[PATCH] largefiles: fixes failure from clone test with --pull

Matt Mackall mpm at selenic.com
Thu Oct 27 17:56:55 CDT 2011

On Wed, 2011-10-26 at 16:22 -0400, Hao Lian wrote:
> On Wed, Oct 26, 2011 at 4:12 PM, Carter, Eli <Eli.Carter at tektronix.com> wrote:
> > It seems that this should instead update the output.  Afterall, we're trying to test the local clone case here.
> >
> > That said, I'm not seeing this test fail here....
> >
> > Eli
> I'm reproducing on Cygwin Windows. I suspected something was
> conspiring to make this happen on Cygwin Windows only, so I was trying
> to avoid updating the output and breaking the test on other platforms.

Yes, what's happening is that permissions on Cygwin are fake. On Unix,
we have:

  $ chmod -R a-w pubrepo
  $ hg clone ../alice/pubrepo pubrepo

and when we attempt to lock the source repo to do the clone
(http://www.selenic.com/hg/file/tip/mercurial/hg.py#l292), we fail and
fall back to a --pull. The test result is currently correct but adding a
--pull is harmless. Done.

Mathematics is the supreme nostalgia of our time.

More information about the Mercurial-devel mailing list