GNU bug report logs -
#7167
23.2; w32-shell-execute doc
Previous Next
Reported by: Mirko Vukovic <mirko.vukovic <at> gmail.com>
Date: Wed, 6 Oct 2010 16:09:01 UTC
Severity: normal
Found in version 23.2
Done: Eli Zaretskii <eliz <at> gnu.org>
Bug is archived. No further changes may be made.
Full log
Message #58 received at 7167-done <at> debbugs.gnu.org (full text, mbox):
> Date: Thu, 07 Oct 2010 20:12:47 +0200
> From: Eli Zaretskii <eliz <at> gnu.org>
> Cc: cschol2112 <at> googlemail.com, sean.sieger <at> gmail.com, 7167 <at> debbugs.gnu.org
>
> > From: Juanma Barranquero <lekktu <at> gmail.com>
> > Date: Thu, 7 Oct 2010 12:15:55 +0200
> > Cc: Sean Sieger <sean.sieger <at> gmail.com>, 7167 <at> debbugs.gnu.org
> >
> > [Sean, I've Cc:ed you because you're doing binary tarballs now]
> >
> > On Thu, Oct 7, 2010 at 05:57, Christoph <cschol2112 <at> googlemail.com> wrote:
> >
> > > Juanma, let me know if I can help troubleshoot this.
> >
> > I think it is an issue with the binary distribution of Emacs 23.2 for
> > Windows, not with 23.2 per se, because I have my own build of it and
> > the problem is not present.
> >
> > Likely the fix will simply be to rebuild the official binary
> > distribution of 23.2.
>
> Not sure, because my binary, which I built myself, also has this
> problem.
>
> Will try to investigate tomorrow.
Found the problem. There's nothing wrong with the Windows build
system per se. The problem is that emacs-23.2 tarball comes with a
src/buildobj.h from a Unix system where it was tarred, which is
baaaaad, and not only on Windows. This causes Make not to create
src/buildobj.h as suitable for the actual build on the target
platform, and the rest is history.
I've just installed a fix in the emacs-23 branch (revno 100091 and a
followup change in 100092) to exclude this file from the tarball.
To fix this locally, just remove src/buildobj.h and rebuild Emacs.
While at that, I would suggest that this code in help-fns.el:
(let ((file (catch 'loop
(while t
(let ((pnt (search-forward (concat " " name "\n"))))
(re-search-backward " S\\(.*\\)")
(let ((file (match-string 1)))
(if (member file build-files)
(throw 'loop file)
(goto-char pnt))))))))
(if (string-match "^ns.*\\(\\.o\\|obj\\)\\'" file)
(setq file (replace-match ".m" t t file 1))
(if (string-match "\\.\\(o\\|obj\\)\\'" file)
(setq file (replace-match ".c" t t file))))
(if (string-match "\\.\\(c\\|m\\)\\'" file)
(concat "src/" file)
file)))))
be made smarter wrt the error message it displays when the function or
variable it looks for is found in etc/DOC, but the file in which it is
defined is not in build-files. (This happens when you click or type
RET on the link to the source file where the function is defined.)
Currently, the error message, which comes from search-forward, is
quite cryptic for a naive user:
Search failed: "^_Fw32-shell-execute
"
It should at least mention the fact that the problem could be with
build-files, or, better, say that w32-shell-execute's definition is in
a file that is not in build-files. Volunteers are welcome to make
this improvement.
This bug report was last modified 14 years and 287 days ago.
Previous Next
GNU bug tracking system
Copyright (C) 1999 Darren O. Benham,
1997,2003 nCipher Corporation Ltd,
1994-97 Ian Jackson.