Message204198
Updated patch. I've also taken the opportunity to increase MAX_ABS_EXP, which fixes the following long-standing bug (not that anyone has noticed it in the last 4 years, apparently):
>>> s = '0.' + '0'*19999 + '1e+20000'
>>> float(s) # should be 1.0
0.1
(The new limit on the number of digits means that the clipping of the exponent can no longer result in incorrect results from strtod.)
Unlike the previous patch, I'm reasonably confident in this one. :-) I'll plan to apply it soon, though it probably won't make 3.4 beta 1. |
|
Date |
User |
Action |
Args |
2013-11-24 11:44:46 | mark.dickinson | set | recipients:
+ mark.dickinson, vstinner, eric.smith, christian.heimes, jkloth, zach.ware |
2013-11-24 11:44:46 | mark.dickinson | set | messageid: <1385293486.48.0.357587596722.issue19638@psf.upfronthosting.co.za> |
2013-11-24 11:44:46 | mark.dickinson | link | issue19638 messages |
2013-11-24 11:44:46 | mark.dickinson | create | |
|