-
Ah, I didn't realise that Psyco would stop at functions marked with 'cannotcompile()' — there's a note in the docs (3.1):
"Note that this does not prevent functions called by object to be compiled, if they are bound or proxied or if the profiler chooses to compile them."
The problem is that Psyco seems to have problems with the Cheetah templating engine (which compiles html templates into...
2008-11-09 23:11:13 UTC in Psyco, the Python Specializing Compiler
-
It would good if I could stop Psyco from trying to optimise any function under a particular packages (in my case, Cheetah). Since the filter function only looks at code objects, which have no context information, this seems to be presently impossible. Alternatively, it would be good if there was an option to stop the optimisation of functions only called by other functions which have been passed...
2008-11-07 02:26:10 UTC in Psyco, the Python Specializing Compiler
-
You need to install python2.4 (can run alongside python2.5), and then change the first line in /usr/bin/gursormaker to "#!/usr/bin/python2.4".
2008-10-05 12:42:41 UTC in Gursor Maker
-
I used the Debian sources (3.2-9 and 3.3-1).
2008-03-20 16:09:48 UTC in XaoS
-
One more point of interest — after selecting HSV, there is no sign of any calculation. Even using large values for the iterations/bailout which, under palette colour settings and most true colour settings, chew up plenty of CPU time and display the boundary trace progress, show nothing under HSV or HSV2. There is also no sign of CPU usage after telling it to recalculate, but there is when I...
2008-03-20 15:18:40 UTC in XaoS
-
Now that I read it again, it seems clear that you actually meant the latter. So I tried rolling back the patch from the 3.3 sources and still have the same problem. I then tried using the normal (UNmodified) 3.2 package and the problem is still there, so it's not necessarily a 3.2->3.3 problem.
2008-03-20 14:07:12 UTC in XaoS
-
I applied the patch to the 3.2 sources, and still have the same problem. Although I'm not sure if I interpreted your instruction properly — should I have applied the patch to the 3.2 sources, use the unmodified 3.2 sources, or undo that patch from the 3.3 sources?.
2008-03-20 09:02:07 UTC in XaoS
-
XaoS version: from official Debian Sid dist, 3.3-1
Machine: NEC Versa M540
Video: ATI Mobility Radeon
Kernel: Custom 2.6.23 with proprietary fglrx driver 8.44.3-1 (although also occurs with kernel radeon driver)
Display runs at 1280x768 res with 24 depth
When I select true colour outcolouring modes HSV/HSV2, I don't get the results I expect. I find a region of interest using the "smooth"...
2008-03-18 06:54:35 UTC in XaoS
