[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: XAA bug in 4.x

PS: Now I have more time to answer.

Branden Robinson wrote:
I personally need a compelling reason to grab stuff from HEAD.  Myself,
I generally only backport bugfixes (some XFree86 committers don't bother
to do this for a stable release branch).

Well. Debian's XFree86 4.3 release is almost 10 months behind other distributions, giving the developers (including humble me) 10 months time for fixing bugs which were discovered after the mass-distros like Mandrake and RedHat threw it out. Don't you think this is a compelling reason?

(Please don't consider this statement any kind of offence. I _really_ appreciate your work for Debian.)

It appears, though, that DanielS is not as conservative.  I guess we'll
see how that works out in this case; in general it's a pretty bad
strategy, as we've seen occasionally.

I can only speak for the SiS driver. This driver has been very stable for a long time. Just count the posts on the XFree mailing lists regarding this driver. You'll hardly see 15 since Jan 2003 (the majority of which were installation problems). Or look at Debian's BTS. The only bug reports for SiS stuff were like "SIS 740 not supported" and such. (Don't get the idea that SiS hardware is seldom; the driver has been downloaded from my website more than 600.000 times since 01/01/2003)

The reason for changing anything (despite the stability) relates to the fact that vendors customize SiS chips and this leads to incompatibilities. Since Jan 2003 (the release of Xfree 4.3), I have included support for such customized systems in not less than 20 cases. Furthermore, companies still produce new machines with SiS hardware, with new types of LCD panels. These sometimes require adaptions as well. 10 months is a long time. Many new machines have been thrown on the market since then.

The reason for pushing the most current driver into Debian's XFree is simply the fact that a lot of people will rely on a "3rd party driver" (being the one from my website) otherwise, because their LCD panel - for example- just turns black...

The other changes since June (which was when Daniel pulled the driver from CVS HEAD the last time), include no major (ie dangerous) things. (Fixes for TV, TV scaling, SiSCtrl interface, improved monitor detection). I myself use the driver on 4 different machines 24/7. One could say it gets enough testing, I guess.


Thomas Winischhofer
thomas AT winischhofer DOT net          http://www.winischhofer.net/
twini AT xfree86 DOT org

Reply to: