[regression tests] failures in path simplication unit tests?

Hi,

please, can someone dealing with the path simplication have a look at
the log attached?

I'm on 10.6 Mac OS X with astraw repo 10/18/2010.

The symlog mismatch is said to by fixed; for the pcolormesh mismatch I
cannot see any visual difference. Might by false alarm. But the
``len(simplified)`` things might be worth being considered I guess. I
have numpy.__version__ 1.4.1. The python running the tests is 32bit
(although compiled fat). Python 2.6. mpl compiled locally of course
by the fat Python, so both archs. gcc-4.2 on 10.6 as usual. ...

Friedrich

mpl-test.txt (6.53 KB)