Subject: Re: KNF for i386 pmap new?
To: None <tls@rek.tjls.com>
From: Mike Hibler <mike@fast.cs.utah.edu>
List: tech-kern
Date: 05/05/1999 14:05:43
> Date: Wed, 5 May 1999 13:13:56 -0400
> From: Thor Lancelot Simon <tls@rek.tjls.com>
> To: Gandhi woulda smacked you <greywolf@starwolf.com>
> Subject: Re: KNF for i386 pmap new?
> Message-ID: <19990505131355.A3905@rek.tjls.com>
>
> On Wed, May 05, 1999 at 12:21:16AM -0700, Gandhi woulda smacked you wrote:
> > On Wed, 5 May 1999, matthew green wrote:
> >
> > #
> > # Is there a reason why the i386 version of pmap-new is not KNF'd? I just
> > # had a look at Chuck's resident count changes and noticed this...
> > #
> > #
> > # chuck doesn't like KNF.
> >
> > My reply was going to be a lot more rude but it's late so suffice to say
> > that rule #1 of porting or hacking applies here:
> >
> > "If you're gonna play in the big sandbox, it's generally considered bad
> > form to kick down someone else's sandcastle to build your own."
> >
> > In short, if you're coding within an existing project, try to keep with
> > style, even if it repulses you.
>
> You have clearly never closely examined the original Hibler VM code.
>
Not entirely sure what you are saying here, but I'll reply anyway :-)
I am a KNF kinda guy and didn't like a lot of aspects of the original Mach
VM code that I was working with. However, I do believe in Rule #1 and tried
to adhere to it but perhaps not at the granularity you would like. When I
was working in files that derived from Mach, I tried to use their coding
conventions. When I was developing new stuff (like the 68k pmap module,
or the swap pager or whatever) I tended toward KNF (though I do deviate in
some aspects). I did not treat the "VM system" as a whole and use Mach
style throughout (yuk) or attempt to convert all the code to KNF.
That said, maybe I just plain failed anyway. If so, sorry! :-)
That code is all gone now anyway isn't it? (sniff, sniff...)