Re: rejecting patches that have an offset

From: Andreas Gruenbacher <andreas.gruenbacher-re5jqeeqqe8avxtiumwx3w@public.gmane.org>
Date: Aug 16, 2011, 22:48 UTC
Message-ID: <1313534889.5598.21.camel@schurl.linbit>
In-Reply-To: <4E49A8EA.5020507-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>

Eric,

On Mon, 2011-08-15 at 17:16 -0600, Eric Blake wrote:

> It would have saved me a lot of time if both 'patch' and 'git apply' 
> could be taught a mode of operation where they explicitly reject a patch 
> that cannot be applied without relying on an offset.

that sounds reasonable. Can you send a patch or at least add a bug on Savannah?

Show 7 quoted lines

> It might also be nice if patch could learn the algorithm that appears to 
> match the git behavior, where when there are multiple points with 
> identical context (viewing just the context in isolation), but where 
> those locations differ in function location (as learned by the @@ header 
> line in the patch file), then the preferred offset is the one in the 
> named function, even if that is not the closes context match to the line 
> number given in the patch file.

Sounds interesting; a patch for that would be great as well.

Thanks, Andreas