- From: Yosi Scharf <syosi@MIT.EDU>
- Date: Mon, 07 Aug 2006 10:33:17 -0400
- To: public-cwm-talk@w3.org
That's because the change wasn't in check.py . I ran a profile and figured out it was spending most of its time canonicalizing. Cutting that out was giving me order of magnitude speedups, with few regressions. I think that's what you are seeing too. Thus, the changes are in llyn.py and query.py . Yosi > Waw! this goes so fast now, much faster than I was experiencing before.. > bash .check -v50 http://eulersharp.sourceforge.net/2006/02swap/gedcomE.n3 > > the observable diff in check.py code is > > [[ > RCS file: /w3ccvs/WWW/2000/10/swap/check.py,v > retrieving revision 1.53 > retrieving revision 1.54 > diff -r1.53 -r1.54 > 15c15 > < __version__ = '$Id: check.py,v 1.53 2006/07/31 21:37:54 syosi Exp > $'[1:-1] > --- > > __version__ = '$Id: check.py,v 1.54 2006/08/02 16:59:14 syosi Exp > $'[1:-1] > 79c79 > < f = load(uri, flags="B") # why B? -DWC > --- > > f = load(uri, flags="B").close() # why B? -DWC > 81a82 > > assert f.canonical is f, `f.canonical` > 570c571 > < raise StopIteration > --- > > # raise StopIteration > ]] > > -- > Jos De Roo, AGFA http://www.agfa.com/w3c/jdroo/
Received on Monday, 7 August 2006 14:33:40 UTC