Re: Driver development vs X development
Egbert Eich <[email protected]>
| Newsgroups | gmane.comp.xfree86.forum |
|---|---|
| Message-ID | <[email protected]> |
There are plans to allow driver release independently from the X release. I'm going to change the location of the driver directories to implement this more easily. We also need to have an SDK just for driver development as we did with 3.x. Please not that this only addresses part of the problem: New features to the core server will not be available sooner this way. This needs to be addressed, too. Egbert. Billy Biggs writes: > I would like to see better handling of adding new drivers and > bugfixing existing drivers. It's unclear to me whether it would be > better to try and split driver and server development more, or release > more often, but I think many of the benefits of a modular server > architecture are lost when we see users not getting bugfixes for drivers > until the next major release. > > I was quite disturbed by this recent thread: > > http://www.mail-archive.com/devel%40xfree86.org/msg00846.html > > Also, there are many driver bugfixes and experimental driver > development that take a long time to get into the X distribution. If > there is an experimental branch of a driver, why not ship both? > > http://www.winischhofer.net/linuxsis630.shtml > http://gatos.sourceforge.net/ > > One final note about drivers, video projects (such as xine, mplayer, > avifile) are doing their own driver architectures, partly because of the > slow X development cycles (see vidix/*_vid/...). Maybe having driver > code more independent of X development would be enough to see some of > this advanced and experimental driver code be shared with code intended > for use inside the X server. > > -- > Billy Biggs > [email protected] > _______________________________________________ > Forum mailing list > [email protected] > http://XFree86.Org/mailman/listinfo/forum