Skip Menu |

This queue is for tickets about the RPC-XML CPAN distribution.

Report information
The Basics
Id: 92931
Status: resolved
Priority: 0/
Queue: RPC-XML

People
Owner: Nobody in particular
Requestors: ANDK [...] cpan.org
Cc: pmqs [...] cpan.org
AdminCc:

Bug Information
Severity: (no value)
Broken in: 0.78
Fixed in: (no value)



CC: PMQS [...] cpan.org
Subject: 70_compression_detect.t SEGV
As per subject. I have CC'd Paul already because it might be that this belongs to the Compress::Zlib queue, but the fail is seen with RJRAY/RPC-XML-0.78.tar.gz. Sample fail report: http://www.cpantesters.org/cpan/report/c4603262-8fae-11e3-b2a6-dfa65d7e0655 More to find in the matrix http://matrix.cpantesters.org/?dist=RPC-XML-0.78 Most of the time I see that non-threaded perls produce the segv and threaded ones survive the tests but I have seen at least one segv also with a threaded perl. Sample stack trace follows Program terminated with signal 11, Segmentation fault. #0 Perl_gp_free (gv=gv@entry=0x2718490) at gv.c:2404 2404 gv.c: No such file or directory. (gdb) bt #0 Perl_gp_free (gv=gv@entry=0x2718490) at gv.c:2404 #1 0x00000000004b95ad in Perl_pp_undef () at pp.c:1027 #2 0x000000000049beb3 in Perl_runops_standard () at run.c:42 #3 0x000000000043a9a8 in S_run_body (oldscope=1) at perl.c:2445 #4 perl_run (my_perl=<optimized out>) at perl.c:2361 #5 0x000000000041dc95 in main (argc=2, argv=0x7fffa5649d98, env=0x7fffa5649db0) at perlmain.c:112 HTH&&Thanks, (Disclaimer: this was discovered by statistical analysis, I'm not speaking as a user)
RT-Send-CC: pmqs [...] cpan.org
I was wondering about this, as I saw the failure reports but not any details. It is possible that the fact that I'm (re)loading Compress::Zlib several times over is confusing either the dynloader or Compress::Zlib itself. I load Compress::Zlib (or attempt to, at least) 3 times total, and it appears that the second attempt is what causes the SEGV. I say this because the first two tests pass (they don't load it, in fact I have a CODE ref in @INC that fakes a load-failure), but tests 3 & 4 seem to not even run, let alone pass or fail. Test 3 is the second time Compress::Zlib gets loaded, and it seems to be where the SEGV is occurring. Note that I myself test with both threaded and non-threaded Perls, and have not seen this error on my dev platforms. Hoping that Paul might have some insight on this. Randy -- Randy J. Ray rjray@blackperl.com randy.j.ray@gmail.com
Hey guys, Looking at the cpantesters matrix, it looks like the issue kicked in with 5.19.7 Just tried building RPC-XML 0.78 with perl 5.19.9 - it triggered the same failure in 70_compression_detect.t Why this is happening is a good question - will need to work backwards through old versions of Compress::Raw::Zlib to see which version introduced the issue. Paul
Think this is an XSLoader issue - hacked the 70_compression_detect.t file to use a dummy module that just has a single XS symbol & reproduced the issue. That took both Compress::* & RPC-XML out of the equation. Will sync my bleadperl & see if the issue still exists. If it does I'll flag it to p5p. Paul
RT-Send-CC: pmqs [...] cpan.org
Good to know that this isn't anything I've done! Will hold this ticket open until you're able to report back on whether this goes to p5p or not. Randy -- Randy J. Ray rjray@blackperl.com randy.j.ray@gmail.com
The bug has been found in perl and fixed in bleadperl: https://rt.perl.org/Public/Bug/Display.html?id=121242 I have verified it works. Closing ticket. Thanks everybody!