Re: [Gcl-devel] help fixing edit_double, failure to print 1d-6 for gcl-2.6.8

[David Billinghurst]

That works.  Thanks.

Happy to try and answer any mingw questions, but I am not an expert.  In 
fact, I have been spectacularly unsuccessful in hacking gcl on windows - 
all I have managed to do is keep gcl-2.6.8pre - and now 2.6.8 - running 
with gcc-3.3.1 on Windows XP.

David

On 8/10/2013 11:41 PM, Camm Maguire wrote:
> Greetings!  I think this should do it:
>
> --- /home/camm/debian/gcl/gclttt/o/print.d      2013-01-15 19:44:49.000000000 
> +0000
> +++ print.d     2013-09-28 14:53:26.000000000 +0000
> @@ -388,6 +390,10 @@
>   
>     truncate_double(b,d,n!=7);
>   
> +  if (isdigit(b[0])) {
> +    b[1]=b[0];
> +    (*ep)++;
> +  }
>     if (b[2]=='0') (*ep)++;
>     b[2] = b[1];
>     p = b + 2;
>
> Please let me know if problems persist.
>
> BTW, can you field a few mingw questions?
>
> Take care,
>
> David Billinghurst <address@hidden> writes:
>
> > The current maxima 5.31.2 release for windows is built with cvs
> > gcl-2.6.8pre from 2013-01-01.  The previous release used gcl-2.6.8,
> > but this has a bug where 1e-6 is printed as 0.0.
> >
> > I tried backporting Camm's patch to 2.6.20pre
> > http://git.savannah.gnu.org/cgit/gcl.git/commit/?h=Version_2_6_10pre&id=b25201deb6e0303e92adb105a610f367745032d4
> > but the resulting maxima has many testsuite failures.
> >
> > I have identified the patch to 2.6.8pre that introduced the bug
> > http://cvs.savannah.gnu.org/viewvc/gcl/o/print.d?root=gcl&r1=1.15.4.1.2.2.12.1.2.1.2.1&r2=1.15.4.1.2.2.12.1.2.1.2.2&pathrev=Version_2_6_8pre
> >
> > I would like to fix 2.6.8 and get a 2.6.8-1 (or whatever) release so
> > that maxima is can again be built with a released version of gcl.
> > Unfortunately is isn't immediately obvious to me what the fix is.  I
> > can probably work it out, but it is easier to ask :-).
> >
> > David