Hi Will I can't put my finger on it, but B::Hooks::OP was last released on 11 Sept 2011, quite a while ago, and there could be a problem with recent versions of perl.
Regards, Jie * William N. Braswell, Jr. <william.brasw...@autoparallel.com> wrote: > Date: Tue, 22 Nov 2016 23:14:11 -0600 > From: "William N. Braswell, Jr." <william.brasw...@autoparallel.com> > To: Jie Gao <j....@sydney.edu.au> > CC: modperl@perl.apache.org > Subject: Re: mod_perl Silent Failure, Very Mysterious > > Hi Jie, > If I temporarily disable B::Hooks::OP::Check, as shown below, then it > tells me it was included from Devel::Declare, which was in turn > (indireclty) included from Method::Signatures::Simple. > I am not the author of either Catalyst or ShinyCMS, so unfortunately I > don't know any more in-depth details off the top of my head. > Does that answer your question? > Thanks,~ Will > wbraswell@cloud-comp0-00:~$ pm_location.sh > B::Hooks::OP::Check/home/wbraswell/perl5/lib/perl5/x86_64-linux-gnu-thread-multi/B/Hooks/OP/Check.pm > wbraswell@cloud-comp0-00:~$ mv > /home/wbraswell/perl5/lib/perl5/x86_64-linux-gnu-thread-multi/B/Hooks/OP/Check.pm > /home/wbraswell/perl5/lib/perl5/x86_64-linux-gnu-thread-multi/B/Hooks/OP/Check.pm.DISABLED > wbraswell@cloud-comp0-00:~/public_html/cloud-comp0-00.autoparallel.com-latest$ > perl -e 'use ShinyCMS;'Can't locate B/Hooks/OP/Check.pm in @INC (you > may need to install the B::Hooks::OP::Check module) (@INC contains > ....) at > /home/wbraswell/perl5/lib/perl5/x86_64-linux-gnu-thread-multi/Devel/Declare.pm > line 18.BEGIN failed--compilation aborted at > /home/wbraswell/perl5/lib/perl5/x86_64-linux-gnu-thread-multi/Devel/Declare.pm > line 18.Compilation failed in require at > /home/wbraswell/perl5/lib/perl5/x86_64-linux-gnu-thread-multi/Devel/Declare/Context/Simple.pm > line 5.BEGIN failed--compilation aborted at > /home/wbraswell/perl5/lib/perl5/x86_64-linux-gnu-thread-multi/Devel/Declare/Context/Simple.pm > line 5.Compilation failed in require at /usr/share/perl/5.22/base.pm > line 97. ...propagated at /usr/share/perl/5.22/base.pm line 106.BEGIN > failed--compilation aborted at > /home/wbraswell/perl5/lib/perl5/x86_64-linux-gnu-thread-multi/Devel/Declare/MethodInstaller/Simple.pm > line 3.Compilation failed in require at /usr/share/perl/5.22/base.pm > line 97. ...propagated at /usr/share/perl/5.22/base.pm line 106.BEGIN > failed--compilation aborted at > /home/wbraswell/perl5/lib/perl5/Method/Signatures/Simple.pm line > 19.Compilation failed in require at blib/lib/ShinyCMS.pm line 21.BEGIN > failed--compilation aborted at blib/lib/ShinyCMS.pm line > 21.Compilation failed in require at -e line 1.BEGIN > failed--compilation aborted at -e line 1. > On 11/22/2016 at 11:03 PM, "Jie Gao" wrote:Hi William > > Where excatly is B::Hooks::OP::Check required / its Check.xs called? > Regards, > Jie > > * William N. Braswell, Jr. wrote: > > > Date: Tue, 22 Nov 2016 06:04:36 -0600 > > From: "William N. Braswell, Jr." > > To: Jie Gao > > CC: modperl@perl.apache.org > > Subject: RE: mod_perl Silent Failure, Very Mysterious > > > > Hello Again Jie, > > I have run the modperl_testreport.pl script per your direction, and > > have attached