Bug #35179 for DBD-Informix: t93lvarchar fails for DBD-Informix on RHEL5 CentOS5

Fri Apr 18 08:58:55 2008 http://terrybowling.pip.verisignlabs.com/ - Ticket created

CC:	John.Drefke [...] verizon.com, rick.lashua [...] verizon.com
Subject:	t93lvarchar fails for DBD-Informix on RHEL5 CentOS5

Message body is not shown because it is too large.

Fri Apr 25 10:26:49 2008 j.leffler [...] acm.org - Taken

Fri Apr 25 11:06:06 2008 j.leffler [...] acm.org - Correspondence added

RT-Send-CC:

dbd.informix [...] gmail.com

Dear Terry, John, Rick, I apologize for the delay starting to respond to this. I got the bug report a week ago, looked at it, and didn't respond, for which I apologize. I've seen this problem before, and was never able to resolve whether the trouble is in DBD::Informix or in ESQL/C. I haven't seen it on my own systems in quite a while (several years, I'd say), and it was a complete pain to track down when I was seeing it. I'm sorry you've run into it again. I am disappointed that it is resurfacing with a 3.00 release of CSDK. I am moderately sure that it is actually a bug in ESQL/C rather than in DBD::Informix, but demonstrating this is hard. Anecdotal evidence: the problem seems to appear and disappear between releases, behaving differently on different systems. My suspicion is that there is an uninitialized variable that matters - and on some machines (like Solaris for SPARC), the relevant value is zeroed and therefore doesn't show the problem but on other machines, it is not and the problem shows intermittently. I will try and get your environment working, and reproduce the problem again. -- Jonathan Leffler <jleffler@earthlink.net> <jleffler@us.ibm.com> Guardian of DBD::Informix 2008.0229 -- see http://dbi.perl.org/ "I don't suffer from insanity - I enjoy every minute of it!"

Fri Apr 25 11:06:11 2008 The RT System itself - Status changed from 'new' to 'open'

Thu Feb 28 02:51:32 2013 j.leffler [...] acm.org - Correspondence added

On Fri Apr 18 08:58:55 2008, http://terrybowling.pip.verisignlabs.com/ wrote: Show quoted text

> This is similar issue to bug #34490 on Ubuntu, but I wanted to make it > known since RHEL is officially supported by IBM that is wasn't just an > Ubuntu problem.

[...] Show quoted text

> t/t92rows...........ok > t/t93lvarchar.......DBD::Informix::st fetchrow_hashref failed: SQL: > -1820: Host variable type has been changed between fetches or puts. at > /var/tmp/DBD-Informix-2008.0229/blib/lib/DBD/Informix/TestHarness.pm > line 669. > t/t93lvarchar.......dubious > Test returned status 1 (wstat 256, 0x100) > DIED. FAILED test 10 > Failed 1/13 tests, 92.31% okay > t/t94bool...........ok

I finally got this bug nailed. It is indeed a bug in ESQL/C. There's a workaround available which is implemented in DBD::Informix 2013.0118. -- Jonathan Leffler <jleffler@earthlink.net> <jleffler@us.ibm.com> Guardian of DBD::Informix 2008.0229 -- see http://dbi.perl.org/ "I don't suffer from insanity - I enjoy every minute of it!"

Thu Feb 28 02:53:44 2013 j.leffler [...] acm.org - Correspondence added 60 min

RT-Send-CC:

dbd.informix [...] gmail.com

This problem (error -1820 on LVARCHAR) was finally reduced to an ESQL/C bug, and a workaround was created that is implemented in DBD::Informix 2013.0118. This bug is closed now. -- Jonathan Leffler <jleffler@earthlink.net> <jleffler@us.ibm.com> Guardian of DBD::Informix 2013.0118 -- see http://dbi.perl.org/ "I don't suffer from insanity - I enjoy every minute of it!"

Thu Feb 28 02:53:45 2013 j.leffler [...] acm.org - Status changed from 'open' to 'resolved'