RE: [Gnu-arch-users] tla star-merge in cycle, how to work around?

[David Allouche]

On Fri, 2005-03-25 at 18:44 +0300, Попков Александр wrote:
> On Friday, March 25, 2005 6:19 PM, David Allouche wrote:
> >> Only the one problem I see now. For example next situation:
> >> 1) Developer merges changes from local to main archive;
> >>    (and arch generate the new patch-log)
> >> 2) All another developers don't change this archive;
> >> 3) At night cron run job for automerge, and this patch-log
> >>    (as a new changeset) will be merged and commited to
> >>    local archive.
> >> 4) Local archive again has changes (last commited patch-log)
> >>    and developer go to point 1)
> >> 
> >> Is exist right method to handle this situation?
> >> Or may be I am use broken ARCH usage schema?
> 
> >You should make your automatic merge script smarter so it detects when
> >the changes applied by the merge are only patchlog additions, and not
> >commit such merges since they would be mere noise.
> 
> Advice me please which command I can use for true detect this?

There is no built-in command that can trivially and reliably provides
you with this information. You would have to do some non-trivial
scripting.

I recommend that you look into pyarch (that I wrote, but is currently
unmaintained as I'm focusing on pybaz) or pybaz, that provide you with
Python bindings or the Arch-Magic bindings (by M. Goikham) for Perl.

-- 
                                                            -- ddaa

signature.asc
Description: This is a digitally signed message part