Re: [Veusz-discuss] speeding up veusz

2009-09-29 Par sujet Jeremy Sanders
Matthew McCormick (thewtex) wrote:

 Yeah, _getLinePoints seems to be one of the buggers.

Okay, I've hacked together a routine to convert arrays to QPolygonF in C++ 
to speed up this loop. It appears to work and is a lot faster than pure 
python.

Unfortunately I've hacked in the support to get it to build in distutils, so 
it won't compile everywhere.

Try the speedups branch from the SVN repository if you want to give it a go. 
You'll need to run setup.py build and install to compile the extension.

There are a few other loops which might be worth converting - the function 
plotter and point plotter.

Jeremy



___
Veusz-discuss mailing list
Veusz-discuss@gna.org
https://mail.gna.org/listinfo/veusz-discuss


Re: [Veusz-discuss] speeding up veusz

2009-09-29 Par sujet Matthew McCormick (thewtex)
On Sun, Sep 27, 2009 at 12:22 PM, Jeremy Sanders
jer...@jeremysanders.net wrote:
 Matthew McCormick (thewtex) wrote:

 Yeah, _getLinePoints seems to be one of the buggers.

 Okay, I've hacked together a routine to convert arrays to QPolygonF in C++
 to speed up this loop. It appears to work and is a lot faster than pure
 python.

 Unfortunately I've hacked in the support to get it to build in distutils, so
 it won't compile everywhere.

 Try the speedups branch from the SVN repository if you want to give it a go.
 You'll need to run setup.py build and install to compile the extension.

 There are a few other loops which might be worth converting - the function
 plotter and point plotter.

 Jeremy

Very cool!  I'm got a lot on my plate this next week, but I'll be
giving it a good look soon.

Matt

___
Veusz-discuss mailing list
Veusz-discuss@gna.org
https://mail.gna.org/listinfo/veusz-discuss