make force-tag gone

Jon Ciesla limb at jcomserv.net
Tue Sep 9 15:36:18 UTC 2008


> On Tue, 9 Sep 2008, Jon Ciesla wrote:
>
>>
>> > Mike McGrath wrote:
>> >
>> >
>> >> In fairness... you should be testing things before you commit them
>> >
>> > Of course.  I've been hit by this a lot with local builds working, but
>> > rawhide builds that fail (for whatever reason, new rpm, wierd build
>> deps
>> > breaking, etc...).
>>
>> I have issues, also, where something succeeds in a local build, is fine
>> oin i386 mock, but dies in ppc, x86_64, etc.
>>
>
> So why would you make a change to the spec file, without bumping the
> release?  Also there's an auditing GPL legal reason (IIRC) that we're
> doing this now.  The bottom line is this:
>
> Make change to a spec file.
>
> Bump release.

Even it there's no successful build?  Is the cvs log not retained, and
useful for auditing purposes?

> Its a simple workflow that provides an audit trail that we believe will
> keep us in compliance with the GPL.  force-tag is sort of nice to have I
> guess, but release bumps happen all the time by everyone its not a high
> barrier to get releases out.
>
> I'm not sure what else to say, its not going to happen guys unless you can
> come up with a different audit trail that will keep us in compliance with
> the GPL and satisfies legal.

So is using TAG_OPTS=-F make tag a problem?

> 	-Mike
>


-- 
novus ordo absurdum




More information about the devel mailing list