On Thu, 07 Jun 2001, Valdis Kletnieks said:
On Thu, 07 Jun 2001 12:01:06 +0900, "Stephen J. Turnbull"
> uname -a: SunOS lift 5.7 Generic_106541-14 sun4u sparc SUNW,Ultra-5_10
Ah yes. GCC, the portable compiler that has to be kept in fascist lockstep
with your operating system. Looks like you built it under 2.5.1, and run
That's not what `portable' means.
it under 2.7. Compile-time errors are probably caused by the GCC
for running fuxincludes and stashing copies in a private directory where
they will cause problems if you ever install patches to your system.
Not in this case, no. Modern GCCs do not run fixincludes on Solaris,
because Solaris's includes are OK these days. However, there are lots of
other platform-specificities that changed between 2.5.1 and 2.7 that
must be compensated for.
It's not GCC's job to provide a working /usr/include -
that's the vendor's
job. The "excuse" that gcc needs to do this to work is a crock.
And how would you suggest coping with grossly nonstandard headers? We
can't order the vendors to fix them, and even if they do there will be
people who never upgrade from the versions with the broken
headers. IMHO, it is better to find a kludge to work on those platforms
than to leave those people without a working GCC.
/Valdis (who boycotts gcc on platforms that have a vendor compiler,
It's a strange reason. Avoid GCC on platforms where the vendor compiler
is better and you're not cross-compiling, sure, but `boycotting' it
because of broken vendor headers (i.e. no fault of GCC's!) is just
(It is true that there should be a way to rerun fixincludes after GCC is
installed; GCC 3, due out in a week, will already warn if your
fixincluded headers are outdated, and discussions on a rerunnable
fixincludes are underway on the GCC list even now. Well, even last week,
the last time I had time to read it.)
`"This code is gross!" meaning "This code has over 144 compilation
--- Correct use of English, from jer