[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Bug-gnucap] patches for gnucap
From: |
Al Davis |
Subject: |
Re: [Bug-gnucap] patches for gnucap |
Date: |
Thu, 25 Mar 2004 23:42:08 -0500 |
User-agent: |
KMail/1.6.1 |
On Monday 22 March 2004 06:16 am, Dan McMahill wrote:
> Hope this is the right place to send this. I recently had
> a problem where I needed to include math.h in one of the gnucap
> source files to build gnucap. Also, I found a bug in one of the
> scripts.
>
> patches are attached.
Yes this is the place for bug reports.
I need more information to properly fix it. I can make some
assumptions, but I would rather ask you to say it directly.
1. You sent a patch. What are the symptoms? I am guessing: In
l_ftos.cc, "isnan" and/or "isinf" undefined on compile. In the test
script, it doesn't accept "==".
2. What version of gnucap? I am guessing: it must be 0.34, because that
is the only one with the "usage" check in the test script, but I
thought the "isnan" was fixed.
3. What computer, OS, etc..? I am guessing: based on other
communication with you, a BSD system, but not FreeBSD. Is it NetBSD or
OpenBSD???
4. What compiler?
I will change the test script. Linux uses "bash" when you ask for "sh".
I verified that "dash" doesn't work, and that your patch fixes it. I
believe that the Linux "dash" is the standard BSD shell. I also
verified the problem and your fix on a Sun. Thank you.
On <math.h> ..... This is not a standard C++ header. It's a standard C
header. The preferred C++ header is <cmath>. Can you check this for
me? If it is "isnan" and "isinf", why does the configure script not
catch it? <cmath> is not included explicitly because it is included by
<complex> on my system. If <math.h> fixes it, why do the trig
functions work? I need more info on this one.
al.