Date: Sun, 5 Mar 2017 08:47:09 +0100 From: Uli Schlachter <psychon@znc.in> To: Matt Turner <mattst88@gmail.com>, Joe Marcus Clarke <marcus@freebsd.org> Cc: xcb@lists.freedesktop.org, freebsd-threads@freebsd.org Subject: Re: [Xcb] pthread_condattr_{init,destroy} stubs in libc? Message-ID: <505bedb8-c585-44a1-de77-a66334472328@znc.in> In-Reply-To: <CAEdQ38E-4DtCBsSpOz5KAvspT_AC6tAV--a%2B0RHNo=YbsrQm3w@mail.gmail.com> References: <CAEdQ38E-4DtCBsSpOz5KAvspT_AC6tAV--a%2B0RHNo=YbsrQm3w@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On 04.03.2017 16:59, Matt Turner wrote: > Context: https://bugs.freedesktop.org/show_bug.cgi?id=72074#c2 [...] I would summarize the problem as "libpthread-stubs.so provides weak symbols, FreeBSD's threading library provides weak symbols, the result is that a mix of stubs and real functions is used". All of this problem goes away with the new patches against pthread-stubs. There is no longer any libpthread-stubs.so whose weak symbols could cause problems. Instead, pthread-stubs.pc will just "make" things link against libpthread directly and everyone is happy. Right? (Of course, having stubs in libc for the functions which can be stubbed would still be a good idea and would avoid single-threaded programs linking against libpthread) Uli -- Sent from my Game Boy.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?505bedb8-c585-44a1-de77-a66334472328>