Jim Meyering wrote: ... >>> I seem is_mb_middle has two bugs. >>> This patch (merged) will be corrected both their bugs. >>> >>> 1. fgrep (2.6.1) hangs on a pattern with invalid sequence. >>> >>> ==> See attachment `fgrep-hang'. >>> >>> 2. A complete multibyte string matches with incomplate >>> multibyte pattern with truncated octet character in fgrep (2.6.1). >>> >>> ==> See attachment `truncated-character. >> >> Thank you for the patch and test cases! >> >> I've adjusted your test scripts slightly, hooked them >> up via tests/Makefile.am and added most of the commit logs. >> I confirm that each test exercises a bug that is fixed by the patch. >> I'll study your patch over the weekend. >> Hmm... or maybe not. Continue reading... > > I could not risk studying your patch (don't want to have to wait for > the copyright assignment process), so wrote my own. Since I've had to > modify the test case (with your patch, grep printed the line, with mine > it does not), I gather that our solutions are fundamentally different.
I suspect that our solutions are not so different after all. It's merely that I changed an early version of your test not to use "-v" anymore, and then forgot I'd done that.
