Author rhettinger
Recipients amaury.forgeotdarc, collinwinter, eric.smith, pitrou, rhettinger, stutzbach, terry.reedy, tim.peters
Date 2010-11-23.01:16:38
SpamBayes Score 2.21623e-06
Marked as misclassified No
Message-id <1290475000.98.0.110170769958.issue9915@psf.upfronthosting.co.za>
In-reply-to
Content
Thanks for the revisions and timing updates.  I'm heartened that the common-case of sorting without a key function isn't negatively impacted.  That result is surprising though -- I thought the concept was manipulate the key and value arrays at the same time instead of just the keys -- did you do more than this, perhaps changing the logic or pattern of comparisons?  If so, I would be *much* more comfortable if Tim reviewed this.

The one part of the current code that would be missed is that it cleanly separates/decouples the key-function logic from the Timsort logic.  Now those are heavily intertwined -- I find the code harder to follow.

Why did the variable names change, pa/pb to ssa/ssb, etc.?

I'm hoping that I'll have a chance to go through the details of the patch in the next couple of weeks.  Unfortunately, the patch is huge and it looks like it mixes in a number of optimizations beyond just moving keys and values in parallel, variable names are changing, comment lines are rewrapped, etc.
History
Date User Action Args
2010-11-23 01:16:41rhettingersetrecipients: + rhettinger, tim.peters, collinwinter, terry.reedy, amaury.forgeotdarc, pitrou, eric.smith, stutzbach
2010-11-23 01:16:40rhettingersetmessageid: <1290475000.98.0.110170769958.issue9915@psf.upfronthosting.co.za>
2010-11-23 01:16:38rhettingerlinkissue9915 messages
2010-11-23 01:16:38rhettingercreate