[asterisk-dev] Confused with release numbers, how to manage incremental patching ?

[Russell Bryant]

Pavel Troller wrote:
>   Sorry for reopening this stupid thread, but WHY THE HELL the patch for
> asterisk-1.6.1.10 is based on 1.6.1.5 ?
>   What should one do, when he carefully incorporates every patch into his
> source tree (full of his own patches, so it's not easy to simply throw out
> and start with a fresh download), he's now standing at 1.6.1.10-rc3 and
> boom, a patch is based on a tree 5 releases back ???
>   Is there a good soul capable of creating a patch from 1.6.1.10-rc3 to
> 1.6.1.10-final ? Similarly, as absolutely logical series of Linux kernel
> patches (in the "incr" subdirectory of the main kernel directory)... This
> policy makes asterisk patch files almost useless...

I agree that the patches are not ideal.  This is effectively the result
of our release script morphing over time.  Our release process is much
more complex than when the script was first written.

The script takes an argument of the "last" release and does a number of
things with that.  One of which is to figure out how to generate the
patch file.  However, it is also used to generate the ChangeLog and
release summary documents.  It needs a list of changes in the branch
since the specified last release.  You can't use the security releases
for that purpose since you'll miss some of the bug fix changes that were
made in the branch.  It sounds like the patch generation needs to be
decoupled from the ChangeLog and release summary generation.

So, yes, the release script needs some love.  If you'd like, you can
file an issue on mantis for it.

Also, for what it's worth, the script is open source in case anyone
wanted to hack on it.  It's "mkrelease" in the repotools repository.

-- 
Russell Bryant
Digium, Inc. | Engineering Manager, Open Source Software
445 Jan Davis Drive NW - Huntsville, AL 35806 - USA
Check us out at: www.digium.com & www.asterisk.org