the output file. > > Also, I have posted this as a mod_perl issue: > > > > https://rt.cpan.org/Ticket/Display.html?id=118900 > > I am attempting to run an already-bug-free Catalyst application > under > > mod_perl, which should work just fine but instead gives a segfault > in > > the Check.xs file. > > The name of the Catalyst application is ShinyCMS, it works just fine > > using the PSGI stand-alone testing web server as well as fastcgi, > and > > mod_perl is the only option which causes a segfault. I must have > > mod_perl support for my specific use of ShinyCMS, so I really really > > need to fix this segfault and get Shiny working under mod_perl. > Yes, > > Shiny is specifically designed to work with mod_perl, as are most > > Catalyst apps. In fact, my simple ">" print statements in > ShinyCMS.pm > > show that Shiny is totally loading and returning control back to > > Apache (or mod_perl?) before the actual segfault occurs. > > My system info: > > $ apache2 -vServer version: Apache/2.4.18 (Ubuntu)Server built: > > 2016-07-14T12:32:26 > > $ perl -vThis is perl 5, version 22, subversion 1 (v5.22.1) built > for > > x86_64-linux-gnu-thread-multi(with 58 registered patches, see perl > -V > > for more detail) > > $ pm_version.pl mod_perl2$mod_perl2::VERSION = 2.000009 > > $ pm_version.pl B::Hooks::OP::Check$B::Hooks::OP::Check::VERSION = > > 0.19 > > $ cat /etc/issueUbuntu 16.04.1 LTS > > I have spent the last week trying all of the following and more, but > > the segfault persists: > > 1. manually-compiled perl v5.22.1 & mod_perl2 v2.000009 w/out > > threads2. mod_perl2 back to v2.0000053. B::Hooks::OP::Check back > to > > v0.17 (maybe earlier)4. move all possible 'use' statements to be > > called as early as possible in Shiny_dependencies.pm > > I also tried Apache v2.2 & mod_perl2 v2.000008 & Perl v5.18.2 in > > Ubuntu v14.04, it did not segfault but instead it served up totally > > blank pages with 0 bytes of data. I have not dug deeper because I > use > > Ubuntu v16.04 in production and I figure it is harder to debug w/out > > an actual segfault to start from. > > I have set up SSH access on a cloud server so you can see and debug > > the segfault directly, all 3 of the usernames below are set as > sudoers > > so please wield your root powers with caution: > > HOSTNAME: cloud-comp0-00.autoparallel.comHOST IP: > > 64.137.255.53USERNAME: shay OR fmoyer OR jgao OR > modperl_teamPASSWORD: > > check your CPAN e-mail address > > Pertinent paths: > > > /home/wbraswell/public_html/cloud-comp0-00.autoparallel.com-latest/etc/apache2/sites-available/cloud-comp0-00.autoparallel.com.conf > > Here is an example gdb showing Check.xs as the likely culprit, > > starting with `sudo` and then `source` to load Apache env vars: > > [[[ BEGIN GDB SESSION ]]] > > zefram@cloud-comp0-00:/$ sudo -i...root@cloud-comp0-00:~# source > > /etc/apache2/envvarsroot@cloud-comp0-00:~# gdb > > /usr/sbin/apache2...Reading symbols from /usr/sbin/apache2...(no > > debugging symbols found)...done. > > (gdb) run -k start -XStarting program: /usr/sbin/apache2 -k start > > -X[Thread debugging using libthread_db enabled]Using host > libthread_db > > library "/lib/x86_64-linux-gnu/libthread_db.so.1".