That's s(l)ick enough for my approval! I like it. I've always wondered how we could avoid the overhead of debug calls when we weren't actually using the debugging.
-- Jason On Saturday 10 September 2005 8:04 pm, James Thompson wrote: > I was thinking about gDebug calls tonight (I'm sick like that :) > > If we were to alter our gDebug() calls to *always* return a value of True. > > Then in our code instead of doing > > gDebug(15,"hi") > > we do > > assert gDebug(15,"hi") > > When optimization is requested (python -O) assert statements are left out > of the generated code. So we would regain all the time lost to those > function calls. We would also be able to put gDebug calls in places where > we've removed them in the past to increase performance. > > A quick test with a hacked forms show this works. > > What do people think? > > Take Care, > James > > > _______________________________________________ > Gnue-dev mailing list > [email protected] > http://lists.gnu.org/mailman/listinfo/gnue-dev _______________________________________________ Gnue-dev mailing list [email protected] http://lists.gnu.org/mailman/listinfo/gnue-dev
