[Python-Dev] Re: Making mutable objects readonly
M.-A. Lemburg
mal@lemburg.com
2001年1月31日 13:10:43 +0100
Tim Peters wrote:
>> Note that even adding a "frozen" flag would add 4 bytes to every freezable
> object on most machines. That's why I'd rather .freeze() replace the type
> pointer and .unfreeze() restore it. No time or space overhead; no
> cluttering up the normal-case (i.e., unfrozen) type implementations with new
> tests.
Note that Fred's weak ref implementation also need a flag on every
weak referencable object (at least last time I looked at his patches).
Why not add a flag byte or word to these objects -- then we'd have
8 or 16 choices of what to do with them ;-)
--
Marc-Andre Lemburg
______________________________________________________________________
Company: http://www.egenix.com/
Consulting: http://www.lemburg.com/
Python Pages: http://www.lemburg.com/python/