Re: PNP patch into kernel when?

Philip Blundell (pjb27@cam.ac.uk)
Tue, 3 Dec 1996 17:50:02 +0000 (GMT)


On Tue, 3 Dec 1996, Ingo Molnar wrote:

> On Tue, 3 Dec 1996, Philip Blundell wrote:
>
> > What exactly is the benefit of renaming request_region() and so on? [...]
>
> couldnt this be renamed in one go, with a script? [if it's only a rename
> that happens there].

Yes, but it will cause needless heartache to people who try to maintain
compatibility with more than one version of the kernel (as is the case for
quite a lot of 3rd-party drivers). You'll end up with every driver having
something like

#if LINUX_VERSION < n
#define request_interrupt request_irq
#endif

and you will upset people who want modules to be binary compatible.

> If the new name saves 1 second of thinking time per new Linux hacker, then
> it is worth it. Just extrapolate to 1 million years or so. (the day when

I'm not sure either that the new name _will_ save any thinking time (is
"interrupt" really any less obvious than "irq"?) or that it will be
particularly worthwhile to save that amount of thinking time. If you want
to save new hackers time, your efforts are better directed into writing
documentation or something.

phil