Red Hat Bugzilla – Bug 102318
FYI: failed ASSERT on color report from ender
Last modified: 2007-04-18 12:56:51 EDT
[15:20] <ender> any of you seen this: [root@ender root]# rpm -ivh
[15:20] <ender> rpm: rpmte.c:518: rpmteColorDS: Assertion `ix < Count' failed.
[15:20] <ender> Aborted
[15:22] <orc_orc> ender: rpm colorizing is terra incoginta 'here be monsters' per JBJ --
looks as though a selfcheck in that code failed
[15:22] <ender> any idea why this would start all of a sudden?
[15:23] <orc_orc> have you bumped rpm since last compile of that ctcs code?
[15:23] <ender> well, this is the first attempt at a ctcs rpm.
[15:23] <ender> the spec is mind numbingly simple..
[15:24] <orc_orc> cerberus ctcs-1.2.11.tar.gz
[15:24] <orc_orc> built for me ... lemme check
[15:24] <ender> no, this is a complete new spec for it. I'm making a custom one for
[15:26] <ender> I'll have to re-examine the spec file to see what could be causing rpm
to flip out.
[15:27] <orc_orc> in rebuilding I see a lot of arch construct warnings from that code
under RHL 9
[15:28] <ender> I'm not compiling any code.
[15:29] <ender> it's not the building of the rpm that causes problems, its when I
attempt to install the rpm that built just fine.
[15:29] <ender> "built"
[15:29] <ender> would there be a problem with a Version: being set to 1.3.0pre10 ?
[15:30] <orc_orc> that was the reason for my RPM question -- I wondered if it was
compiled under a different rpm version than the install was occurring under, with a
possible change in colorization bit handling in RPM
[15:30] <orc_orc> the Version form is valid
[15:31] <ender> nope, same rpm is used.
[15:31] <ender> as the user, I issue rpmbuild -ba ctcs.spec, it builds, then as root I -ivh
it, and boom.
[15:31] <orc_orc> I am repliacting so I may strace with my older version
[15:33] <orc_orc> ender: happy to look over -- put SRPM somewhere for anon FTP
[15:34] <ender> orc_orc: http://geek.j2solutions.net/stuff/ctcs-1.3.0pre10-1.src.rpm
[15:40] <orc_orc> ender: hmmm -- yours appears to be a binary packaging SRPM --
with a 'from sources' on the older version build, I show several dependencies popping
[15:40] <orc_orc> lemme get the later sources ...
[15:41] <ender> from sources? huh?
[15:42] <ender> yeah, I'm just binary packaging the ctcs stuff. I grabbed the latest .tgz
of ctcs from the sf.net website, and since part of the burnin is compiling the code on
the spot there is no need to distribute precompiled code. I just don't get why it's
causing rpm to freak.
[15:49] <orc_orc> ender: Please try:
http://www.herrold.com/ctcs-1.3.0pre10-1.i386.rpm built from:
[15:50] <orc_orc> ender: you may wish to file a transcript of this for JBJ -- he may want
to see why the ASSERT failed
more than six months old - long since overtaken by events - closing; please
reopen if appropriate