bug-make
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: different behavior with -j2 on solaris


From: Paul D. Smith
Subject: Re: different behavior with -j2 on solaris
Date: Wed, 16 May 2001 18:23:53 -0400

> With all things being equal, -j2 causes an apparent failure of
> the $(wildcard ...) function for us.

Please always specify which version of GNU make you're using.

It's very likely that you're running into a known problem with signals:
the jobserver requires non-restartable system calls, but the code is not
safe for this: in some places it doesn't check for EINTR results.  This
can cause system calls to fail and not be noticed.

I have a comprehensive solution to this problem which _doesn't_ involve
adding EINTR checks to every system call in GNU make (ouch!).

In the meantime if you like you can try to fix it by applying the patch
you'll find attached to PR/2233 in the bug database.  It seems in
particular you're falling afoul of the readdir() part.

However, note that this is _not_ the way I'm proceeding to fix this bug
"for real": there are many more system calls in GNU make that could fail
than are fixed by this patch--but it seems to help many people, so maybe
it will work for you as well.

The bug database is available at http://bugs.gnu.org/cgi-bin/gnatsweb.pl
and you can login as user "guest", password "guest" (no quotes of
course).

-- 
-------------------------------------------------------------------------------
 Paul D. Smith <address@hidden>          Find some GNU make tips at:
 http://www.gnu.org                      http://www.paulandlesley.org/gmake/
 "Please remain calm...I may be mad, but I am a professional." --Mad Scientist



reply via email to

[Prev in Thread] Current Thread [Next in Thread]