grep -F skips an extra character after matched middle of multi-byte character. As a result, fails to match at the next position. A test case for this bug is already included in tests/sjis-mb.
For example, Following test is failure on grep-2.19 or later. $ printf '\203AA\n' >in $ env LC_ALL=ja_JP.SHIFT-JIS src/grep -F A in We expect that it matches at the second A and outputs the line, but doesn't output.
From bd5e4650a4d6af7322c623acff647444138e8360 Mon Sep 17 00:00:00 2001 From: Norihiro Tanaka <[email protected]> Date: Tue, 18 Nov 2014 13:36:42 +0900 Subject: [PATCH] grep: grep -F fails to match at the next position after matched middle of a multi-byte character grep -F skips an extra character after matched middle of multi-byte character. As a result, fails to match at the next position. A test case for this bug is already included in tests/sjis-mb. src/kwsearch.c (Fexecute): Skip correctly after matched middle of a multi-byte character. --- NEWS | 4 ++++ src/kwsearch.c | 2 +- 2 files changed, 5 insertions(+), 1 deletion(-) diff --git a/NEWS b/NEWS index 1597528..6c5d8e1 100644 --- a/NEWS +++ b/NEWS @@ -48,6 +48,10 @@ GNU grep NEWS -*- outline -*- of a multibyte character when using a '^'-anchored alternate in a pattern, leading it to print non-matching lines. [bug present since "the beginning"] + grep -F no longer fails to match at the next position after matched + middle of a multi-byte character. + [bug introduced in grep-2.19] + grep -E rejected unmatched ')', instead of treating it like '\)'. [bug present since "the beginning"] diff --git a/src/kwsearch.c b/src/kwsearch.c index aa965f6..c0cf0ad 100644 --- a/src/kwsearch.c +++ b/src/kwsearch.c @@ -135,7 +135,7 @@ Fexecute (char const *buf, size_t size, size_t *match_size, { /* The match was a part of multibyte character, advance at least one byte to ensure no infinite loop happens. */ - beg = mb_start; + beg = mb_start - 1; continue; } beg += offset; -- 2.1.3
