Perception of LFS
no.spam at allowed.here
Sun Jan 12 18:41:19 PST 2003
On Sun, 2003-01-12 at 18:23, Ian Molton wrote:
> On Sun, 12 Jan 2003 11:40:57 +0000 (UTC)
> matthias at winterdrache.de (Matthias Benkmann) wrote:
> > Well, he does have a point. Just look at the stuff we get on
> > lfs-support, look at the people who are building LFS. If I'm not
> > mistaken, "recompile gcc and/or binutils" is a very frequent answer on
> > lfs-support.
> Well, if they built with '-O3 -useeverytrickinthebook' then yes, but for
> a normal -O2 gcc / glibc build, I think that LFS is likely to provide
> more solid results than most distros.
> besides, these people should concentrate on FIXING things so that
> miscompiles become rarer, not bitching when someone broke their
> toolchain doing something that SHOULD have worked, but didnt due to
> (more often than not) a GCC optimisation related bug...
Hey now, in the case of gcc and glibc (the two packages where playing
with CFLAGS is near to suicide and people STILL do it) the instructions
that come with the source code are *very* explicit about not screwing
around with compiler flags. Glibc even gives a set of optimizing flags
which they know to work, apparently in hopes of preventing people from
coming up with screwy ones of their own. It simply doesn't seem fair to
ask the developers to spend time working up ways to prevent people from
deliberately miscompiling their code when just telling people flat out
"doing x will break this" can accomplish the same task.
Unsubscribe: send email to listar at linuxfromscratch.org
and put 'unsubscribe lfs-dev' in the subject header of the message
More information about the lfs-dev