Re: Help compiling gcc 4.2.0

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Lee Rhodes wrote:

> How should I interpret the above?  In searching the gcc.log and g++.log
> files I did not find undefined references to _ungetc, __srget_r (wrt
> msg00948.html below), but did find a few other undefined references that did
> result in FAILs: e.g., _mcount,__monstartup.

If you had run into the problem you would have gotten an error building
libgcc that would have aborted the bootstrap, so you wouldn't have been
able to run the testsuite.  I guess 4.2 is too old to hit that issue.

As for interpreting the results... well, some would say that 4.x is
still not quite ready for prime time on Cygwin/MinGW, which is why
neither project has released gcc4 binary packages yet.  There are still
some outstanding problems with using gcc 4.x on these platforms, the
most major of which is lack of shared target libraries.  Until that's
fixed exception handling will be broken in various ways, among other
issues.  It may be that none of these issues applies to your particular
use case, in which case you can use gcc4 all you want, but from a
project standpoint it doesn't make sense to release a compiler that
kinda-sorta-almost works except for sometimes, so that's why both
projects still ship gcc 3.4.

> script.  Can I tar.gz up the gcc.sum, gcc.log, g++.sum, g++.log files and
> send it to gcc-testresults@xxxxxxxxxxx?  Or is there a particular structure
> for these results.

You can look at the archives online, but the list pretty much consists
entirely of the output of contrib/test_summary, so I wouldn't.  You can
still post the output if you don't have a command line mail client
installed, it's just plain text.  I'd suggest if you're going to go to
the trouble to post the results that you run the whole testsuite and not
just check-gcc.  You can do this with "make -k check" from the
toplevel.  The check-gcc rule doesn't test any target libs.

> P.S. Sorry about the dual post, I didn't realize the lists were connected.

They're not connected, it's just that starting the same thread in more
than one place without CCing the other generally just wastes time
because people have to either reply twice or read two copies of the
thread.

Brian

[Index of Archives]     [Linux C Programming]     [Linux Kernel]     [eCos]     [Fedora Development]     [Fedora Announce]     [Autoconf]     [The DWARVES Debugging Tools]     [Yosemite Campsites]     [Yosemite News]     [Linux GCC]

  Powered by Linux