[Gnash-dev] First commit-triggered builds (kind of bogus)

[strk]

On Thu, Dec 16, 2010 at 09:27:17AM +0000, Benjamin Wolsey wrote
(on gnash-commit)

> The branch, master has been updated
>        via  5c257844a719b2beafbeb49009d517f3a484f23c (commit)
>        via  e3c4fa707d8c2eec62e45df40bae0ff24bb1aefa (commit)
>        via  2340e2ee2eedc2ac79431e0a08fea7f564f4d142 (commit)
>        via  bd05dbd503249a352a7d53844d16bfbae7d26509 (commit)
>        via  6a28902da8ad153f587601b4c9520e8343081786 (commit)
>       from  011c29e10ae1377cf92d7202c84e407524c20f24 (commit)

FYI: this is the first commit really triggering builds on buildbot.

Unfortunately, the change notified only reports the very first
commit [1] rather than the last one.

[1] http://gnashdev.org:8010/changes/4

As a result, the update steps of the two triggered builds
is fetching the older commit in the push, rather than
the newer, reporting (in [2] and [3])

 fatal: Could not parse object '6a28902da8ad153f587601b4c9520e8343081786'.
 ....
 HEAD is now at 6a28902 Const correct.

[2] 
http://gnashdev.org:8010/builders/lucid-linux-x86_64/builds/44/steps/git/logs/stdio
[3] 
http://gnashdev.org:8010/builders/squeeze-linux-x86/builds/70/steps/git/logs/stdio

Sounds like something that should be fixed in buildbot upstream.

--strk;

  ()   Free GIS & Flash consultant/developer
  /\   http://strk.keybit.net/services.html