>: top of > > ShinyCMS.pm>: in ShinyCMS.pm, about to use Moose>: in ShinyCMS.pm, > > about to use Catalyst::Runtime>: in ShinyCMS.pm, about to use > > Catalyst>: in ShinyCMS.pm, about to use CatalystX::RoleApplicator>: > in > > ShinyCMS.pm, about to use Method::Signatures::Simple>: in > ShinyCMS.pm, > > about to call config()>: in ShinyCMS.pm, have __PACKAGE__ = > ShinyCMS>: > > in ShinyCMS.pm, about to call setup()...>: in ShinyCMS.pm, returned > > from setup()>: in ShinyCMS.pm, about to return 1 > > Program received signal SIGSEGV, Segmentation > fault.0x00007fffebabcb10 > > in ?? ()(gdb) bt#0 0x00007fffebabcb10 in ?? ()#1 > 0x00007ffff3e5302b > > in Perl_newUNOP (my_perl=my_perl@entry=0x5555577ddde0, > > type=type@entry=17, flags=, flags@entry=8192, first=0x5555562141b8) > at > > op.c:4811#2 0x00007ffff3e54a1d in Perl_newCVREF > > (my_perl=my_perl@entry=0x5555577ddde0, flags=flags@entry=8192, o=) > at > > op.c:9367#3 0x00007ffff3e8b686 in Perl_yylex > > (my_perl=my_perl@entry=0x5555577ddde0) at toke.c:6693#4 > > 0x00007ffff3e97228 in Perl_yyparse > > (my_perl=my_perl@entry=0x5555577ddde0, gramtype=gramtype@entry=258) > at > > perly.c:322...#44 0x000055555558709f in main () > > (gdb) startThe program being debugged has been started already.Start > > it from the beginning? (y or n) yTemporary breakpoint 1 at > > 0x5555555867b0Starting program: /usr/sbin/apache2 -k start -X[Thread > > debugging using libthread_db enabled]Using host libthread_db library > > "/lib/x86_64-linux-gnu/libthread_db.so.1". > > Temporary breakpoint 1, 0x00005555555867b0 in main () > > (gdb) break perl_parseFunction "perl_parse" not defined.Make > > breakpoint pending on future shared library load? (y or [n]) > > yBreakpoint 2 (perl_parse) pending. > > (gdb) continueContinuing. > > Breakpoint 2, perl_parse (my_perl=0x55555581ae50, > > xsinit=0x7ffff420a280, argc=2, argv=0x7ffff7f88070, env=0x0) at > > perl.c:14731473 perl.c: No such file or directory. > > (gdb) watch -l PL_check[17]Hardware watchpoint 3: -location > > PL_check[17] > > (gdb) continueContinuing. > > Hardware watchpoint 3: -location PL_check[17] > > Old value = (OP *(*)(PerlInterpreter *, OP *)) 0x7ffff3e4d9e0 New > > value = (OP *(*)(PerlInterpreter *, OP *)) 0x7fffebabcb10 > > 0x00007fffebabcd0e in hook_op_check (type=type@entry=OP_RV2CV, > > cb=cb@entry=0x7fffeb8b6d30 , user_data=user_data@entry=0x0) at > > Check.xs:6666 PL_check[type] = check_cb; > > (gdb) info threads Id Target Id Frame * 1 Thread > > 0x7ffff7fe2780 (LWP 6196) "/usr/sbin/apach" 0x00007fffebabcd0e in > > hook_op_check (type=type@entry=OP_RV2CV, cb=cb@entry=0x7fffeb8b6d30 > , > > user_data=user_data@entry=0x0) at Check.xs:66 > > (gdb) info sourceCurrent source file is Check.xsCompilation > directory > > is > > > /home/wbraswell/.cpanm/work/1479735542.24787/B-Hooks-OP-Check-0.19Located > > in > > > /home/wbraswell/.cpanm/work/1479735542.24787/B-Hooks-OP-Check-0.19/Check.xsContains > > 106 lines.Source language is c.Producer is GNU C11 5.4.0 20160609 > > -mtune=generic -march=x86-64 -g -O2 -fwrapv -fno-strict-aliasing > -fPIC > > -fstack-protector-strong.Compiled with DWARF 2 debugging format.Does > > not include preprocessor macro info. > > (gdb) continueContinuing.