Bug 1074217

Summary: Build irssi plugin from github.com/cryptodotis/irssi-otr
Product: [Fedora] Fedora Reporter: Martin Milata <mmilata>
Component: irc-otrAssignee: Conrad Meyer <cse.cem+redhatbugz>
Status: CLOSED RAWHIDE QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: unspecified Docs Contact:
Priority: unspecified    
Version: rawhideCC: cse.cem+redhatbugz, pwouters
Target Milestone: ---   
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2014-04-01 15:02:34 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
Possible patch
none
specfile for github-based irssi-otr package
none
Patch accompanying the specfile none

Description Martin Milata 2014-03-08 23:46:05 UTC
Created attachment 872321 [details]
Possible patch

Description of problem: When loading the plugin, irssi segfaults.

Version-Release number of selected component (if applicable):
irssi-0.8.16-0.2.rc1.fc21.x86_64
irssi-otr-0.3-9.git9ea5cc4.fc21.x86_64
libotr-4.0.0-2.fc21.x86_64

How reproducible: 100 %

Steps to Reproduce:
1. run irssi
2. /load otr

Actual results: Segmentation fault.

Expected results: Plugin is loaded.

Additional info:
Attaching untested patch.
The repo at https://github.com/cryptodotis/irssi-otr contains two years worth of fixes for the plugin (including this one). What do you think about basing the package on it? The downside is that it does not contain the plugin for xchat.

Comment 1 Conrad Meyer 2014-03-09 01:21:13 UTC
Hm, that repo dropped xchat here:
https://github.com/cryptodotis/irssi-otr/commit/748c26c93d50002fbcbe405da13481b096aac921

Unfortunate, but I don't use Xchat anymore anyway (and it has gone bitrotted upstream anyway).

I am open to rebasing on the github sources.

Comment 2 Conrad Meyer 2014-03-09 14:31:45 UTC
Btw, I am traveling soon and don't have access to my usual development environment for a week. So if you are a provenpackager or can find one to do the rebase that is welcome. Or if you are patient, it can wait :).

Thanks.

Comment 3 Martin Milata 2014-03-09 14:39:22 UTC
If the xchat plugin still works, it would perhaps be better to leave it in the distribution, drop the irssi-otr subpackage, and create new irssi-otr package based on the github sources.

If it doesn't then I guess it can be dropped altogether and irc-otr rebased on the github sources.

Unfortunately I haven't been able to boot current rawhide into graphical environment so I can't say which is the case.

Comment 4 Paul Wouters 2014-03-10 14:59:02 UTC
the reason we had not yet switched to the github version was that it would mean dropping xchat support from the package. I had talked to the new upstream to either fixup the xchat parts as well, or perhaps pick a different name for their irssi-otr plugin.

Seeing this has not really changed since I met the new irssi-otr people, I'm okay with dropping xchat support and changing upstream for irssi-otr.

Comment 5 Martin Milata 2014-03-11 11:19:09 UTC
Created attachment 873064 [details]
specfile for github-based irssi-otr package

Attaching a specfile (based on upstream specfile) that can be used as a base for the new package.

Comment 6 Martin Milata 2014-03-11 11:20:35 UTC
Created attachment 873065 [details]
Patch accompanying the specfile

Comment 7 Martin Milata 2014-03-11 15:50:28 UTC
It seems xchat-otr-0.3-9.git9ea5cc4.fc21.x86_64 is broken in rawhide - loading the plugin fails with:
/usr/lib64/xchat/plugins/libxchatotr.so: undefined symbol: ircctx_by_peername

Comment 8 Martin Milata 2014-04-01 10:35:24 UTC
Ping, any update? (I'm not a provenpackager btw.)

Comment 9 Conrad Meyer 2014-04-01 12:16:20 UTC
Sorry, other things keeping me busy... you have a proposed spec / patch? This should be pretty quick.

Comment 10 Martin Milata 2014-04-01 12:38:14 UTC
No problem, I don't mind waiting, just wanted to bring this up in case you forgot.

The specfile and patch are in the attachments. I can do the dist-git changes in a private branch somewhere if that would be of any help.

Comment 11 Conrad Meyer 2014-04-01 12:44:03 UTC
Ok, build kicked off here:
http://koji.fedoraproject.org/koji/taskinfo?taskID=6695810

I mostly followed your spec, with a few changes:
- Didn't change the package name since that requires much more bureaucracy than a simple update
- Dropped commit/shortcommit bits and just used the github release URL
- Added Obsoletes tag for irssi-otr < 1.0.0-1

Hopefully that build succeeds (it does locally) and this bug can be resolved. Thanks for your patience and help.

Comment 12 Martin Milata 2014-04-01 12:50:32 UTC
> - Dropped commit/shortcommit bits and just used the github release URL

Using the commit hash is suggested in the packaging guide: https://fedoraproject.org/wiki/Packaging:SourceURL#Github

Comment 13 Conrad Meyer 2014-04-01 13:01:11 UTC
You're right, but I'm going to complain about the policy anyway.

When non-git upstreams release tarballs, they *often* change the underlying source and release a new tarball with the same version number (same problem as moving a git tag). So it seems inconsistently strict to require this of git upstreams.

</rant> And at least git makes this sort of hash readily checkable.

Fixed here:
http://koji.fedoraproject.org/koji/taskinfo?taskID=6695859

Thanks again.

Comment 14 Conrad Meyer 2014-04-01 13:18:41 UTC
Ok, and actually upload the sources...

http://koji.fedoraproject.org/koji/taskinfo?taskID=6695915

Comment 15 Martin Milata 2014-04-01 14:02:02 UTC
Works fine!

Hopefully my last concern: what was previously containted in the irssi-otr subpackage is now in the main irc-otr package. It would be best to request package rename, but you implied that you don't have time for that. Can the package at least have Provides: as described in http://fedoraproject.org/wiki/Packaging:Guidelines#Renaming.2FReplacing_Existing_Packages ?

Comment 16 Conrad Meyer 2014-04-01 14:29:08 UTC
Hm, whatever I was reading before seemed to imply Provides: was unnecessary or deprecated or something. Of course, I don't remember which document I was reading...

Fixed, thanks:
http://koji.fedoraproject.org/koji/taskinfo?taskID=6696092

Comment 17 Martin Milata 2014-04-01 15:10:16 UTC
Great, thanks for your time!