On Sun, Mar 31, 2002 at 09:11:27PM -0500, Pete Zaitcev wrote:
> > Why not?  What's wrong with 2.5?  In fact, that's where all new work
> > should be going on right now, not the 2.4 tree.
> 
> The pressure to stay with stable kernels is simply too strong.
> If you develop for unstable series, you must track a quickly
> changing codebase, and this saps the time and effort which could
> be used productively. Not to mention data integrity concerns.
> 
> A number of projects, small and big, were developed on stable
> series. For instance, s390 was merged into 2.2 first. I did
> ymfpci for 2.2 first. I do all current development on 2.4.

Both of those are separate projects.  As USB is now in three different
kernel trees (at three different levels of functionality), we can't just
ignore the development kernel branch and do our development in 2.4
without hurting a lot of different people in the process (assuming you
mean to push the changes back to the main kernel all along the way.)
Sad to say, but people rely on their USB devices to work properly (and
too many companies are making machines right now with no PS2 or serial
ports on them...)

Hey, 2.5 is the development tree, and 2.4 and 2.2 are stable trees.
Which one just sounds like the proper one for new USB development?  :)

> I see a problem with that, but I do not see a solution.
> Greg does well to resist, I suppose. Perhaps, UML might
> be helpful.

UML looks very good for core kernel work, I don't think the virtual USB
driver that someone wrote for it works anymore, so doing driver work
might be difficult.

thanks,

greg k-h

_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

Reply via email to