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

[Nick Roberts]

Stefan Monnier writes:
 > >> Hmm... are you sure it worked in Emacs-21?
 > > Yes.
 > 
 > Sure, sure, sure, in the very same circumstances?

You seem sceptical.

 > >> There has been changes in this part of the code, but IIRC they're
 > >> fairly minor.  And in any case we've never made use of the text added by 
 > >> the
 > >> "-p" flag.
 > 
 > > No, but maybe the text is inadvertantly used in trying to match up the
 > > function.
 > 
 > That would be a bug, but of course, it could be.
 > 
 > >> If you can find a case that works in Emacs-21 and fails now, it's probably
 > >> a bug, but then you'll need to give us a precise recipe.
 > 
 > > I've already given one.
 > 
 > > Evaluate (setq diff-switches "-cp")
 > 
 > diff-switches is not used by diff-mode, so this should make no difference.

It makes a difference to what gets inserted into the buffer.

 > > Put the diff that I sent earlier in a file temp.diff and 'C-x 4 a" behaves
 > > as I explained.
 > 
 > C-x 4 a will first look for the corresponding source file and then run C-x
 > 4 a there, so the result depends on the gdb-ui.el file you happen to have in
 > the same directory.

Maybe this recipe's no good but you could create your own change.

 > PS: There was still a bug that I just fixed, which may explain what you
 >     were seeing, although in my case it manifested itself by signalling an
 >     error rather than silently ignoring the function name.  Could be due to
 >     local changes.

I'm getting thoroughly confused.  Forget this report for the moment.  I'll
come back again when I have something concrete.

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