but it would probably help if you could send a followup to the problem report and add a backtrace so Omer or someone else can look at it. Best regards, Stefan [1]: http://www.freebsd.org/cgi/query-pr.cgi?pr=ports/139867
Responsible Changed From-To: freebsd-ports-bugs->stefan Track this.
On 2010-05-16 09:49, Stefan Walter wrote: > Hi Bartosz, hi Omer! > > Oh, hi! I almost forgot about this PR ;) Thanks for answer! > Bartosz: Are you still seeing the problem you reported in ports/139867 > [1], or has it been resolved in the meantime? > The problem is still there, with world rebuilded yesterday and rebuilded isoqlog, with and without make.conf variables set - it still drops core with segmentation fault. > I do not have a fix for the segmentation fault and I don't use isoqlog, > but it would probably help if you could send a followup to the problem > report and add a backtrace so Omer or someone else can look at it. > > Best regards, > Stefan > > [1]: http://www.freebsd.org/cgi/query-pr.cgi?pr=ports/139867 > I'm new to unix debugging so a litlle guide would be helpful if what I provided below isn't enough. I've recompiled isoqlog with "CFLAGS=-pipe -g" and execute it to catch core. Here's the dbg output: # gdb /usr/local/bin/isoqlog isoqlog.core GNU gdb 6.1.1 [FreeBSD] Copyright 2004 Free Software Foundation, Inc. GDB is free software, covered by the GNU General Public License, and you are welcome to change it and/or distribute copies of it under certain conditions. Type "show copying" to see the conditions. There is absolutely no warranty for GDB. Type "show warranty" for details. This GDB was configured as "amd64-marcel-freebsd"...(no debugging symbols found)... Core was generated by `isoqlog'. Program terminated with signal 11, Segmentation fault. Reading symbols from /lib/libc.so.7...(no debugging symbols found)...done. Loaded symbols for /lib/libc.so.7 Reading symbols from /libexec/ld-elf.so.1...(no debugging symbols found)...done. Loaded symbols for /libexec/ld-elf.so.1 #0 0x000000000040af20 in ?? () (gbd) Should I recompile something to add debugging symbols? If yes - what and how? Backtrace shows a lot of cryptic lines and ends with: (gdb) backtrace #0 0x000000000040af20 in ?? () #1 0x0000000000408d8b in ?? () #2 0x0000000000409366 in ?? () #3 0x00000000004034c3 in ?? () (...) #695 0x0000000000000001 in ?? () #696 0x00007fffffffea80 in ?? () #697 0x000000000000001c in ?? () Cannot access memory at address 0x800000000000
Hi Bartosz, Bartosz Stec, 17.05.10, 12:06h CEST: > I'm new to unix debugging so a litlle guide would be helpful if what I > provided below isn't enough. I've recompiled isoqlog with "CFLAGS=-pipe > -g" and execute it to catch core. Here's the dbg output: the program is compiled with debugging symbols that way, but the binary is still stripped during "make install". Can you please set WITH_DEBUG=yes in /etc/make.conf, reinstall the port without overriding CFLAGS and then acquire a new backtrace? Regards, Stefan p.s.: Please send followups to the PR as plain text only, not HTML, as it's hardly readable in the web interface.
