2012/6/14 Felipe Sateler <fsate...@gmail.com>:
> On Thu, Jun 14, 2012 at 6:08 PM, Dan S <danstowell+de...@gmail.com> wrote:
>> 2012/6/1 Felipe Sateler <fsate...@gmail.com>:
>>> On Fri, Jun 1, 2012 at 9:01 AM, Felipe Sateler <fsate...@gmail.com> wrote:
>>>>
>>>> On Jun 1, 2012 7:04 AM, "Dan S" <danstowell+de...@gmail.com> wrote:
>>>>>
>>>>> 2012/6/1 Jonas Smedegaard <d...@jones.dk>:
>>>>> > Hi Dan,
>>>>> >
>>>>> > On 12-06-01 at 08:56am, Dan S wrote:
>>>>> >> 2012/5/31 Felipe Sateler <fsate...@debian.org>:
>>>>> >> > I see that MathJax is included in the binary packages. Can't we use
>>>>> >> > libjs-mathjax?
>>>>> >>
>>>>> >> I don't really know mathjax myself but it looks like, technically, it
>>>>> >> could be patched to use libjs-mathjax, as long as our custom config
>>>>> >> file "HelpSource/MathJax/config/scmathjax.js" was installed in the
>>>>> >> appropriate place.
>>>>> >>
>>>>> >> However, the rendered help is supposed to be self-contained in the
>>>>> >> sense that it can be uploaded to a website (we do this with
>>>>> >> http://doc.sccode.org ) so if the URLs were rewritten to use
>>>>> >> /usr/share/javascript/mathjax/... then this would break. (And the help
>>>>> >> is supposed to be re-rendered by users - when the user installs an
>>>>> >> extension the rendering process is re-run locally.) So this makes me
>>>>> >> think that we would prefer not to. Is that OK with you?
>>>>
>>>> Another option is to create a linkfarm pointing to the real mathjax 
>>>> sources.
>>>> This would avoid the problem, wouldn't it?
>>>>
>>>> I think Raphael Hertzog wrote a tool precisely for this. I'm not on a
>>>> computer now, though, so I can't search for it.
>>>
>>> Found it
>>> http://packages.debian.org/unstable/dh-linktree
>>
>> Thanks - looks promising. I've edited the scripts as follows, but
>> dh-linktree is not being invoked. Anyone know the best way to get it
>> invoked?
>
> You seem to have mixed cdbs and short form dh...

Thanks. I've spent a lot of time this morning trying to get it
working. So far I've got this:

diff --git a/debian/control b/debian/control
index 278f220..e771be0 100644
--- a/debian/control
+++ b/debian/control
@@ -11,6 +11,8 @@ Uploaders: Alexandre Quessy <alexan...@quessy.net>,
 Build-Depends: cdbs (>= 0.4.73~),
  cmake,
  debhelper,
+ dh-linktree,
+ libjs-mathjax,
  emacs23 | emacsen,
  libasound2-dev [linux-any],
  libavahi-client-dev,
diff --git a/debian/rules b/debian/rules
index a015e2c..ccace6e 100755
--- a/debian/rules
+++ b/debian/rules
@@ -20,6 +20,9 @@ DEB_INSTALL_MANPAGES_supercollider-vim     =
debian/scvim.1 debian/sclangpipe_ap

 DEB_CMAKE_EXTRA_FLAGS = -DDSO_VISIBILITY=on -DSUPERNOVA=on
-DSC_EL_BYTECOMPILE=OFF

+install/supercollider-common::
+       dh_linktree -v --sourcedirectory=$(DEB_DH_INSTALL_SOURCEDIR)
-psupercollider-common
+
 install/supercollider-emacs::
        install -d $(DEB_DH_INSTALL_SOURCEDIR)/usr/share/applications
        install -m 644 debian/scel.desktop
$(DEB_DH_INSTALL_SOURCEDIR)/usr/share/applications
diff --git a/debian/supercollider-common.linktrees
b/debian/supercollider-common.linktrees
new file mode 100644
index 0000000..f126b43
--- /dev/null
+++ b/debian/supercollider-common.linktrees
@@ -0,0 +1 @@
+embed /usr/share/javascript/mathjax usr/share/SuperCollider/HelpSource/MathJax


The behaviour is:

dh_linktree -v -psupercollider-common
        install -d
debian/supercollider-common/usr/share/SuperCollider/HelpSource/MathJax
        rm -f 
debian/supercollider-common/usr/share/SuperCollider/HelpSource/MathJax/MathJax.js
        ln -sf ../../../javascript/mathjax/MathJax.js
debian/supercollider-common/usr/share/SuperCollider/HelpSource/MathJax/MathJax.js


but install fails because those relative symlinks are not right, and
creates dangling symlinks.
Any clues welcome, or working examples (there are very few in the package repo).

Dan

_______________________________________________
pkg-multimedia-maintainers mailing list
pkg-multimedia-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-multimedia-maintainers

Reply via email to