Hi,
I am a partial committer for Ruby tools in the Subversion project (and a
Debian user/translator).
> I already patched around this by adding a 'sleep 1' in there, but
> apparently this is insufficient on m68k. I've changed it to 'sleep 2'
> for the next upload, but the correct fix is a retry loop of some sort.
> Since I don't know ruby, anyone who does know ruby is invited to help
> us out: see 'debian/patches/ruby-test-svnserve-race' in the source
> package.
I'm going to write a patch (and propose to apply it in the upstream)
tomorrow. I'm planning to write a patch to make SvnTestUtil::setup_svnserve
wait until the connection is accepted.
> (Note, the impact of the 'sleep N' is about N*64 seconds of build time
> on every build, every architecture. So I don't want to just make N
> really large.)
Exactly. SvnTestUtil::setup_svnserve is called in the start-up of all
the test methods in test/test_{client,delta,error,fs,info,ra,repos,wc}.
Thanks,
-nori
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
