RE: stream.eof/AIX/pthread.h

["Paul Moore" <paul dot moore at centrify dot com>]

-pthread stops the error in the simple test case, but I have more
complex case (what I am really trying to build, the loop is identical
but the linkage is huge) and in that one the position of pthread.h makes
no difference and -pthread is always present

the test case is a complete working (or failing) sample in one file -
put in test.cpp and do "g++ test.cpp" 

yes thats the posting i was referring to

 

-----Original Message-----
From: Jonathan Wakely [mailto:jwakely.gcc@gmail.com] 
Sent: Thursday, February 21, 2008 3:58 PM
To: Paul Moore
Cc: libstdc++@gcc.gnu.org
Subject: Re: stream.eof/AIX/pthread.h

On 14/02/2008, Paul Moore <paul.moore@centrify.com> wrote:
> gcc 4.1.1 on AIX 5.2
>
>
>  without pthread included this code works fine. the GetLine function 
> ends  up with eof set, the main function sees eof and stops  with 
> pthread included eof gets set in GetLine but when I pop back to  main 
> eof is not set. When GetLIne is called again eof is set on entry!
>  ie the first cout<< in Getline shows eof is set, but the return to 
> main  has it not set etc.

Does passing (or not passing) -pthread to gcc make any difference either
with or without <pthread.h> ?  And is the testcase you posted a single
file?

http://gcc.gnu.org/ml/libstdc++/2007-05/msg00012.html looks like the
previous report.

Jonathan