Re: Bloat? (khttpd)

[Martin Dalecki]

"H. Peter Anvin" wrote:
> 
> Alexander Viro wrote:
> >
> > On Thu, 23 Dec 1999, H. Peter Anvin wrote:
> >
> > > Martin Dalecki wrote:
> > > > >
> > > > > Performance, and *modularity*.  khttpd doesn't affect anything else in
> > > > > the kernel, and so is a rather harmless optional feature.
> > > >
> > > > There where times where for example /proc was an option too...
> > > >
> > >
> > > /proc was never modular.  It touches *everything*, which makes it
> >
> > No, now it doesn't. It used to, and that made it a whole friggin' pain in
> > ass whenever we did anything near VFS. Right now _nothing_ outside of
> > fs/proc should care about procfs internals. There are 4 or 5 places that
> > still do and IMO each of such cases is a bug.
> 
> True for the filesystem.  Not true for the data sources!!

And I still remember well the whole process of continual breakge of
top and ps after any unevitable change to this stuff. Thereafter it
seems 
anybody  just agreed on: "Don't tuch this heap of sh*t it will break
DAHAX*!".
Instead of fixing it...

*) DAHAX - Dumb Ad Hoc Appliaction X.

--Marcin