Not to be pedantic, but the major concern others are voicing is that is that queue size is not reliable and is therefore a potential source of hard to find threading bugs by naive users. Why not just rename q.size() to the unweildy name of q.est_size()?
On Jan 14, 2008 11:59 AM, Guido van Rossum <[EMAIL PROTECTED]> wrote: > Sounds like we have all-round agreement. Go for it. > > On Jan 14, 2008 11:55 AM, Raymond Hettinger <[EMAIL PROTECTED]> wrote: > > > I'm not sure people who ignore the big "Because of > > > multithreading semantics, this is not reliable" warnings > > > should be catered to. Since others have contributed use-cases > > > for qsize()'s advisory information, it should probably stay around. > > > > I concur. > > > > I do recommend we dump q.empty() and q.full(). The right way is to trap > > the Empty and Full exceptions. If needed qsize() is available to make your > > own less reliable checks. > > > > More than just simplifying the API, the improvement makes it easier to roll > > your own Queue (like a priority queue style). Currently, we require that > > six methods get overridden (_init, _empty, _full, _qsize, _put, and _get). > > It would be nice to lower the burden to just the basic four. > > > > > > Raymond > > > > > > -- > --Guido van Rossum (home page: http://www.python.org/~guido/) > _______________________________________________ > > Python-3000 mailing list > [email protected] > http://mail.python.org/mailman/listinfo/python-3000 > Unsubscribe: > http://mail.python.org/mailman/options/python-3000/charles.merriam%40gmail.com > _______________________________________________ Python-3000 mailing list [email protected] http://mail.python.org/mailman/listinfo/python-3000 Unsubscribe: http://mail.python.org/mailman/options/python-3000/archive%40mail-archive.com
