Mailing List Archive
tlug.jp Mailing List tlug archive tlug Mailing List Archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]Re: [tlug] Japanese input on Debian Testing
- Date: Thu, 17 Nov 2011 12:39:09 +0900
- From: "Stephen J. Turnbull" <stephen@example.com>
- Subject: Re: [tlug] Japanese input on Debian Testing
- References: <201111151528.08036.daniel.ramaley@example.com> <8762ikkb8q.fsf@example.com> <20111116114334.GA15023@example.com>
Mattia Dongili writes: > > I *cannot* believe that they haven't fixed this yet! The POSIX locale > > should be a *hint* (and a very valuable one, because 99% of users are > > either monolingual or their other language is English == spellable > > with ASCII, but nevertheless *just* a hint). This has been known to > > be broken for multilingual users since the early 1990s. :-( > > my experience is that sun's (oracle's?) java has the problem, > open{jdk,jre} icedtea doesn't. All software which uses the locale and/or other environment variables to set the input method suffers from this. It's a P aimed directly at your A, even if you haven't felt it yet. The problem is the design of POSIX locales, which is tuned to convenience for monolingual users. That's fine, most users *are* monolingual, but a system that sets input methods according to the POSIX locale at startup *must* provide a convenient escape for multilingual users -- and, except for Emacsen[1], I don't know any runtime selection mechanism that's truly convenient. For example, the insane bullshit of having to register a new input method with GTK, and restart applications, to use it. Footnotes: [1] And even there, Richard "POSIX_ME_HARDER" Stallman[2] screwed GNU Emacs users by forcing them to use the horribly awkward XIM instead of linking directly to libraries like Wnn and Canna. I implemented a proof-of-concept switching mechanism for XEmacs, but never put it in the mainline because we do link directly to those libraries, so users never complained. GNU Emacs users eventually reimplemented the network protocols for Wnn and Canna in Lisp, but I don't know what they could do for Anthy etc that insist on linking .so's for security reasons. [2] Richard normally rejects standards for the sake of standards, arguing that the GNU System should be its own standard, and follow public standards where it chooses. Why he chose to be Dick in this case is beyond me.
- References:
- [tlug] Japanese input on Debian Testing
- From: Daniel A. Ramaley
- [tlug] Japanese input on Debian Testing
- From: Stephen J. Turnbull
- Re: [tlug] Japanese input on Debian Testing
- From: Mattia Dongili
Home | Main Index | Thread Index
- Prev by Date: Re: [tlug] Japanese input on Debian Testing
- Next by Date: [tlug] ASUS U30SD sound issues
- Previous by thread: Re: [tlug] Japanese input on Debian Testing
- Next by thread: [tlug] ASUS U30SD sound issues
- Index(es):
Home Page Mailing List Linux and Japan TLUG Members Links