1033294 - Errors applying patches on tb-try when following instructions from the wiki

Status: RESOLVED INVALID

(MailNews Core :: Build Config, defect)

Description

[Mike Hommey [:glandium]]

I was pushing stuff to try-c-c today and got repeated failures to apply the m-c patch I had in my tree. The problem is I kept getting the same set of slaves, and when client.py runs and updates the mozilla/ tree, it doesn't cleanup beforehand, so the patch applied in the previous build is still applied, and thus fails to reapply.

To circumvent, I applied this:

--- a/client.py
+++ b/client.py
@@ -342,17 +342,17 @@ def do_hg_pull(dir, repository, hg, rev,
                          retryMax=options.retries)
     else:
         cmd = [hg, 'pull', '-R', fulldir] + hgopts
         if repository is not None:
             cmd.append(repository)
         check_call_noisy(cmd, retryMax=options.retries)
 
     # update to specific revision
-    cmd = [hg, 'update', '-r', rev, '-R', fulldir ] + hgopts
+    cmd = [hg, 'update', '-C', '-r', rev, '-R', fulldir ] + hgopts
     if options.verbose:
         cmd.append('-v')
     # Explicitly never retry 'hg update': otherwise any merge failures are ignored.
     # This command is local: a failure can't be caused by a network error.
     check_call_noisy(cmd, retryMax=0)
 
     check_call([hg, 'parent', '-R', fulldir,
                 '--template=Updated to revision {node}.\n'])

I'm not sure if this is something that can be done generally, though. In particular, I'm worried about the impact on non-automation workflows.