Re: [buildd] Etch?
On Fri, 4 Aug 2006, Geert Uytterhoeven wrote:
> On Fri, 4 Aug 2006, Finn Thain wrote:
> > On Fri, 4 Aug 2006, Geert Uytterhoeven wrote:
> > > On Fri, 4 Aug 2006, Finn Thain wrote:
> > > > On Fri, 4 Aug 2006, Geert Uytterhoeven wrote:
> > > > >
> > > > > Since most of the problems are caused by compiler issues, what
> > > > > guarantees that a
> > > > > release-without-packages-that-caused-obvious-problems doesn't
> > > > > contain non-obvious problems caused by those same compiler
> > > > > issues?
> > > >
> > > > Was there ever a release with no bugs? No known bugs?
> > >
> > > Of course not.
> > >
> > > But an unreliable toolchain is a guarantee for more bugs.
> > I think I see what you meant: we might still meet the release criteria
> > for etch while there are still packages unbuilt due to toolchain bugs.
> > Now that _would_ be a problem for Gentoo ;-)
> Sorry if I didn't make it clear what I meant. I'll try to explain. An
> unreliable toolchain causes:
> 1. Compile failures -> obvious
> 2. Broken applications that are used a
> lot: basic tools (e.g. ls), system tools (e.g. ifconfig), things run
> during boot up (e.g. fsck), build tools (e.g. bison) -> probably
> easy to find they are broken
> 3. Broken applications that are rarely used -> non-obvious
> So while we will know about the packages in categories 1 and 2, there
> may be lots of silent breakage in category 3.
Yes, I know. And in my response, I was just trying make the point that
category 3 bugs may or may not be known, and the known ones may or may not
be "release critical".
My guess is, the number of unknown bugs is likely to be inversely
proportional to size of the user base.
> Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- firstname.lastname@example.org
> In personal conversations with technical people, I call myself a hacker. But
> when I'm talking to journalists I just say "programmer" or something like that.
> -- Linus Torvalds