Author mark.dickinson
Recipients christian.heimes, gregory.p.smith, haypo, mark.dickinson, pitrou
Date 2009-02-17.17:36:51
SpamBayes Score 8.36147e-11
Marked as misclassified No
Message-id <1234892214.89.0.450614489214.issue4258@psf.upfronthosting.co.za>
In-reply-to
Content
Updated non-optimized patch.  The only real change is that I've moved 
some of the configuration stuff around (so not worth re-benchmarking 
this one);  I hope that I've now got the division of labour correct:

- configure script simply parses the --enable-big-digits option and sets
  PYLONG_DIGIT_SIZE in pyconfig.h to 15 or 30, or leaves it undefined
  if that option wasn't given to configure

- PC/pyconfig.h doesn't define PYLONG_DIGIT_SIZE

- pyport.h chooses a suitable value for PYLONG_DIGIT_SIZE if it's not
  already defined

- Include/longintrepr.h just follows orders: it expects 
PYLONG_DIGIT_SIZE
  to be defined already, and complains if PYLONG_DIGIT_SIZE=30 but the
  necessary integer types aren't available.

Thanks for all the benchmarking.

I'd probably better check on python-dev before pushing this in, since 
it's a new feature.  I hope no-one wants a PEP. :-)
History
Date User Action Args
2009-02-17 17:36:54mark.dickinsonsetrecipients: + mark.dickinson, gregory.p.smith, pitrou, haypo, christian.heimes
2009-02-17 17:36:54mark.dickinsonsetmessageid: <1234892214.89.0.450614489214.issue4258@psf.upfronthosting.co.za>
2009-02-17 17:36:53mark.dickinsonlinkissue4258 messages
2009-02-17 17:36:52mark.dickinsoncreate