W dniu 2010-05-18 10:44, Stefan Walter pisze: > Hi Bartosz, > > Bartosz Stec, 17.05.10, 12:06h CEST: > > >> I'm new to unix debugging so a litlle guide would be helpful if what I >> provided below isn't enough. I've recompiled isoqlog with "CFLAGS=-pipe >> -g" and execute it to catch core. Here's the dbg output: >> > the program is compiled with debugging symbols that way, but the binary is > still stripped during "make install". Can you please set WITH_DEBUG=yes in > /etc/make.conf, reinstall the port without overriding CFLAGS and then > acquire a new backtrace? > > Thanks for the hint, here's backtrace ( I've changed domain name to domain.com for privacy): # gdb /usr/local/bin/isoqlog isoqlog.core GNU gdb 6.1.1 [FreeBSD] Copyright 2004 Free Software Foundation, Inc. GDB is free software, covered by the GNU General Public License, and you are welcome to change it and/or distribute copies of it under certain conditions. Type "show copying" to see the conditions. There is absolutely no warranty for GDB. Type "show warranty" for details. This GDB was configured as "amd64-marcel-freebsd"... Core was generated by `isoqlog'. Program terminated with signal 11, Segmentation fault. Reading symbols from /lib/libc.so.7...done. Loaded symbols for /lib/libc.so.7 Reading symbols from /libexec/ld-elf.so.1...done. Loaded symbols for /libexec/ld-elf.so.1 #0 0x000000000040b50a in sortDomainUsersFrom (domain=0x800a1ccd0 "domain.com") at Data.c:224 224 if (tmp1->from_cnt > tmp2->from_cnt) (gdb) bt #0 0x000000000040b50a in sortDomainUsersFrom (domain=0x800a1ccd0 "domain.com") at Data.c:224 #1 0x0000000000404fa8 in createDailyHtml (f=0x800a3b400 "/usr/local/www/sarg/exim/domain.com/2010/5/18.html", d=0x800a1ccd0) at Html.c:206 #2 0x0000000000404910 in createHtml () at Html.c:83 #3 0x000000000040361a in main (argc=1, argv=0x7fffffffea48) at main.c:74 (gdb) Maybe this could be helpful too: # cat isoqlog.conf #isoqlog Configuration file logtype = "exim" # log type qmai-multilog, qmail-syslog, sendmail, postfix logstore = "/var/log/exim/mainlog" # domainsfile = "/usr/local/etc/isoqlog.domains" # outputdir = "/usr/local/www/sarg/exim" # html output directory htmldir = "/usr/local/share/isoqlog/htmltemp" langfile = "/usr/local/share/isoqlog/lang/polish" hostname = "my.domain.com" maxsender = 100 maxreceiver = 100 maxtotal = 100 maxbyte = 100 > p.s.: Please send followups to the PR as plain text only, not HTML, as > it's hardly readable in the web interface. > Sorry for that, my thunderbird client was set to auto choose format and I forgot to manually change it to plain text. -- Bartosz Stec
Bartosz Stec, 18.05.10, 12:55h CEST: > Thanks for the hint, here's backtrace ( I've changed domain name to > domain.com for privacy): Thanks for the backtrace. Omer, can you have a look at this problem? Regards, Stefan
stefan 2010-06-27 17:36:13 UTC FreeBSD ports repository Modified files: mail/isoqlog Makefile Log: Mark as BROKEN on amd64: produces segmentation faults when run. PR: 139867 Submitted by: Bartosz Stec <admin@kkip.pl> Approved by: maintainer timeout (8 months) Feature safe: yes Revision Changes Path 1.12 +7 -1 ports/mail/isoqlog/Makefile _______________________________________________ cvs-all@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/cvs-all To unsubscribe, send any mail to "cvs-all-unsubscribe@freebsd.org"
Hi Bartosz, looking at the build log on an amd64 system, I see a lot of warnings like this one: Data.c: In function 'sortDomainUsersFrom': Data.c:221: warning: cast to pointer from integer of different size Data.c:222: warning: cast to pointer from integer of different size It looks like incorrect use of pointers Data.c and in Html.c (at least these are the obvious ones). It happens to work on i386 because the pointer size matches the size of int there. Since Omer does not respond, I assume the isoqlog project is unmaintained and someone else would have to fix it... How good are your C skills? ;-) (I actually think it shouldn't take too much to fix it, but I don't know when I would get to it, especially since I don't use this software at all.) Regards, Stefan
On 2010-06-30 16:14, Stefan Walter wrote: > Hi Bartosz, > > looking at the build log on an amd64 system, I see a lot of warnings like > this one: > > Data.c: In function 'sortDomainUsersFrom': > Data.c:221: warning: cast to pointer from integer of different size > Data.c:222: warning: cast to pointer from integer of different size > > It looks like incorrect use of pointers Data.c and in Html.c (at least > these are the obvious ones). It happens to work on i386 because the > pointer size matches the size of int there. > > Since Omer does not respond, I assume the isoqlog project is unmaintained > and someone else would have to fix it... How good are your C skills? ;-) > (I actually think it shouldn't take too much to fix it, but I don't know > when I would get to it, especially since I don't use this software at > all.) > > Regards, > Stefan > Now it looks like a challenge to me ;-) My C skills are also 'umaintained' for some years, maybe it's a good opportunity to change that. Your explanation looks sane to me, I'll see what I can do. Cheers, Bartosz Stec
Responsible Changed From-To: stefan->freebsd-ports-bugs Return this PR to the pool - I currently do not have the time to take care of it.
