Front page | perl.perl5.porters |
Postings from June 2011
Howdy people. Exams are mostly over until next Wednesday, so progress will pick up again this week. During the weekend I pushed a hopefully cleaner version of the pad commits, and in the next couple of hours I'll be doing the same with the stash/GV branch ( https://github.com/Hugmeir/gsoc-pad-utf8-safety/tree/gsoc-stash-utf8-safety, for now). Currently there's a bunch of exposed HEKs in the stash branch that should probably end up hidden behind macro calls; any thoughts on how to tackle that would be very welcome. There are three pad-related warnings that aren't clean at the moment; One I have fixed locally and I'm waiting for the test suite to finish before committing, but the other two are in toke.c and are called through yyerror, which isn't clean. So those are a TODO for whenever I get there. While going through the pad commits I managed to get our to play nice with the clean stashes/GVs/pads, which took a grand total of half a line to implement, so that's working now; And while going through the stash commits, I found out why things suddenly required G_AUTOLOAD, so that's fixed too. Huzzah! That's about it for the week though. Moving on, I realize now that I should have addressed the following a lot sooner, but well, hindsight is 20/20. This comment from hv.h is nagging me: /* FIXME - all of these should use a UTF8 aware API, which should also involve getting the length. */ I have no idea what that means. What would be a UTF8 aware API for the Hv* macros? This week I'll be working on getting the stash repo up to shape. Also, tests; There's a huge number of warnings that still need UTF-8 test cases (and an equally huge amount that still aren't clean), which make up for a sizable portion of my TODOs.Thread Next