Re: [Python-Dev] Re: ATTENTION! Releasing Python 2.2.2 in a few weeks
Thanks. In addition, I was hoping to hear about your timeline (when do you expect to release PyTie?) and a hint on the 3rd party packages you're thinking of adding. Also a list of target platforms for which PyTie must absolutely work. (Note e.g. that we just discovered a problem with Solaris and the latest version of binutils (2.13), which seems to be used by the latest GCC version (3.2 IIRC) but is also separately downloadable. The bug is in binutils 2.13. Is this *combination* (Solaris + binutils 2.13) a target platform? If so, you might want to use a different approach than we plan to do for Python 2.3 and 2.2.2 (which is merely to bail out if a certain test dumps core during configuration).
--Guido van Rossum (home page: http://www.python.org/~guido/)
Do you know when a fixed binutils is due? This may explain one bug report I have right here at Strakt. Solaris is the preferred platform of our beta-testing customer, Chalmers, so I would like PyTie to run on as many Solaris-including hardware and software platforms as possible. I'll bring this up in a meeting. Right now are you advising people not to use GCC 3.2 or binutils 2.13? or do you have other advice for them which you can steer me towards?
Laura
Laura> Do you know when a fixed binutils is due? This may explain one Laura> bug report I have right here at Strakt. Solaris is the Laura> preferred platform of our beta-testing customer, Chalmers, so I Laura> would like PyTie to run on as many Solaris-including hardware Laura> and software platforms as possible. I'll bring this up in a Laura> meeting. Right now are you advising people not to use GCC 3.2 Laura> or binutils 2.13? or do you have other advice for them which Laura> you can steer me towards?
On my machine, gcc 3.2 works just fine -- it's binutils 2.13 that is the culprit. Use 2.12.1 instead (but be sure to install the configure.in patch I posted earlier today).
I hope to get a patch from the binutils developers today for testing; if it works, I expect that the patch will be in binutils 2.13.1, which I understand is to be released shortly.
Laura Creighton lac@strakt.com writes:
Do you know when a fixed binutils is due?
The bug hasn't been acknowledged by binutils maintainers yet; gcc maintainers report many problems with binutils, but have not identified any specific problem.
So, unless somebody looks down into the details and studies the resulting binaries, it may be a matter of months for a fix to appear. Until then, binutils 2.13 should be avoided on Solaris.
Right now are you advising people not to use GCC 3.2 or binutils 2.13? or do you have other advice for them which you can steer me towards?
gcc 3.2 is fine, binutils 2.13 is not - use 2.12 or the system tools instead.
Regards, Martin
Martin> So, unless somebody looks down into the details and studies Martin> the resulting binaries, it may be a matter of months for a fix Martin> to appear. Until then, binutils 2.13 should be avoided on Martin> Solaris.
Also, note that if you already have binutils 2.13, it is not enough just to reinstall 2.12; you have to rebuild gcc also.
Martin> Laura Creighton lac@strakt.com writes: >> Do you know when a fixed binutils is due?
Martin> The bug hasn't been acknowledged by binutils maintainers yet; Martin> gcc maintainers report many problems with binutils, but have not Martin> identified any specific problem.
Martin> So, unless somebody looks down into the details and studies the Martin> resulting binaries, it may be a matter of months for a fix to Martin> appear. Until then, binutils 2.13 should be avoided on Solaris.
Perhaps on Solaris the Python configure script should detect the presence of binutils 2.13 and barf if it's found? Something like
if [ `uname` = 'SunOS' ] ; then v=`as --version 2>/dev/null \ | head -1 \ | sed -e 's/.* ([^.]*.[^.]*.[^.]*).*/\1/` if [ $? -eq 0 ] ; then # got the gnu version of as - Sun as doesn't grok --version if [ $v = '2.13.0' ] ; then barf fi fi fi
seems like it should come close to working.
Skip
Skip> Perhaps on Solaris the Python configure script should detect the Skip> presence of binutils 2.13 and barf if it's found?
I've already suggested a slightly different test, that has the advantage of allowing a patched 2.13 (and of detecting a broken 2.13.1 should it still be broken).
participants (4)
-
Andrew Koenig -
Laura Creighton -
martin＠v.loewis.de -
Skip Montanaro