I mechanical changed the code to compile w/o warning http://people.freebsd.org/~arved/stuff/patch-isoqlog Can you try to test if the program now works better for you?
State Changed From-To: feedback->open Feedback received
Hi Bartosz, have you given Tilman's patch (see [1]) a try? Does it fix the problem with mail/isoqlog for you? Regards, Stefan [1]: http://www.freebsd.org/cgi/query-pr.cgi?pr=ports/139867
Unfortunately Tilmans patch doesn't change anything in my case: # gdb /usr/local/bin/isoqlog isoqlog.core GNU gdb 6.1.1 [FreeBSD] Copyright 2004 Free Software Foundation, Inc. GDB is free software, covered by the GNU General Public License, and you are welcome to change it and/or distribute copies of it under certain conditions. Type "show copying" to see the conditions. There is absolutely no warranty for GDB. Type "show warranty" for details. This GDB was configured as "amd64-marcel-freebsd"... Core was generated by `isoqlog'. Program terminated with signal 11, Segmentation fault. Reading symbols from /lib/libc.so.7...done. Loaded symbols for /lib/libc.so.7 Reading symbols from /libexec/ld-elf.so.1...done. Loaded symbols for /libexec/ld-elf.so.1 #0 0x000000000040b50a in sortDomainUsersFrom (domain=0x800a1ccd0 "domain.com") at Data.c:224 224 if (tmp1->from_cnt > tmp2->from_cnt) (gdb) bt #0 0x000000000040b50a in sortDomainUsersFrom (domain=0x800a1ccd0 "domain.com") at Data.c:224 #1 0x0000000000404fa8 in createDailyHtml ( f=0x800a3b400 "/usr/local/www/sarg/exim/domain.com/2010/11/15.html", d=0x800a1ccd0) at Html.c:206 #2 0x0000000000404910 in createHtml () at Html.c:83 #3 0x000000000040361a in main (argc=1, argv=0x7fffffffea40) at main.c:74 Backtrace output is the same as before. Here's build log, and warnings are still there: http://pastebin.com/eAVda465 I patched a port by hashing BROKEN line and then I fetched patch file to mail/isoqlog/files. I also tried to patch manually, which was succesfull, but with the same result.
Hi Bartosz, Bartosz Stec, 15.11.10, 14:02h CET: > I patched a port by hashing BROKEN line and then I fetched patch file to > mail/isoqlog/files. I also tried to patch manually, which was > succesfull, but with the same result. looking at Tilman's patch, I see it's not a patch to be placed into the files/ directory, but a patch to modify the port. Please copy the patch file to the port's directory and run "patch < patch-isoqlog". It should create a couple of files in the files/ directory. After that, retry to build/install the port. Regards, Stefan
>> I patched a port by hashing BROKEN line and then I fetched patch file to >> mail/isoqlog/files. I also tried to patch manually, which was >> succesfull, but with the same result. > looking at Tilman's patch, I see it's not a patch to be placed into the > files/ directory, but a patch to modify the port. Please copy the patch > file to the port's directory and run "patch< patch-isoqlog". It should > create a couple of files in the files/ directory. After that, retry to > build/install the port. > > Regards, > Stefan You're right, I tried it several ways but none of them was right I suppose :) Patching directly from port directory after '#make extract' fixed warnings and segfault is gone too. Thanks for your good work Tilman!
stefan 2010-11-15 15:24:54 UTC FreeBSD ports repository Modified files: mail/isoqlog Makefile Added files: mail/isoqlog/files patch-Data.c patch-Data.h Log: - Fix segmentation faults on amd64. - Reset maintainer due to multiple long response timeouts. PR: 139867 Submitted by: Bartosz Stec <admin@kkip.pl> Patch by: arved Revision Changes Path 1.13 +2 -5 ports/mail/isoqlog/Makefile 1.1 +119 -0 ports/mail/isoqlog/files/patch-Data.c (new) 1.1 +20 -0 ports/mail/isoqlog/files/patch-Data.h (new) _______________________________________________ cvs-all@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/cvs-all To unsubscribe, send any mail to "cvs-all-unsubscribe@freebsd.org"
State Changed From-To: open->closed Patch committed, thanks!