Date: Tue, 28 Nov 2006 00:59:36 +0900 From: Norikatsu Shigemura <nork@FreeBSD.org> To: Mikhail Teterin <mi+kde@aldan.algebra.com> Cc: ports@FreeBSD.org, Norikatsu Shigemura <nork@FreeBSD.org> Subject: Re: ports/devel/icu: PTHREAD_LIBS clean Message-ID: <20061128005936.5ae4b851.nork@FreeBSD.org> In-Reply-To: <200611271032.18366@aldan> References: <20061127222709.aa35ab22.nork@FreeBSD.org> <200611271032.18366@aldan>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, 27 Nov 2006 10:32:17 -0500 Mikhail Teterin <mi+kde@aldan.algebra.com> wrote: > On Monday 27 November 2006 08:27, Norikatsu Shigemura wrote: > = Hi mi! > = I found a problem that icu don't reflect PTHREAD_LIBS=ANY(like -lthr). > = Please check following patch. > Thanks, I'll try it... Do you need it to be thread-aware though? It is built > by default without threads at all (because threads are/were busted on > FreeBSD/ia64). Now, there is a argument about libthr is default on 7.x. We test libthr environment little(maybe only on sparc64). But we don't almost test it on i386. So I'm testing libthr environment on 7-current/i386. It is a rule that BLENDING THREAD LIBRARIES IS TOO DANGER. So if we try to get libthr environment, we must select following approch. 1. PTHREAD_LIBS=-lthr So I removed libpthread.*, I'm testing. 2. WITHOUT_LIBPTHREAD=yes >> /etc/src.conf and make world ru@ committed this way. So I can do it. But we can only use on 7-current. If you'll try to get libthr environment, you can only select 1. 3. libpthread.so.2 -> libthr.so.2 in /etc/libmap.conf IS TOO BAD I don't think this approch is good, believe that I'll get unstable environment. > ICU developers think, threads ought to be enabled, but there really is no > software at the moment, which calls into ICU from multiple threads at the > same time, AFAIK... I don't think trying to use multiple threads at the same time. But if applications use libthr, libraries must use libthr, too.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20061128005936.5ae4b851.nork>