Date: Tue, 24 May 2011 01:42:49 +0000 From: Alexey Dokuchaev <danfe@FreeBSD.org> To: Matthias Andree <mandree@FreeBSD.org> Cc: cvs-ports@FreeBSD.org, ports-committers@FreeBSD.org Subject: Re: cvs commit: ports/security/yara Makefile distinfo ports/security/yara/files patch-libyara__proc.c Message-ID: <20110524014249.GA93724@FreeBSD.org> In-Reply-To: <4DDAA478.3030203@FreeBSD.org> References: <201105231554.p4NFsBdt024494@repoman.freebsd.org> <20110523170435.GA15903@FreeBSD.org> <4DDAA478.3030203@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, May 23, 2011 at 08:16:24PM +0200, Matthias Andree wrote: > Am 23.05.2011 19:04, schrieb Alexey Dokuchaev: > > Could you perhaps explain the rationale behind pcre -> re2 (never heard > > of it before) change? I mean, lots of ports depend on devel/pcre, which > > seems to be de-facto regex library, and having two similar (well, that's > > maybe an underestimation) libraries installed in a system without some > > sufficiently strong reason seems at least questionable. (Unless, of > > course, this port deliberately requires re2 and cannot be built against > > pcre or easily converted.) > > PCRE is questionable because the standard regexp matcher is regex.h/libc. Lots of ports depend on pcre, which means that whatever regex.h/libc offers is not enough. In this context I refer to pcre as standard library when Perl-compatible (read: real) regexps are wanted. > Seriously, we should question the use of software that has awkward > worst-case behaviour... I wouldn't mind. I am firmly advocating careful and uniform dependecy selection across Ports Collection. ./danfe
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20110524014249.GA93724>