Fileutils compilation problem

Leonardo Valeri Manera lv2 at
Wed Jul 11 19:47:44 PDT 2001

Gerard Beekmans wrote:

> Oh that I totally agree with. Perhaps I wasn't clear in my previous
> email, I was just mentioning that two gcc versions in the lfs-book isn't
> an option. I rather fix a package than install a different gcc.

agree. however, even though glibc ain't supposed to build with gcc 3.0, 
and it does, c++ on gcc 3.0 is actually quite flaky. noone here at the 
university who i know has been able to compile kde with it, and i had to 
make groff with a blank CFLAGS.

while i abhor the idea of having a rhat-style "second compiler"...if you 
wanna use gcc3.0 there's little to do about it.

I was thinking of installing gcc 3.0 for c only, and gcc 2.95.3 for c++ 
only on my that doable? then one stes two slighly different 
flags for c and c++, and off you go.

> Which is appreciated, but right now fixing packages doesn't do a whole
> lot of good: gcc-3 is a very broken compiler so it's not advisable to
> use it right now, even to give something back.

come on now. it's not that bad if you stick to c, and i *won't* give up 
athlon-optimized code ;)

if you forget about using stuff like -malign-double, which always seems 
to throw up compilatoin errors, that is.

> The main reason simply being that most patches will probably be removed
> when a newer gcc is released, so it would be kind of a waste of time now
> to patch everything up, etc.

i completely agree. doing pathces to make programs work with a compiler 
that will be updated in 2/3 months (maybe less? anyone know how far gcc 
3.0.1 or whatever is in the future?) is a) a waste of time b) quite 
silly really.

IMHO, if a package just won't compile at all, and the fixes needed are 
any more than sub-trivial, then either don't use it, or <aarrgh> use 
another compiler, cuz' the bug is in the compiler, not in the program - 
aprat from a few exceptions - ie: you're not fixing, you're 
work-arounding the bugs, which is *not* a good idea. and a horrific 
programming tactic too.

plus, anyone for making a patch for an important (read: large) program 
like kde/gnome/x is probably still going to be working by the time the 
fixed compiler comes out. it'be an insanity to even consider.

still. gcc 3.0 and i've had no c problems that were actually related to 
the compiler. :)


Unsubscribe: send email to lfs-discuss-request at
and put unsubscribe in the subject header of the message

More information about the lfs-dev mailing list