Author lemburg
Recipients Mark.Shannon, jtaylor, lemburg, pitrou, rhettinger, serhiy.storchaka, tim.peters
Date 2015-03-07.13:43:17
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <1425735798.04.0.747732037604.issue23601@psf.upfronthosting.co.za>
In-reply-to
Content
There seem to be quite a few other places where this simple optimization could make sense as well, perhaps even going as far
as converting all uses of PyMem_MALLOC to PyObject_MALLOC.

There was a time when the small memory allocator did not return
free arenas to the system allocator, but those are long ago.

The only detail to watch out for is not mixing the malloc/free APIs. In particular, memory which is allocated in the interpreter and then deallocated elsewhere in extensions needs to continue using the PyMem_* APIs.
History
Date User Action Args
2015-03-07 13:43:18lemburgsetrecipients: + lemburg, tim.peters, rhettinger, pitrou, Mark.Shannon, jtaylor, serhiy.storchaka
2015-03-07 13:43:18lemburgsetmessageid: <1425735798.04.0.747732037604.issue23601@psf.upfronthosting.co.za>
2015-03-07 13:43:18lemburglinkissue23601 messages
2015-03-07 13:43:17lemburgcreate