Hi,
At Wed, 11 Sep 2002 13:21:32 -0400,
Glenn Maynard wrote:
> I havn't heard of this, and I'd be surprised that Tomohiro doesn't seem
> to know about it (at least he's never mentioned it). Could you give an
> example or a reference? (This would be a major bug, since round-trip
> compatibility is extremely important.)
I have never heard about this problem, if we use the same system to
forward and backward conversions.
I think, if we use the same mapping table for both of forward and
backward conversions, such problem cannot occur (except for lacking
of conversion for a codepoint). And, I think glibc uses the same
mapping table for EUC-JP -> UTF-8 and UTF-8 -> EUC-JP conversions.
Different from most other encodings, EUC-JP (and other JIS-based
encodings) surely has problem of mapping tables. For example,
each of Microsoft, Apple, IBM, Sun, Glibc, and so on uses different
tables. JIS-based text data will suffer cross-platform incompatibility
in future when Unicode become more popular.
There are several possibilities of such a problem even when we think
only about Debian. It is because there are some softwares which have
their own mapping tables. The followings are examples of softwares
which have their own mapping tables:
GNU libc
XFree86
Java
Tcl/Tk
Thus, EUC-JP --(Java-based software on Debian)--> UTF-8 --(GNU libc-
based software on Debian)--> EUC-JP might lose round-trip compatibility.
I think Debian GNU/Win32 will be a chaos on this point.
---
Tomohiro KUBOTA <[EMAIL PROTECTED]>
http://www.debian.or.jp/~kubota/
"Introduction to I18N" http://www.debian.org/doc/manuals/intro-i18n/
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
