Aaron Griffin wrote:
On Mon, Jul 28, 2008 at 8:46 AM, Allan McRae <allan.mcrae@qimr.edu.au> wrote:
Hi all,
Does anybody have a ncurses build list? I am building ncurses with only the wide-character libraries which should hopefully clear up a few of the locale issues with applications using it (e.g. FS#4418, FS#6384, FS#9872). The current situation is sort of between doing this and not (we have only the wide character include files but both libraries).
After some initial problems (don't install an ncurses without a real libncurses.so before fixing readline & bash!), I have this up and working and most of the needed core rebuilds done. I'm thinking I will push it to a separate repo first so that I can get some testing of this by people with other locales before I push it to [testing].
Are you sure this actually fixes these issues? I suggested the same thing a while back, but Roman pointed out something regarding Cyrillic chars or something... I don't know the details, but it had to do with the fact that not all locales are really UTF friendly.
So I see.... (http://bugs.archlinux.org/task/4418#comment14208). Am I 100% percent sure? No. But I am quite confident (say 97%). From what I have read wide-character libraries work with both multibyte and 8-bit locales, while normal libraries work only in 8-bit locales. Anyway, that niggling 3% is why I don't want to push it directly to testing.... How about I completely rebuild the stuff in core and stick it somewhere. Then people can make a testing chroot, install the 10 or so updated core packages and check that man pages work properly in their locales. If that doesn't break, then I will go for the whole rebuild.
Also, let me see if our old TODO lists have an ncurses list.
That would be great. Thanks. Allan