Date: Wed, 13 Jul 2016 19:02:59 +0000 (UTC) From: "Andrey A. Chernov" <ache@FreeBSD.org> To: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-head@freebsd.org Subject: svn commit: r302781 - head/gnu/usr.bin/grep Message-ID: <201607131902.u6DJ2xAu008280@repo.freebsd.org>
next in thread | raw e-mail | index | archive | help
Author: ache Date: Wed Jul 13 19:02:59 2016 New Revision: 302781 URL: https://svnweb.freebsd.org/changeset/base/302781 Log: After removing collation for [a-z] ranges in r302512, do it here too. Instead of trying to expand whole range at regcomp() stage as we do, GNU regex allocates separate ranges [start,end] set each character is checked against, so collation is possible and turned on for ranges here. When something like that will be implemented or our obsoleted regex code will be replaced to something like TRE, and in case we decide to use collation in [a-z] ranges, all changes related to r302512 can be backed out, but now we need consistency. Modified: head/gnu/usr.bin/grep/dfa.c Modified: head/gnu/usr.bin/grep/dfa.c ============================================================================== --- head/gnu/usr.bin/grep/dfa.c Wed Jul 13 18:51:19 2016 (r302780) +++ head/gnu/usr.bin/grep/dfa.c Wed Jul 13 19:02:59 2016 (r302781) @@ -2547,8 +2547,13 @@ match_mb_charset (struct dfa *d, int s, wcbuf[2] = work_mbc->range_sts[i]; wcbuf[4] = work_mbc->range_ends[i]; +#ifdef __FreeBSD__ + if (wcscmp(wcbuf, wcbuf+2) >= 0 && + wcscmp(wcbuf+4, wcbuf) >= 0) +#else if (wcscoll(wcbuf, wcbuf+2) >= 0 && wcscoll(wcbuf+4, wcbuf) >= 0) +#endif goto charset_matched; }
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201607131902.u6DJ2xAu008280>