[PATCH] sed: fix minor multibyte parsing bug

[Assaf Gordon]

Hello,

The attached patch fixes two somewhat obscure bugs in parsing of multibyte 
characters under very specific conditions.

Sadly it seems that "ja_JP.SJIS" locale under Mac OS X is buggy, and it treats 
invalid sequences like '\203\060' as valid, so one test fails on Mac OS X 
(similar to the situation with 'invalid-mb-sea-umr.sh').

Perhaps a more robust solution is to add a detailed M4 configuration file to 
detect valid Japanese locales, and also pull-in gnulib's mbrtowc replacement.

Comments welcomed,
 - assaf

0001-sed-fix-minor-multibyte-parsing-bug.patch
Description: Binary data