Why Did Mercurial Add Files During a Merge Between Branches?

Wujek Srujek wujek.srujek at gmail.com
Tue Oct 2 19:08:11 UTC 2012


Is this a joke about the blackbox extension, or are you serious? I think
this is a splendid (I love this word) idea! I don't know if it would be of
use to you only or others like me, but I'm looking forward to it. This
could make my day job much easier as I happen to be the unofficial hg
support at my company, and problems like: "'I am sure I didn't do it, I
swear I didn't use mq' and then I scroll the console up and see he actually
did do it" are very common ;d

wujek

On Tue, Oct 2, 2012 at 9:00 PM, Matt Mackall <mpm at selenic.com> wrote:

> On Tue, 2012-10-02 at 18:16 +0000, Jensen, Aaron wrote:
> > > > @    19984 on default     <-- Merge that re-added 165 files
> > > > |\
> > > > | o    19983 on stable
> > > > | |\
> > > > | | o  19982 on stable   <-- designer's original change
> > > > | | |
> > > > +-+---o  19981 on default
> > > > | | |
> > > > | o |  19980 on stable
> > > > | |/
> > > > o |  19979 on default
> > > > | |
> > > > o |  19978 on default
> > > > |\|
> > > > | o  19977 on stable
> > > > | |
> > >
> > > What does this command say?
> >
> > >  hg log -r "heads(::19983 and ::19979)"
> >
> > It returns changeset 19977.
> >
> >
> > > What ancestor does merge --debug report?
> >
> >   searching for copies back to rev 15849
> >   unmatched files in local:
> >    ... snip ...
> >   unmatched files in other:
> >    ... snip ...
> >   all copies found (* = to merge, ! = divergent, % = renamed and
> deleted):
> >    ... snip ...
> >   checking for directory renames
> >    ... snip ...
> > resolving manifests
> >  overwrite: False, partial: False
> >  ancestor: 19977:19a9511ffc1a, local: 19979:1bec4d8209cc+, remote:
> 19983:8fae885db78c
> >    ... snip ...
> >
> >
> > > Can it be reproduced on the designer's machine?
> >
> > I am unable to reproduce.
>
> Then you're probably left with either a hardware glitch or undiagnosed
> user error[1].
>
> >   In talking with more users, I'm getting reports that this has happened
> before: merges between branches causing deleted files to reappear.
>
> There are some graph topologies where surprising things like this can
> happen; this case is not one of them. Generally speaking, you need
> criss-crossing or leap-frogging merges that make the merge ancestor
> choice ambiguous but here there's only one choice of ancestor.
>
> [1] Some day I'll finish the blackbox extension that records everything
> a user does for forensic purposes.
>
> --
> Mathematics is the supreme nostalgia of our time.
>
>
> _______________________________________________
> Mercurial mailing list
> Mercurial at selenic.com
> http://selenic.com/mailman/listinfo/mercurial
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20121002/8db7761c/attachment-0002.html>


More information about the Mercurial mailing list