On Fri, 2011-11-11 at 09:40 +1100, Dmitry Smirnov wrote: > >> This is to substitute built-in r8169 driver if it doesn't work well. > > > > Please don't. If this gets added to Debian then every time there is a > > bug in in-tree r8169 then users will be told 'oh, try r8168'. And if > > the bug is not present in r8168 then they may not report the bug at all, > > and it will take longer to fix. > > Nobody argue that r8169 should be improved. > You've made a good point that including r8168 might delay fixing r8169. > > On the other hand without alternative driver one might suspect faulty > hardware, in which case time will be wasted and diagnosis will be > delayed. > When reporting bugs people should be able to compare and say "hey, > this works and that doesn't". They can do that now, as we don't stop them from installing the version Realtek distributes. > While fixing drivers is very important, as far as I understand, Debian > role is to provide usable system to people. Yes. A usable *system*. > I'm surprised how easily you prepared to refuse the remedy in favour > of (rather obscure) benefit to driver development. > > Having alternative driver breaks nothing and put nothing at risk. Not directly, no. > One of the things I like about Debian is that we have plenty of > alternatives so for example nobody tell we should kick all web servers > but Apache. Users don't want alternative drivers; they just want a working driver installed automatically. The answer to 'this package is buggy' should not be 'never mind, try another one'. > Having another driver available for our users is not against our traditions. > > > The major reason for out-of-tree driver updates is to support new > > hardware on older kernel versions, but we already do that through > > backports. > > Right, but in this case we can't backport so at the moment for some > unfortunate owners of Realtek cards alternative driver is the only > option. [...] I just *did* backport r8169 from 3.0 into squeeze. Ben. -- Ben Hutchings The program is absolutely right; therefore, the computer must be wrong.
Attachment:
signature.asc
Description: This is a digitally signed message part