Nick Craig-Wood wrote:
I tried it on a program I wrote to solve a puzzle (Rush Hour). (Actually it solves the meta-puzzle - trying to make the hardest possible Rush Hour puzzle.) After a bit of twiddling (remove psyco and profiling) I got it to start compiling, but unfortunately it compiled for about an hour (in iterative type analysis..) used up 600 MB of RAM printing an '*' every 10 minutes or so then gave an error message and gave up. Unfortunately I shut the window by accident, but the error message was something about not being able to resolve types I think with a list of 20 or so unresolved types. Can you give a hint as to how I debug this? I presume my program has some instances of non static types which is causing the problem, but it is going to be a very long debugging session if it takes me an hour each cycle ;-) The program is about 700 lines of python (excluding comments).
Split it into pieces and compile each separately. Recurse. tjr -- http://mail.python.org/mailman/listinfo/python-list