24 Nov
2011
24 Nov
'11
11:37 a.m.
After going over the code a few more times, and after a quick chat with Bazsi, I cleaned up the patch considerably. So much so, that I'm actually starting to like it! Functionality remained the same, but the code behind it is shorter and contains less magic. Instead of playing tricks with both scratch buffers and allocating gchar* strings all over the place, the transformation functions now operate directly on scratch buffers, and we steal the results. The patch coming in a follow-up message is something I consider ready for review. The only thing left to do is adding a couple of test cases (there's already a simple one included, but I'd like a few more).