>: top of ShinyCMS.pm>: in ShinyCMS.pm, > about > > to use Moose>: in ShinyCMS.pm, about to use Catalyst::Runtime>: in > > ShinyCMS.pm, about to use Catalyst>: in ShinyCMS.pm, about to use > > CatalystX::RoleApplicator>: in ShinyCMS.pm, about to use > > Method::Signatures::Simple>: in ShinyCMS.pm, about to call > config()>: > > in ShinyCMS.pm, have __PACKAGE__ = ShinyCMS>: in ShinyCMS.pm, about > to > > call setup()...>: in ShinyCMS.pm, returned from setup()>: in > > ShinyCMS.pm, about to return 1 > > ... > > Program received signal SIGSEGV, Segmentation > fault.0x00007fffebabcb10 > > in ?? ()(gdb) bt#0 0x00007fffebabcb10 in ?? ()#1 > 0x00007ffff3e5302b > > in Perl_newUNOP (my_perl=my_perl@entry=0x555559dd8ed0, > > type=type@entry=17, flags=, flags@entry=8192, first=0x55555824eb98) > at > > op.c:4811#2 0x00007ffff3e54a1d in Perl_newCVREF > > (my_perl=my_perl@entry=0x555559dd8ed0, flags=flags@entry=8192, o=) > at > > op.c:9367...#44 0x000055555558709f in main () > > (gdb) print PL_check[17]$1 = (Perl_check_t) 0x7fffebabcb10 > > (gdb) quit > > [[[ END GDB SESSION ]]] > > I have attached the unabbreviated gdb session and the system info in > > text files, for your reference. > > I am anxious (near desperate) to get this bug fixed so that I can > move > > forward with my work. I will eagerly assist in any way I can, > please > > just let me know what I should do and I will do it! > > Thanks so much for your help. > > Perling,~ Will > > > > On 11/16/2016 at 3:17 AM, "Jie Gao" wrote: For the completeness > > of issue reporting, please run the following script and post the > > output: > > #!/usr/bin/perl > > use strict; > > use warnings FATAL => 'all'; > > use ModPerl::TestReport; > > ModPerl::TestReport->new(@ARGV)->run; > > Jie > > ------------------------- > > From: William N. Braswell, Jr. [william.brasw...@autoparallel.com] > > Sent: Tuesday, November 15, 2016 8:01 PM > > To: Jie Gao > > Cc: modperl@perl.apache.org > > Subject: Re: mod_perl Silent Failure, Very Mysterious > > Hello Jie & Friends, > > I got the core dump out of gdb and also some hopefully-helpful > > output. > > Here is the output from gdb (not the core dump which is binary of > > course): > > http://hastebin.com/okaxacuriq.scala > > Then again maybe we have a problem because I'm missing debugging > > symbols inside /usr/sbin/apache2? > > _Program received signal SIGSEGV, Segmentation fault. > > 0x00007fffeed67a00 in ?? ()_ > > Note the _??_ part in the output above... > > Where do I go from here? > > Thanks, ~ Will > > On 11/15/2016 at 12:14 AM, "William N. Braswell, Jr." wrote: Hi > > Jie, > > I have captured the output of `strace -f /usr/sbin/apache2 -e > trace8 > > -k start`: > > http://hastebin.com/unupocozop.pl > > Looks like it is dying with a segfault, but I'm not sure what is > > causing this??? > > Thanks so much for your help!!! :-) > > ~ Will > > [[[ BEGIN PASTE SNIPPET ]]] > > [pid 6550] stat("blib/lib/DynaLoader.pmc", 0x7ffdda87f280) = -1 > > ENOENT (No such file or directory) [pid 6550] > > stat("blib/lib/DynaLoader.pm", 0x7ffdda87f1b0) = -1 ENOENT (No such > > file or directory) [pid 6550] stat("/etc/perl/DynaLoader.pmc", > > 0x7ffdda87f280) = -1 ENOENT (No such file or directory) [pid 6550] > > stat("/etc/perl/DynaLoader.pm", 0x7ffdda87f1b0) = -1 ENOENT (No such > > file or directory) [pid 6550] > > stat("/usr/local/lib/x86_64-linux-gnu/perl/5.22.1/DynaLoader.pmc", > > 0x7ffdda87f280) = -1 ENOENT (No such file or directory) [pid 6550] > > stat("/usr/local/lib/x86_64-linux-gnu/perl/5.22.1/DynaLoader.pm", > > 0x7ffdda87f1b0) = -1 ENOENT (No such file or directory) [pid 6550] > > stat("/usr/local/share/perl/5.22.1/DynaLoader.pmc", 0x7ffdda87f280) > = > > -1 ENOENT (No such file or directory) [pid 6550] > > stat("/usr/local/share/perl/5.22.1/DynaLoader.pm", 0x7ffdda87f1b0) = > > -1 ENOENT (No such file or directory) [pid 6550] > > stat("/usr/lib/x86_64-linux-gnu/perl5/5.22/DynaLoader.pmc", > > 0x7ffdda87f280) = -1 ENOENT (No such file or directory) [pid 6550] > > stat("/usr/lib/x86_64-linux-gnu/perl5/5.22/DynaLoader.pm", > > 0x7ffdda87f1b0) = -1 ENOENT (No such file or directory) [pid 6550] > > stat("/usr/share/perl5/DynaLoader.pmc", 0x7ffdda87f280) = -1 ENOENT > > (No such file or directory) [pid 6550] > > stat("/usr/share/perl5/DynaLoader.pm", 0x7ffdda87f1b0) = -1 ENOENT > (No > > such file or directory) [pid 6550] > > stat("/usr/lib/x86_64-linux-gnu/perl/5.22/DynaLoader.pmc", > > 0x7ffdda87f280) = -1 ENOENT (No such file or directory) [pid 6550] > > stat("/usr/lib/x86_64-linux-gnu/perl/5.22/DynaLoader.pm", > > {st_mode=S_IFREG|0644, st_size=10748, ...}) = 0 [pid 6550] > > open("/usr/lib/x86_64-linux-gnu/perl/5.22/DynaLoader.pm", O_RDONLY) > = > > 5 [pid 6550] ioctl(5, TCGETS, 0x7ffdda87ef50) = -1 ENOTTY > > (Inappropriate ioctl for device) [pid 6550] lseek(5, 0, SEEK_CUR) > > > = 0 [pid 6550] read(5, "# Generated from DynaLoader_pm.P"..., > 8192) > > = 8192 [pid 6550] --- SIGSEGV {si_signo=SIGSEGV, > si_code=SEGV_MAPERR, > > si_addr=0x7f491064fa00} --- [pid 6550] chdir("/etc/apache2") > = > > 0 [pid 6550] rt_sigaction(SIGSEGV, {SIG_DFL, [], > > SA_RESTORER|SA_INTERRUPT, 0x7f4918dea3d0}, {SIG_DFL, [], > > SA_RESTORER|SA_RESETHAND, 0x7f4918dea3d0}, 8) = 0 [pid 6550] > > kill(6550, SIGSEGV) = 0 [pid 6550] rt_sigreturn({mask=[]}) > > > = 139951873083648 [pid 6550] --- SIGSEGV {si_signo=SIGSEGV, > > si_code=SI_USER, si_pid=6550, si_uid=0} --- [pid 6549] +++ exited > > with 0 +++ +++ killed by SIGSEGV (core dumped) +++ > > [[[ END PASTE SNIPPET ]]] > > On 11/14/2016 at 11:54 PM, "Jie Gao" wrote: Add the "-f" flag to > > strace and see what errors you get. > > Regards, > > Jie > > * William N. Braswell, Jr. wrote: > > > Date: Mon, 14 Nov 2016 23:27:53 -0600 > > > From: "William N. Braswell, Jr." > > > To: modperl@perl.apache.org > > > Subject: mod_perl Silent Failure, Very Mysterious > > > > > > Hello everyone, I'm trying to enable a web app (Catalyst > ShinyCMS) > > to > > > run under mod_perl, it is definitely supposed to work but it > fails > > > without any errors, and I've been trying for several days to get > it > > > working. > > > My system info: > > > * Perl 5, version 22, subversion 1 (v5.22.1) built for > > > x86_64-linux-gnu-thread-multi > > > * libapache2-mod-perl2 v2.0.9-4ubuntu1 > > > * Apache/2.4.18 (Ubuntu) > > > * $Apache::Test::VERSION = 1.39 > > > * $Catalyst::VERSION = 5.90112 > > > * $ShinyCMS::VERSION = 0.007 > > > I know the app already functions properly when running via the > > > stand-alone test server script (PSGI) or fastcgi, and I am able > to > > run > > > a trivial Catalyst test application successfully under mod_perl, > so > > > this issue seems to have something to do with the interaction > > between > > > the ShinyCMS app and mod_perl. > > > I have added print debug statements in my app, so when I call my > > app > > > from within the Apache site conf file then I can see the debug > > > statements prove that it is successfully going all the way > through > > my > > > app, and it even returns control to the embedded code w/in the > > Apache > > > site conf file, but it still just silently fails anyway with the > > > following non-descriptive error: > > > "The apache2 instance did not start within 20 seconds. Please > read > > the > > > log files to discover problems" > > > Unfortunately the log files do not provide any additional > > information > > > whatsoever! > > > Here are my Apache config files and the output of trying to start > > > Apache manually, you can see my ">" statements which prove that > my > > > ShinyCMS app code is definitely executing all the way through and > > > returning control to Apache: > > > http://hastebin.com/rulixitole.php > > > http://hastebin.com/wevicagano.rb > > > This is a really mysterious issue, when I run `apache2ctl -e > debug > > -k > > > start` as seen in the link above, it goes through the motions as > if > > it > > > is starting correctly, and it does not give any error messages, > but > > > when apache2ctl is done there simply are not any apache processes > > > running... > > > I only found 1 person on google with a possibly-similar issue, > but > > he > > > told me that he never found a solution: > > > https://github.com/timbunce/devel-nytprof/issues/89 > > > I also found 1 person on the mod_perl mailing list archives with > > > another possibly-similar issue, but I have no way to follow up, > and > > > I'm not sure of the best way to totally rebuild everything from > > > source: > > > > > > https://mail-archives.apache.org/mod_mbox/perl-modperl/200207.mbox/%3c4a19f19a1bb4d51183220020ed06971401288...@exch2.iso-ne.com%3E > > > I have asked for help on IRC #httpd, user Unbeliever suggested I > > > ascertain my real /usr/sbin/apache2 command line and run it > through > > > strace, then post to this mailing list. > > > I used -x in apache2ctl which showed me I was simply sourcing > > > /tmp/apache2/envvars and then running `/usr/sbin/apache2 -e > trace8 > > -k > > > start` which are the exact same args as accepted by the > apache2ctl > > > script. > > > I have now run `strace /usr/sbin/apache2 -e trace8 -k start` > which > > > generates 2.5MB of text output, I have pasted the last few pages > of > > it > > > on hastebin, I can post the whole thing for download if needed: > > > http://hastebin.com/idemogexiq.erl > > > I am not sure how to interpret the results of the strace command > > > output and debug this issue, help please? > > > Any and all assistance is greatly appreciated! :-) > > > Thanks,~ Will
