Re: [Quilt-dev] [PATCH] Add option to use 'wiggle' to try to apply rejected patches.

[Chris Mason]

On Thursday 20 January 2005 07:43, Andreas Gruenbacher wrote:
> Hello,
>
> On Thursday 20 January 2005 11:25, Chris Larson wrote:
> > Greetings,
> >
> > Here's a small patch that I found useful.  It adds a commandline option
> > '-w', that when specified, calls out the 'wiggle' command to attempt to
> > apply the rejects.  Wiggle can apply patches that patch cannot, and
> > places conflicts in the file, not unlike the way SCM tools do, which is
> > more intuitive for those not very experienced at patching tools.
>
> Unified reject files somewhat improve the readability of reject files for
> me, but I agree that merging rejects still is a pain. I'm not happy with
> wiggle though: It's buggy; I don't trust it, and I wouldn't recommend it to
> anyone.
>
> Chris Mason has been working on a "better wiggle" that seems to be in
> usable state already. I'm not sure what the best way is to integrate those
> things though. Right now I tend to prefer something like a separate wiggle
> command that goes through the top patches' reject files. On the other hand
> it may be better to integrate this into the push command; I really don't
> know yet.
>

Wiggle and rej both make it easier to merge rejects, but it is still always a 
manual process.  The tools just get the reject hunks close to where they 
should be in the file, from there you still need to at least verify the 
results.

You always need to open up the reject and look at it to make sure the merge 
went properly.  So, I prefer and interactive command that runs rej, opens the 
reject for you, and deletes the foo.c.rej file when done.

I put my perl hack for this in with the current version of rej (v 0.11), it is 
named qp, which is just a wrapper around quilt push.  qp is mostly 
undocumented, but I can clean it up if people find it interesting.

rej is usable but needs some fine tuning for cases where there is really no 
good place to merge things.  If anyone finds rejects where it does completely 
the wrong thing, please send me test cases.

Releases are at:
ftp://ftp.suse.com/pub/people/mason/rej/

-chris