[Buildbot-commits] [Buildbot] #1055: missing a no-clobber option when svn fail to update? (and other rcs?)

[Buildbot]

#1055: missing a no-clobber option when svn fail to update? (and other rcs?)
-------------------+--------------------------------------------------------
Reporter:  blaize  |       Owner:       
    Type:  defect  |      Status:  new  
Priority:  major   |   Milestone:  0.8.+
 Version:  0.8.2   |    Keywords:  svn  
-------------------+--------------------------------------------------------
Changes (by dustin):

  * keywords:  svn update clobber => svn
  * milestone:  undecided => 0.8.+


Old description:

> I'm running buildbot with an svn revision control system on winxp.
> Periodically the svn update step fails.  Then  maybeDoVCFallback() is
> called in
>
> \buildbot_slave-0.8.1-py2.6.egg\buildslave\vcs.py
>
> Which does a clobber.  Which deletes my source tree.
>
> Now, my problems are (or seem to be to me):
>
>   i) when the SVN step fails it never recovers from this (it clobbers but
> doesn't seem to checkout the source when it does).
>
>   ii) I have a 4Gb svn trunk dir (yes that is crazy) and I'd much much
> much rather the SVN step just failed and continued to fail waiting on
> human intervention.
>
> There doesn't seem to be an obvious way to turn this behaviour off - is
> that right?  Amber Yust took a quick look and seems to think that this is
> indeed a bug.
>
> thanks

New description:

 I'm running buildbot with an svn revision control system on winxp.
 Periodically the svn update step fails.  Then  maybeDoVCFallback() is
 called in

 \buildbot_slave-0.8.1-py2.6.egg\buildslave\vcs.py

 Which does a clobber.  Which deletes my source tree.

 Now, my problems are (or seem to be to me):

   i) when the SVN step fails it never recovers from this (it clobbers but
 doesn't seem to checkout the source when it does).

   ii) I have a 4Gb svn trunk dir (yes that is crazy) and I'd much much
 much rather the SVN step just failed and continued to fail waiting on
 human intervention.

 There doesn't seem to be an obvious way to turn this behaviour off - is
 that right?  Amber Yust took a quick look and seems to think that this is
 indeed a bug.

 thanks

--

Comment:

 What should Buildbot do in this situation?  I assume (i) is related to the
 original subversion failure.  How can Buildbot distinguish this situation
 from other subversion failures (such as an update that fails because a
 generated file is in the way)?

-- 
Ticket URL: <http://buildbot.net/trac/ticket/1055#comment:1>
Buildbot <http://buildbot.net/>
Buildbot: build/test automation