IA64 port status update
faucon.millenium at gmail.com
Sat Feb 26 06:51:23 CST 2005
As the release preview is closing in, I think that some status update
on the state of the IA64 port might be of interest.
As you may know, the ia64 flavour of Hoary is available from the
archive, and is in a somehow decent shape. Specifically:
- State of the port (not considering universe/multiverse):
* As of today, we have 1 package FTBFS: syck
* We have 4 packages outdated:
* We have 32 packages uninstallable (mostly due to openoffice/language support
problems, see below)
So as you can see, the figures are pretty nice.
- Outstanding issues (not considering universe/multiverse):
As of today, we have two main blockers:
OO.o simply doesn't work on 64bit. A workaround would be to use the same
tweak as amd64: install OO.o with ia32-libs. That should work,
but I need help
to add lib32gcc1 support to ia32-libs.
As a side note, ia32-libs contains an ugly bug: #6923
Of course, a broken OO.o means language-support is uninstallable. Hence the
above mentioned 32 problems: these are somehow "dep wait" ones.
* Mozilla Firefox Locales:
This is a relatively new issue. Mozilla Firefox works just fine
until one tries to
install a locale support, in which case update-mozilla-chrome-registery
segfaults. I've been able to reproduce that bug with various
versions of Firefox
(Debian 1-2 and 1-6, Ubuntu 1-6), so I'm not quite sure it's a mozilla bug.
I suspect some side effect of not having language support
installed on ia64.
Any help in that area would be warmly welcomed.
- Kernel status:
The ia64 kernel is working just fine, as far as I can tell (at
least it's not worse than
anywhere else :)
- Installer status:
After quite a lot of debugging and tweaking, we now have a fully
Ubuntu installer, modulo some minor troubles when installing over
serial line (see
James' report about that).
As far as I know, all HP Itanium 2 machines are supported. Itanium 1
will fail during
X startup (missing Nvidia proprietary driver), but that's a
non-issue since Itanium
1 is not on our target list (these were sort of prototype machines anyway).
Special kudos to Colin and LaMont for their invaluable help in the
progress of the
port, thanks guys! 8-)
There you are. I think that the two main blocker issues would probably
be easily (or so I hope) solved if someone with enough insight in
ia32-libs and language support could give me some hints as how to fix
these, and then the port would look really good!
Ubuntu, Debian and Kernel Hacker
(1) Things will get worse before they get better.
(2) Who said things would get better?
More information about the ubuntu-devel