My adventures in Python today

Instead of doing a bunch of separate posts on what I did today that was python-related, I figured I would do package them up into a single post. Hopefully it won't be too random. =)

To begin with, Georg merged in some rewrites of old tests thanks to some GHOP work. As the contributor forms from students trickle in we should eventually end up with no more old tests that base their passing or failing on stdout.

I also discovered that the MacPorts build of sqlite does not play nicely when you run test_ctypes and test_sqlite one after another. After discovering a bug in setup.py where the directories listed in CPPFLAGS were being added in reverse order and figuring out the right thing to download from sqlite.org (the amalgamation tar file with shell.c and the makefile) I confirmed it was not a Python issue at all. So that's nice, but it sure took a while to confirm that fact.

I also managed to fix a race condition in test_queue after Georg did. =) I didn't update my checkout after I finished dealing with my sqlite issue so I didn't notice he fixed the same thing I did.

For my tutorial at PyCon on working with the core I have begun to try to help out the Windows users out there. I sent an email to get some info from people and Christian Heimes has already replied with a lot of info.

Finally, I discovered there are a lot of todo markers in Python. Using the TODO bundle in TextMate while adding marker support for XXX shows that there are 2159 markers. Now granted some of them are listed multiple times (e.g., anything in configure.in is listed at least three times), but that is still a lot of places in code with a comment saying that something should be dealt with.