configure checks for presence of "source-highlight", "highlight" and "vim", and in case one of them is found its path is embedded in /usr/bin/gtkdoc-fixxref. From build.log: checking for source-highlight... no checking for highlight... no checking for vim... /usr/bin/vim Therefore after "emerge vim; emerge gtk-doc; emerge -C vim" gtkdoc-fixxref will fail. Please note that this is _not_ a dupe of bug 333313, as it happens with stable vim (7.2.442) too.
I would suggest you to report this to upstream also if possible. Thanks :-)
(In reply to comment #1) > I would suggest you to report this to upstream also if possible. Thanks :-) I can do of course. A hint from the Gnome team about what would be your preferred solution would be helpful, though.
something like "--with-highlight=(none|vim|source-highlight|highlight|auto)" would be good
For the record: A related problem was faced in bug 306569
*** Bug 338366 has been marked as a duplicate of this bug. ***
Is there any progress in solving this bug? After removing glitz and if the doc USE flag ist, several packages e.g. glib, libbononoboui, etc fail to reemerge.
media-gfx/gtkimageview-1.6.4 also fails to emerge because of missing /usr/bin/vim.
don't bother reporting all failures, we know that any package using gtk-doc might fail due to this bug. What we need is a patch.
Attached a tentative patch to the upstream bug. Let's see what they say about it.
I'm about to commit the necessary changes to the ebuild, but I have a question I'd like to get feedback on, which highlighter should we pick. I'm opposed to exposing all of them through use flags or doing guess work in the ebuild. I wanted to go with source-highlight but it doesn't have sufficient keywording yet. Should we go with vim instead ?
(In reply to comment #10) > I'm about to commit the necessary changes to the ebuild, but I have a question > I'd like to get feedback on, which highlighter should we pick. I'm opposed to > exposing all of them through use flags or doing guess work in the ebuild. I > wanted to go with source-highlight but it doesn't have sufficient keywording > yet. Should we go with vim instead ? > I also don't think having USE flags for each highlighter looks nice, but it would have the advantage of letting people (like me) to prevent installation of a new package (dev-util/source-highlight) as I will have app-editors/vim installed anyway :-/
(In reply to comment #10) > I'm about to commit the necessary changes to the ebuild, but I have a > question I'd like to get feedback on, which highlighter should we pick. > I'm opposed to exposing all of them through use flags or doing guess work > in the ebuild. I wanted to go with source-highlight but it doesn't have > sufficient keywording yet. Should we go with vim instead ? The upstream package can work with either, so why don't you want to make it controllable by USE flags? An unconditional dependency on vim doesn't look right to me. (And it would also have the strange consequence that emacs would depend on vim, via gtk+ and gtk-doc. Last time we did such a thing we had an outcry from users ...) Besides, isn't source-highlight the tool best suited for the job?
A use flag is needed to make sure you don't have to do guessing work in the ebuild to find what highlighter to select and I'm sure we'll get a couple of users complaining about highlighting dependency forced if we go with that.
(In reply to comment #13) > A use flag is needed to make sure you don't have to do guessing work in the > ebuild to find what highlighter to select and I'm sure we'll get a couple of > users complaining about highlighting dependency forced if we go with that. Yeah, I'm all for USE flags here. (Reading my previous comment #12 again, I see that I didn't express it very clearly.)
+*gtk-doc-1.15-r2 (14 Nov 2010) + + 14 Nov 2010; Gilles Dartiguelongue <eva@gentoo.org> + -files/gtk-doc-1.10-no-m4.patch, -gtk-doc-1.11.ebuild, + -files/gtk-doc-1.11-fix-index-id-gen.patch, + -files/gtk-doc-1.11-quote-filenames-with-space.patch, + -files/gtk-doc-1.11-tests-fixes.patch, -gtk-doc-1.13-r2.ebuild, + -files/gtk-doc-1.13-fixxref-compat.patch, + -files/gtk-doc-1.13-scan-touch-module-overrides.patch, + +gtk-doc-1.15-r2.ebuild, + +files/gtk-doc-1.15-allow-selection-highlighter.patch, metadata.xml: + Fix automagic highlighting program detection, bug #334489. Handle + multi-slot python. Clean up old revisions. Thanks for reporting.
