threads

[Hrvoje Niksic]

"Gordon McMillan" <gmcm at hypernet.com> writes:

> Hrvoje Niksic writes:
> 
> > "Barry A. Warsaw" <bwarsaw at cnri.reston.va.us> writes:
> > 
> > > You might be interesting in digging around for Greg Stein's
> > > free-threading patches.  Greg did a lot of work, I think in the
> > > Python 1.4 days, to remove the central interpreter lock and make all
> > > the data structures and such thread safe.
> > 
> > Why would I want to do that? 
> 
> See that other thread.

Huh?  What other thread?  Maybe the question wasn't clear enough.

What I mean is: at this time, my need for simultaneous thread
execution in Python is not so urgent that to warrant forking off (no
pun intended).  Once I do need it, I'd prefer to be able to choose
Python than to dismiss Python on the grounds of not properly
supporting threading.

> > If Greg's patches haven't been applied to the base, surely mine
> > would not be either.
> 
> As I recall, Greg's patches didn't make it because (1) they were late 
> and (2) they were slow.

Interesting.  What was the source of slowness?  (If you _remove_
locks, I'd expect things to go faster, even in a single-threaded
world.)  Maybe the patches simply introduced a different set of
problems?