Piet van Oostrum piet at cs.uu.nl
Tue Aug 27 04:56:54 EDT 2002

• Previous message (by thread): Question about Python threads
• Next message (by thread): Question about Python threads
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

---

>>>>> vvainio at tp.spt.fi (Ville Vainio) (VV) writes:

VV> Greg Ewing <see_reply_address at something.invalid> wrote in message news:<3D6AE99C.5060202 at something.invalid>...
>> a GIL. Even without refcounting, lack of a GIL would
>> require locks around many other things, e.g. anything
>> which mutates an object (inserting/deleting list
>> items, etc.) Also object allocation, which I suspect

VV> I don't think object mutation would be that much of a problem - it's
VV> natural for programmers to *assume* that manupulating such data
VV> structures is not thread safe, and expecting programmers to use
VV> explicit mutexes or use thread safe versions of the objects wouldn't
VV> really be too radical. At least I automatically assume that any
VV> manipulation of common data is not threadsafe.

But you don't want the datastructures to become invalid (dangling pointers
etc.) so the interpreter must be careful anyway.
-- 
Piet van Oostrum <piet at cs.uu.nl>
URL: http://www.cs.uu.nl/~piet [PGP]
Private email: P.van.Oostrum at hccnet.nl

---

• Previous message (by thread): Question about Python threads
• Next message (by thread): Question about Python threads
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

---

More information about the Python-list mailing list