[Svnmerge] svnmerge: command execution failed (exit code: 1)

[Mike Laster]

On 9/1/06, Daniel Rall <dlr at collab.net> wrote:
> On Tue, 29 Aug 2006, Mike Laster wrote:
>
> > On 8/29/06, Madan U Sreenivasan <madan at collab.net> wrote:
> > > On Tue, 29 Aug 2006 01:52:53 +0530, Mike Laster <mlaster at gmail.com> wrote:
> > >
> > > > I'm getting a reprodcable crash in the latest svnmerge:
> > > >
> > > > svnmerge r20074
> > > >   modified: 2006-06-13 01:59:38 -0400 (Tue, 13 Jun 2006)
> ...
> > > Could you try reproducing this on some publicly available (or generated)
> > > data for us? I could use this to fix this problem.
> >
> > I haven't had any luck reproducing it yet outside of my repository.  It doesn't
> > seem to be entirely size related since I found bigger transactions in
> > the Subversion
> > repository that didn't choke it.  I've added some debugging print statements
> > and it seems that the readlines() call in launch is what is dying.
> >
> > p.wait() is returning 256 instead of the normal 0 for a successful run.
> >
> > I'm running on Mac OS X 10.4.7 using Python 2.4.3
>
> What causes p.wait() to return 256 (e.g. a timeout reached)?

I changed this in the launch method:

    if os.name not in ['nt', 'os2']:

to

    if os.name not in ['nt', 'os2', 'posix']:

In order to use the second implementation and it seems to work this way.

So the problems seems to be related to using popen2.Popen4() vs os.popen4()