Building the GStreamer SDK offline

[Josh Doe]

Has anyone built the GStreamer SDK on an offline machine? I've just done it
on Windows, but it was very tedious since cerbero relies so much on cloning
repos and downloading tarballs. Here's a rough outline of what I had to in
case anyone else has to do this.

All the Windows dependencies were easy to install, though for mingw I had
to install it on an online machine then copy it over (making sure to
install msys-wget which isn't installed by default). Bootstrapping was a
bit of an annoyance, as I had to manually download the archives from
bootstrap/windows.py and either copy them so wget wouldn't try to download
them or extract it myself, commenting out the redundant bits in
bootstrap/windows.py.

For building recipes I modified commands/info.py to spit out the URLs for
any tarballs required by the recipes, then downloaded them on an online
machine with wget -i. I had to do some more minor modifications to use the
downloaded files. For git repos I just did a mirror clone of all gstreamer
and gstreamer-sdk repos. Thankfully most of the remotes can be updated just
by setting DEFAULT_GIT_ROOT in cerbero/config.py to the local mirror,
however for the 1.0 recipes I had to do a find and replace since they
directly use the upstream URL.

The biggest annoyance was that every GStreamer repo references the common
repo by it's upstream URL, so I had to fix that for every recipe by
running git config submodule.common.url to point to the local mirror.

I'll probably end up doing this again on another machine, so wanted to see
if anyone else has tried this and has a better way. If not I might just
patch cerbero to spit out a list of all zip/tarball/git URLs, and to fix
the other issues I mentioned above.

-Josh
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/gstreamer-devel/attachments/20130501/d8d93106/attachment.html>