Backtracking
I had a sneaky feeling that my testing of the Python 2 to 3 changes was not sofficient. So I downloaded the pyflakes static tester and used it to fix stsci.distutils and pandokia. Found one error of mine and several others which had been in the code base. More
Back on Track
I finished checking all the code I had changed so far with pyflakes, fixing the code, and retesting. I checked in the changes I made to pandokia. I tested the stsci.tools package running under python 3 and it threw some errors that I will have to test tomorrow. More
Two Bugs Squashed
I was able to fix the Python 3 bugs in convertgeis.py and convertwaveredfits.py. The other geis file bugs should be similar to convertgeis.py, but I haven't had a chance to check them yet. More
Mostly done
I finished testing stsci.tools under Python 3 and fixing the problems. I haven't committed the changes, as there was one conflict, which I will rsolve tomorrow. More
Done and Done
I committed the changes to stsci.tools and then converted stsci.ndimage and stsci.convolve, which were small and went fairly quickly. More
Pyfits Converted
I converted the pyfits code to run under python 2 and 3 and I am debugging the test errors. More
Test Issues
I spent most of the day figuring out the proper way to run the tests under nosetests. I made some progress, but still have some issues. More
Mostly Fixed
I resolved the issues I was having with nosetests and committed the changes to pyfits. I started modifying stsci.stimage but need to figure out how to define a new type/class in C from Python 3 before proceding. More
Win Some, Lose Some
I committed the changes to pyfits and stsci.sphere. I gave up for the time being on stsc.stimage. I couldn't figure out how to convert the old style class code to new style class code. I made the Python 3 changes to stwcs, but I am having some problems with the tests. Two errors were reported on ... More
Interlude
Found and fixed the bug in the spehrical geometry package. Took longer than expected, because I couldn't understand the algorithm. I solved it by bisection. The fix is in my github repository, but haven't yet merged it. I am waiting for the Travis tests to run. More
Backporting
I backported the changes in the astropy version of sphere into the stsci package and am debugging it. I also had a first look at the tweakreg code with the intent of fixing a bug. More
Spherical Aberration
I spent the day looking for the error in the spherical aberration code. It looks like the problem is that the perpendicular (normal) vector does not have a unit magnitude. I will investigate further tomorrow. More
Fixed the Fix
Getting the fix to the problem with the spherical code proved to be harder than expected, because adding a call to normalized the unnormalized vector exposed a bug in the normalization code. Both fixes have been uploaded to github, but there's a problem with the merge that I still need to resolve. More