quilt for patch management?

[Andrew John Hughes]

2009/5/13 Robert Schuster <thebohemian at gmx.net>:
> Hi Andrew,
>
> Andrew Haley schrieb:
>>
>> OK.  Why is it better to do
>>
>> patch -p0 < <patch>
>>
>> or
>>
>> quilt import -p0 <patch>
> You do the import once and after that quilt maintains the list of
> applied and not-applied patches. This allows going back and forth
> between the various patches conveniently (push/pop).
>
>> I don't get it.  Seriously, I'm baffled: how would it make any difference?
> Updating patches could be done in a well described way.
>
>>> Furthermore how do you update a patch at the moment? Manually as well?
>>> quilt has a refresh command for this. It basically works like a poor man
>>> SCM sandbox.
>> Updating patches is probably a bad thing.  It makes more sense to apply
>> the patch and your changes to pristine sources and regenerate.
> Sure this is what it would be. But how is this done at the moment. What
> steps are needed to do that on the command line? As I see this stuff
> needs manual work.
>

I guess I (and it seems a few others) just don't grok how this would
help and not introduce additional burden.

I can see how it might be helpful if I need to stack up patches to
apply to IcedTea as a whole.  There's probably a case for Mercurial
queues there in maintaining my own stack of pending patches I've sent
to the list.  I'm not sure if push and pop would be enough to manage
it though.  I may have three patches on the stack for instance and 1
and 3 get approved, but 2 is still waiting -- is this workable?

For IcedTea generally, I just don't get it.  The patches are applied
at build time by users, not by developers.  Any stack would somehow
have to be available in every checkout, including tarballs, and we'd
have to add an additional dependency on quilt.  That only seems to
hurt.

> But finally the initial mail was more of a: "Hey, I know this little
> tool to make work easier. We use it in OE in the same context. Should I
> think about implementing it or not?"
>
> As it looks people don't want it. So I let it be.
>

It's worth bringing it up again, as this has been mentioned a few
times in various forms - if not here, then on IRC.  It really needs
someone who uses it to explain why it's better than sliced bread
though... :)

> Regards
> Robert
>
>

-- 
Andrew :-)

Free Java Software Engineer
Red Hat, Inc. (http://www.redhat.com)

Support Free Java!
Contribute to GNU Classpath and the OpenJDK
http://www.gnu.org/software/classpath
http://openjdk.java.net

PGP Key: 94EFD9D8 (http://subkeys.pgp.net)
Fingerprint: F8EF F1EA 401E 2E60 15FA  7927 142C 2591 94EF D9D8