Re: 2.9.11?

[Daniel J Sebald]

Daniel J Sebald wrote:
> A S Hodel wrote:
>
> > 2.9.10 (and the current CVS) appear to work for me, but "make check"  
> > hangs for me on the system calls on Mac OS X 10.4.9 on an Intel-mac.   
> > I'm on a different machine right now, so I don't have further details  
> > than that available.
>
>
> I tried "make check" last night too.  It also hanged.  I'm using linux, 
> Fedora Core 3.

Hmm, doesn't always hang, however.  Only sometimes.  Is there something not good 
about this test?  Here is probably what is hanging:

%!  do
%!     if (!isunix())
%!       errno (0);
%!     endif
%!     s = fgets (out);
%!     if (ischar (s))
%!       idx++;
%!       str{idx} = s;
%!     elseif (errno () == EAGAIN)
%!       sleep (0.1);
%!       fclear (out);
%!     else
%!       done = true;
%!     endif
%!   until (done)

But what happens if an error is return, say the sort routine outside of Octave 
isn't complete by time of the first "ischar()" test.  The errno() returns EAGAIN 
and in the process of waiting 0.1 seconds with sleep() the sort application 
finishes.  But then the "fclear(out)" might clear "data available" so that 
EAGAIN comes back the next time, after which point it is simply an infinite loop.


The only other unusual results are the scripts/elfun/acscd.m and asecd.m tests 
with the division by zero problem that hangs.  And this one, which I don't know 
if is really a "compilation" bug:

  test_quad.m ............................................ ABNORMAL RETURN FROM 
DQAGP

Dan