Re: 'C-x 4 a' broken in diff buffer

[Nick Roberts]

 > >> > ISTR it also fails for changed functions (not just added ones).
 > >> I believe Per has fixed this very bug earlier today.
 > > Well it doesn't seem to supply a function with "-cp" any longer.
 > > Also I notice that when it does find a function e.g with "-c", it's often
 > > the wrong one, but I think this might always have been the case.
 > 
 > I'm sorry to tell that I have no idea what you're talking about.
 > Care to spell it out loud for me ?

Presumably you must have the gist of my first post to make your reply.  Per's
ChangeLog entry says:

        * diff-mode.el (diff-sanity-check-hunk): Don't reject the hunk
        just because the diff was produced using "-p" (--show-c-function).

Im my .emacs, I have 

(setq diff-switches "-cp")

so I'm guessing that that 'C-x 4 a' failing had something to do with the line
"-p" adds:

> *** gdb-ui.el 08 Feb 2007 10:06:25 +1300      1.203
> --- gdb-ui.el 08 Feb 2007 10:13:21 +1300      
> *************** With arg, use separate IO iff arg is pos
                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

which, as we've said before, is not reliable for languages other than C.

I find that, with the above value of diff-switches, 'C-x 4 a' fails to identify
a function to put in the ChangeLog, while it found one in Emacs 21.1.

-- 
Nick                                           http://www.inet.net.nz/~nickrob