mi+mx at aldan.algebra.com
Thu Jul 17 13:50:16 PDT 2003
On Thu, 17 Jul 2003 22:18:38 +0200
Michael Nottebrock <michaelnottebrock at gmx.net> wrote:
=> On Thursday 17 July 2003 22:11, Alexander Kabaev wrote:
=> > -Werror? As doctor said: if it hurts, DON'T DO THAT.
=> In the kdelibs case, it's definitely _not_ -Werror
=Whatever it is, I haven't seen one shred of evidence of "GCC issues" in
=your messages, just complaints. Just an example: bad code generated is
=GCC issue, more strict C++ compliance requirements - not. So what of
=these two did you mean?
First of all, thank you very much for integrating the new GCC into
FreeBSD. The pentium4-specific fixes and optimizations, as well as other
compiler's features and improvements are much appreciated.
Here is how to reproduce the problem, Michael is talking about. Simply
try to build the kdelibs3 (or kdegraphic3, or kdenetwork3) port. It will
die soon enough with a C++ error. It look like, indeed, a stricter C++
compliance issue, but it is not, because:
. it is triggered by something in /usr/include/c++/3.3 itself
. it goes away if you remove the ``-pedantic'' from the Makefiles
(find work/kdelibs* -name Makefile | \
xargs sed -i "" -e 's,-pedantic,,')
Note, that it is, indeed, just -pedantic, not the -pedantic-errors.
So much so, I was suggesting to our KDE team to add the post-patch entry
to the bsd.kde.mk, that would remove ``-pedantic'' automaticly.
More information about the freebsd-current