libiconv status report #3

[Gabor Kovesdan]

Hello,

this week's progress:
- Clean up in the esdb and csmapper databases.  This means, finally we 
have working Makefiles to install them correctly on FreeBSD. This seems 
to be a misc part of the work but actually, it is important that we can 
easily add/modify conversion tables. I've tried to generalize the 
building of these pieces of metadata and made a mini-framework for that.
- Added a TESTING-HOWTO as we can easily install iconv now.
- Fixed some issues showing up in the GNU regression tests by adding 
-INTERNAL and -SWAPPED encoding forms for UTF/UCS family.  The former 
always means using the machine byte order, while the latter one is the 
swapped byte order.  GNU regression tests use UCS-4-INTERNAL and by 
implementing it in a compatible way, lot of regressions went away in 
other encodings, as well.
- While here, added some more aliases used in GNU iconv for the UTF/UCS 
family.
- Finally, really fix BOM handling.
- Nits in the test suite.
- Some investigations about some encoding variants.

Short-term TODO: The basic encodings seem to be well supported, except 
UTF-8 and UTF-7, which still have some issues according to the GNU 
regression test. The next important task is to fix UTF-8 before I can 
pass to the CJK encodings. I don't consider UTF-7 very important because 
it is kind of a legacy encoding but it would be nice to finish the whole 
UTF/UCS family, so I'll look into it, as well.

Cheers,

-- 
Gabor Kovesdan
FreeBSD Volunteer

EMAIL: gabor at FreeBSD.org .:|:. gabor at kovesdan.org
WEB:   http://people.FreeBSD.org/~gabor .:|:. http://kovesdan.org