[Quilt-dev] Re: [PATCH] `quilt header -a' corrupts diffstat-carrying headers

[Jean Delvare]

Le mardi 26 janvier 2010 04:53, Don Mullis a écrit :
> Jean Delvare <address@hidden> writes:
> > I never use "quilt header -a" myself, and I have to
> > admit I am curious what others may be using it for. It doesn't seem
> > very useful to me.
> 
> What I'd like to use it for is batched appending to a patch series,
> e.g. for Linux kernel:
> 
>   while quilt top
>   do
>     echo "Signed-off-by: address@hidden" | quilt header -a
>     quilt pop
>   done

Blindly signing patches is bad ;)

The above could be easily handled by a custom script. You don't have
to rely on quilt for everything.

> > The only behavioral change I would consider is the following: if
> > a diffstat section is present _and_ it is preceded by "---" _and_ it
> > is at the end of the header comment, then "quilt header -a" could be
> > made to insert the additional comment right before the "---". But
> > even that might go against the expectations of some users.
> 
> How about leaving the behavior of -a as it always was, and creating a
> '-i' option that would have the functionality you describe?  That way,
> there would be no chance of violating existing expectations.  I would be
> willing to prepare the patch.

I'm not a big fan of adding new options for very specific needs. I
wouldn't even have taken -a in the first place. I think that custom
header edition needs are better handled by "quilt header" redirected
to a temporary file, custom edition of that file, and lastly "quilt
header -r" from that file.

The problem with your "-i" option is that it would insert right before
the diffstat section (if there is such a section.) But others might
have other insertion needs, for example at the top of the header (to
insert a "From:" line), or before the start of the body (to add a
"Subject:" line), or at the end of the body, or between "---" and the
diffstat (hidden comment.) And I'm only thinking of the header format
I use for Linux kernel patches. There must be users out there with
completely different expectations. I really don't think we want to
have a new insert option for every possible case.

-- 
Jean Delvare
Suse L3