[slime-devel] Re: Finding source of functions compiled with slime-compile-defun.

[Luke Gorrie]

Peter Seibel <peter at javamonkey.com> writes:

> Helmut Eller <e9626484 at stud3.tuwien.ac.at> writes:
>
>> Peter Seibel <peter at javamonkey.com> writes:
>>
>>> Again, having not looked into it in any great detail, is there any way
>>> to communicate source location back to Lisp so it can find the source
>>> of functions compiled with C-c C-c?
>>
>> excl::*source-pathname* seems to be the right place for Allegro CL.  I
>> committed some code to set this variable when compiling with C-c C-c.
>
> So I finally got around to checking out this change. Looks good except
> I still get warnings from Allegro about redefining a function when I
> use C-c C-c after, say C-c C-k. Poking around a bit it seems that the
> problem is the location gets set differently by COMPILE-FILE and your
> code. The patch below makes two changes--when the buffer we are
> compiling from is a file buffer we pass the full filename rather than
> just the buffer name and we take of the character position. It seems
> to work with Allegro 6.2 but I don't know if there's some other reason
> that the buffer passed to swank:compile-string-for-emacs should be
> just the buffer name. If not I'd like to commit this.

Currently excl::*source-pathname* is being used in a hacky way to
record the buffer and position that a definition was C-cC-c'd in, so
that M-. can later extract/parse this information by way of
find-fspec-location. Thus the "<buffername>:<position>" format is
significant because it gets parsed back.

So your patch would muffle the warning at the expense of breaking M-.
since it records only filename rather than buffer and position.

Seems to me the ideal would be to find a different way to record the
source location instead of encoding it into the "pathname" string.
Failing that perhaps another kludge is needed to muffle this warning..

-Luke