Bug 421385 – net-im/jabberd fails to build with net-libs/gnutls-3

• Status: RESOLVED FIXED
• Product: Gentoo Linux

[Description Diego Elio Pettenò (RETIRED) 2012-06-16 13:37:13 UTC]

See the log file in the URL

[Comment 1 Pacho Ramos 2012-06-16 15:35:04 UTC]

This was finally not dropped, but upstream keeps inactive for ages and I couldn't find any distribution still shipping it from where we could get a fix for this. If a patch is not provided, we should probably finally kill it (also, if a patch appears, we should probably also think about really deserving the effort as probably it will break again in the future since this package is stalled for years)

[Comment 2 Tony Vroon (RETIRED) 2012-11-20 10:52:11 UTC]

Created attachment 330048 [details, diff]
jabberd-1.6.1.1-gnutls-extra-header-removed.patch

This had a broken configure test (subtle, but so obvious now that you see it...). Newer gnutls provides the GPG functionality, but without the "extra" header & start-up call. This builds. Please test.

[Comment 3 Tony Vroon (RETIRED) 2012-11-20 10:52:42 UTC]

Created attachment 330050 [details]
jabberd-1.6.1.1-r2.ebuild

Proposed ebuild. Only adds the application of one patch.

[Comment 4 Tony Vroon (RETIRED) 2012-11-20 11:16:50 UTC]

I can confirm that this builds against both GnuTLS 2.10 & 3; but am awaiting functionality tests from the community at this point. Should this be a viable fix I propose we add this ebuild to the tree with a recommendation to migrate away to ejabberd. It would offer people a slightly more graceful exit, and allow the gnutls 2.10 removal to occur. Thoughts?

[Comment 5 Markos Chandras (RETIRED) 2012-11-20 11:20:27 UTC]

(In reply to comment #4)
> I can confirm that this builds against both GnuTLS 2.10 & 3; but am awaiting
> functionality tests from the community at this point. Should this be a
> viable fix I propose we add this ebuild to the tree with a recommendation to
> migrate away to ejabberd. It would offer people a slightly more graceful
> exit, and allow the gnutls 2.10 removal to occur. Thoughts?

I think there is no point in trying to keep this package alive. We will hit other problems sooner or later

[Comment 6 Tony Vroon (RETIRED) 2012-11-20 11:41:57 UTC]

Of course we will. But by making it possible to run jabberd now, we allow people a graceful exit. We get a chance to add an ewarn saying "This package is going to go away, your suggested replacement is ejabberd. Please begin your migration now."
Rather then "*bam* installed package is masked, goodbye!".

[Comment 7 Markos Chandras (RETIRED) 2012-11-20 12:17:17 UTC]

Masking a package provides a graceful exit of 30 days :). I don't mind either way

[Comment 8 Tony Vroon (RETIRED) 2012-11-20 18:04:17 UTC]

Community interest has been below expectations. Let's remove and keep the patch on the bug only then. Will leave you to it.

[Comment 9 Markos Chandras (RETIRED) 2012-11-20 19:20:04 UTC]

Thanks. Removed