Bug 501294 – dev-ml/utop-1.11 - src/top/utop.el:801:1:Error: Symbol's function definition is void: utop-choose

• Status: RESOLVED TEST-REQUEST
• Product: Gentoo Linux

[Description José Romildo Malaquias 2014-02-14 13:12:45 UTC]

When compiling utop-1.11 on my ~amd64 system, I am getting the error:

 * Compiling GNU Emacs Elisp files ...

In utop-choose:
src/top/utop.el:748:5:Warning: reference to free variable `error'
src/top/utop.el:801:1:Error: Symbol's function definition is void: utop-choose


Reproducible: Always

[Comment 1 José Romildo Malaquias 2014-02-14 13:13:16 UTC]

Created attachment 370394 [details]
emerge --info

[Comment 2 Jeroen Roovers (RETIRED) 2014-02-14 13:55:21 UTC]

Please attach the entire build log to this bug report.

[Comment 3 José Romildo Malaquias 2014-02-14 17:10:12 UTC]

Created attachment 370408 [details]
dev-ml:utop-1.11:20140214-130508.log

[Comment 4 Alexis Ballier 2014-12-03 16:24:41 UTC]

I have no clue about emacs stuff, if emacs team is willing to fix it, be very welcome because I can't :/

[Comment 5 Ulrich Müller 2014-12-03 17:34:47 UTC]

Has Emacs support ever worked for this package? There is this line in src_install():

    elisp-site-file-install "${FILESDIR}/${SITEFILE}" || die

but no ${FILESDIR} exists for this package.


As for the byte-compile error:
- After fixing the malformed condition clause in line 748 I get another error:
     In toplevel form:
     utop.el:801:1:Error: Symbol's function definition is void: utop-choose
- After fixing that too (by moving function utop-choose into an
  eval-and-compile form) it turns out that it needs some unspecified
  dependencies (tuareg-mode?).
- There is also a warning about function replace-in-string (called in
  line 344) being undefined.


@José: Is the goal to have working Emacs support, or just to make the package compile? If the latter is the case, I suggest to remove Emacs support from this package. And the problem should be reported upstream.

[Comment 6 José Romildo Malaquias 2014-12-03 19:49:38 UTC]

> @José: Is the goal to have working Emacs support, or just to make the
> package compile? If the latter is the case, I suggest to remove Emacs
> support from this package. And the problem should be reported upstream.

The goal is to have working Emacs support, I think.

[Comment 7 Ulrich Müller 2014-12-04 09:09:01 UTC]

Created attachment 390902 [details, diff]
utop-1.11-emacs-bytecomp.patch

This is what I have so far. It doesn't fix the warning about replace-in-string being undefined, though. Really this bug should be fixed upstream.

[Comment 8 Alexis Ballier 2014-12-04 09:44:14 UTC]

(In reply to Ulrich Müller from comment #7)
> Created attachment 390902 [details, diff] [details, diff]
> utop-1.11-emacs-bytecomp.patch
> 
> This is what I have so far. It doesn't fix the warning about
> replace-in-string being undefined, though. Really this bug should be fixed
> upstream.

You have free hands here, if you feel it's better to drop emacs support or to fix it. Please commit what you think is better.

Thanks a lot for the help.

[Comment 9 Ulrich Müller 2014-12-04 14:32:46 UTC]

*** Bug 453732 has been marked as a duplicate of this bug. ***

[Comment 10 Alexis Ballier 2015-06-29 09:23:57 UTC]

utop 1.18 has some improvements on emacs side, please test