Re: Bug #593838: AX_CFLAGS_GCC_OPTION misuses AS_VAR_PUSHDEF variable

[Eric Blake]

On 10/01/2010 10:03 AM, Paolo Bonzini wrote:
> I think the correct answer is "do whatever it has been doing until
> 2.67". :) And the testsuite should be updated to have the appropriate
> coverage, and I want a pony.

I can't promise the pony :)  But given the arguments that have been 
raised on this list about regression fixes vs. development, I can 
promise this:

I'd like to alter development to allow multiple git branches, similar to 
how Ralf has been doing things on automake.  branch-2.68 will track 
regression fixes to be applied to 2.68; that way I can release 2.68.1 in 
the near future, if no other regressions are reported in the next week 
or so.  maint will track changes to be applied to the current branch-xxx 
release branch and master.  And we'll have to change the policy on 
master to allow it to accept merge commits (right now, if I remember 
correctly, savannah has a hook that rejects merge commits on master), so 
that master is always the merge of maint and latest development, and 
thus always a superset of branch-2.68.

That assumes, of course, that I can easily fix the regression of 
AS_PUSH_VAR accepting a var-name with = in it ;)  It may not be today, 
but it's certainly on my list of things to tackle during the first week 
of October.

--
Eric Blake   address@hidden    +1-801-349-2682
Libvirt virtualization library http://libvirt.org