Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 02 Jun 2020 18:42:21 +0000
From:      bugzilla-noreply@freebsd.org
To:        bugs@FreeBSD.org
Subject:   [Bug 246630] stable/11 regression: base.txz reproducibility depends on number of cpu cores
Message-ID:  <bug-246630-227-GKKaeWOt5U@https.bugs.freebsd.org/bugzilla/>
In-Reply-To: <bug-246630-227@https.bugs.freebsd.org/bugzilla/>
References:  <bug-246630-227@https.bugs.freebsd.org/bugzilla/>

next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D246630

--- Comment #24 from Dimitry Andric <dim@FreeBSD.org> ---
(In reply to Alexandre Ganea from comment #23)

This is the tricky part! I've only been able to reproduce this problem until
now within the FreeBSD build system. That is, the result of one particular =
file
being compiled is dependent on whether make is running with or without -j (=
and
in some cases, the value of the -j option).

As of now, I can "fix" the problem by using #define CLANG_SPAWN_CC1 1 in
lib/clang/include/clang/Config/config.h, which effectively disables the
in-process cc1 by default. At least then you always get the same object file
output.

But as to *why* the result differs when CLANG_SPAWN_CC1=3D0, and being invo=
ked
under different variations of make or make -j, it is not clear yet.

One thing which Bryan Drewery just mentioned on IRC is that make -j messes =
with
stderr, causing it to no longer be a tty. I think that must have some sort =
of
influence, but it's going to be tricky to debug (especially if you want to =
send
debug output to stderr :).

--=20
You are receiving this mail because:
You are the assignee for the bug.=



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-246630-227-GKKaeWOt5U>