possible problem in xf86_video_siliconmotion ?
chuckr at telenix.org
Wed Aug 27 17:59:50 PDT 2008
Paulo Cesar Pereira de Andrade wrote:
> Chuck Robey wrote:
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>> I had stopped compiling xorg for a week while I sealt with some other
>> items, and
>> in the meantime, it looks like the video driver
>> xf86-video-siliconmotion grew a
>> compile problem, at least here. The file smi_video.c has a problem
>> with line
>> 1226, the macros XF86_VERSION_CURRENT and XF86_VERSION_NUMERIC break,
>> the include of xf86Version.h is missing. Here, it installs into my
>> X11R7/inlcude/xorg. I added that include, it fixed the problem
>> If it's broken for you, that's what fixed it for me. If it builds for
>> you, then
>> disregard this. I just did a git-pull, it seemed that it wasn't fixed
>> when I
>> checked for this mail.
> I think it was not updated in the recent change to remove
> xf86Version.h. I have it corrected in my work tree, but anyone can
> can correct that in the mean time.
> I am working on a merge of the xorg driver with Alex Deucher's 501
> repository and code from smi. Still getting it to work, after that
> I plan to make separate commits (right now over 6k lines diffs)
Paulo, my inability to get git to work for me is making me unable to really say
whether problems I find are real, or just something that my git is gifting me
with. Now that I know it's wrong, do you know the URL for the bugs, so I can
mail in the tiny patch this needs? I assume that this list is very much the
wrong place to put it, I just wanted to verify it's not my git faking me out.
Another one I just found, in the xserver, the reason I couldn't get my options
to work is because I think I found a fairly big bug in one of the main option
functions. In hw/xfree86/common/xf86Option.c, function xf86CollectInputOptions,
the 3rd parameter is named extraOpts, but it's not used at all in the function
(seems to be an oversight), even though the comment advertises it. If you agree
it's busted, I'll make up a patch and send it (also to the URL I asked for
above). It's a pretty straightforward thing that probably breaks a huge number
of drivers, if git isn't laughing at me again.
Damn I wish you folks used cvs or svn ... I assume that unified diffs (my
favorite) are acceptable?
More information about the xorg