Date: Tue, 24 Oct 2006 19:35:23 +0400 From: Andrey Chernov <ache@FreeBSD.ORG> To: Mikhail Teterin <mi+kde@aldan.algebra.com> Cc: Mikhail Teterin <mi+mx@aldan.algebra.com>, current@FreeBSD.ORG, tjr@FreeBSD.ORG Subject: Re: replacing FreeBSD's -lgnuregex with GNUlib's version Message-ID: <20061024153523.GA73555@nagual.pp.ru> In-Reply-To: <200610240749.11234@aldan> References: <200609202304.25537@aldan> <200609261302.40964.mi%2Bmx@aldan.algebra.com> <20060926184447.GA17862@nagual.pp.ru> <200610240749.11234@aldan>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, Oct 24, 2006 at 07:49:10AM -0400, Mikhail Teterin wrote: > = Someone must test its compatibility with GNU regex and understand in > = details nature of their changes/fixes/differences. Without this work we > = can't blindly replace stable code with unknown one just for reason it is > = actively maintained. > > What kind of test would be deemed sufficient? I don't have anything at hand, but I saw a tests in some regex implementations I don't remember now. Perhaps someone else knows good regex test suits? The common bottle neck is locale: collating, multibyte and character classes handling. I can test it excepting multibyte, our multibyte-enabled developers needed. What must be tested before as primary target: general POSIX compatibility. What must be tested in second: GNU regex compatibility. -- http://ache.pp.ru/
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20061024153523.GA73555>