Re: [TIMINGS] Re: 2.1.xxx makes Electric Fence 22x slower

Gregory Maxwell (linker@z.ml.org)
Tue, 8 Sep 1998 16:14:36 -0400 (EDT)


On Tue, 8 Sep 1998, Linus Torvalds wrote:

> On Tue, 8 Sep 1998, Alex Buell wrote:
> > On Mon, 7 Sep 1998, Linus Torvalds wrote:
> >
> > > David, before you get any more hung up about fuzzy hashing, let's just
> > > make one thing clear:
> > >
> > > Code Freeze
> >
> > So why did you add the QNX filesystem then?
>
> Because it doesn't impact anything else if it is broken. You will find no
> downside to it, because if you don't need it you don't configure it in.
> And if you need it, whether it is broken on non-existent it doesn't work,
> while putting it in it _may_ work well enough for you.
>
> In constrast, things like the VMA management impacts _everybody_.
>
> Linus

Dont bother, they dont get it.

You need to understand what a feature freeze is for: It's to stop the
influx of bugs so you can fix the ones that are already there.

Linus mostly obeys this spirt. Things linke QNXfs dont impact anything if
they are turned off, so they dont mask the existing bugs.

Since it's marked expermental, off by default, and most people turn it on,
it's as though it were a seperate patch in terms of administrative
overhead.

To achieve a true feature freeze you'd have to prevent people from
patching their kernels at all. What? You say that that is okay because
they can mention what patches they have on? Fine. Then I say that they can
mention they have QNXfs or other expirmental features turned on.

Feature freeze means that things like fuzzy hash list dont go in, unless
there is a bug that can't be fixed by a very simple change.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/faq.html