Re: 1.3.2: signgam problem still present in plotutils, gnuplot etc.

[Christopher Faylor <cgf at redhat dot com>]

On Wed, Jun 13, 2001 at 05:09:38PM -0400, J. Johnston wrote:
>Christopher Faylor wrote:
>> On Tue, Jun 12, 2001 at 09:35:38PM +0200, Frank Wuebbeling wrote:
>>>>Hmm.  I don't recall anyone submitting a patch for this.
>>>
>>>The mailing on the CYGWIN list I referred to had a patch, to which
>>>several people agreed it worked.  It almost also did the job for
>>>plotutils, meaning it revealed a problem in plotutils, removing that
>>>one made it work (the problem was that plotutils assumes that signgam
>>>is a variable, and redeclares it - which fails if it is not and is
>>>obsolete anyway since math.h is responsible for declaring it).
>>>
>>>Again, the page was
>>>http://www.cygwin.com/ml/cygwin/2001-03/msg01837.html.  I quote the
>>>main idea.  I have no idea whatsoever what the change really does, so I
>>>don't want to submit a patch - I just propose reconsidering the old
>>>posting which had one.  I incorporated the fix in math.h.  Here's the
>>>original posting:
>>
>>Ok.  That's not a patch.  I have no idea where this definition would
>>go.  I'm sure that I could probably track it down but, if it is
>>important to you, you should start a thread on the newlib mailing list,
>>asking for the change since this is apparently a newlib construction.
>>
>>I've cc'ed newlib and redirected the discussion there.
>
>Newlib changed math.h in January to have an errno-like solution for
>signgam whereby a function gets called.  Is math.h up to date in the
>user's Cygwin and is it being included by the source code in question?
>See: http://sources.redhat.com/ml/newlib-cvs/2001-q1/msg00013.html for
>details on the patch applied in January.

We're certainly using the most recent versions of newlib in the cygwin
releases.  I can't speak to the user's specific problem, however.

cgf