gnuplot-beta

Gnuplot 4.4.2 (fixes some problems with 4.4.1)

[sfeam (E. Merritt) <eam...@gm...>]

I have just uploaded to SourceForge a source tarball and documentation for 
Gnuplot version 4.4.2.

Yes, this follows version 4.4.1 very quickly.  
It fixes three problems that were reported immediately after the 4.4.1
release. Two of these seemed serious enough to justify an immediate fix.

1) The color specification syntax "lc N" was broken in 4.4.1.
   This was a regression from 4.4.0

2) A "refresh" command either from the command line or implicitly from
   mousing would cause a segfault if a previous "replot" command had
   been interrupted.

3) There was an error in the clipping of filled curves when a single
   polygonal fill segment straddled one corner of the plot box.


My apologies for missing a report from Christoph Junghans that pointed
out the "lc N" regression before 4.4.1 was actually released.  
Also my apologies to Tatsuro Matsuoka, whose Windows binaries for
4.4.1 are now out of date only a few days after they were announced.

		Ethan Merritt (gnuplot development team)

Re: Gnuplot 4.4.2 (fixes some problems with 4.4.1)

[Tatsuro M. <tma...@ya...>]

Hello

--- "sfeam (Ethan Merritt)"  wrote:

> Also my apologies to Tatsuro Matsuoka, whose Windows binaries for
> 4.4.1 are now out of date only a few days after they were announced.

OK. 

I will close current snapshots immediately and prepare new binaries for 4.4.2.

Regards

Tatsuro  

--------------------------------------
Get the new Internet Explorer 8 optimized for Yahoo! JAPAN
http://pr.mail.yahoo.co.jp/ie8/

Re: Gnuplot 4.4.2 (fixes some problems with 4.4.1)

[Tatsuro M. <tma...@ya...>]

Hello

--- Tatsuro MATSUOKA  wrote:

> Hello
> 
> --- "sfeam (Ethan Merritt)"  wrote:
> 
> > Also my apologies to Tatsuro Matsuoka, whose Windows binaries for
> > 4.4.1 are now out of date only a few days after they were announced.
> 
> OK. 
> 
> I will close current snapshots immediately and prepare new binaries for 4.4.2.

I have uploaded new binaries for windows, cygwin and djgpp.
http://www.tatsuromatsuoka.com/gnuplot/Eng/gp442/

Regards

Tatsuro


--------------------------------------
Get the new Internet Explorer 8 optimized for Yahoo! JAPAN
http://pr.mail.yahoo.co.jp/ie8/

Re: Gnuplot 4.4.2 (fixes some problems with 4.4.1)

[sfeam (E. Merritt) <eam...@gm...>]

On Monday 27 September 2010, Tatsuro MATSUOKA wrote:
> 
> I have uploaded new binaries for windows, cygwin and djgpp.
> http://www.tatsuromatsuoka.com/gnuplot/Eng/gp442/

Thank you very much for your very quick preparation of these binaries.

I tried to test the binaries for gp442win32
using wine running on linux.  

- The wgnuplot.exe binary seems to work well.

- Unforunately, the gnuplot.exe binary does not respond to <cr>.

Is this a failure of the wine emulation, or is it a problem on true Windows also?
Although this appears like the same failure that was present in 4.5 recently,
I do not understand how that could happen.  The change the caused the error
in 4.5 was never applied to 4.4.

I did not test the gp442djpp binary because it can not run under linux.

Should I upload all of these to SourceForge, or do we need a fix
for the win32 gnuplot.exe?

    thanks,

		Ethan

> Regards
> 
> Tatsuro
> 
> 
> --------------------------------------
> Get the new Internet Explorer 8 optimized for Yahoo! JAPAN
> http://pr.mail.yahoo.co.jp/ie8/
> 

Re: Gnuplot 4.4.2 (fixes some problems with 4.4.1)

[Tatsuro M. <tma...@ya...>]

Hello

--- "sfeam (Ethan Merritt)"  wrote:

> On Monday 27 September 2010, Tatsuro MATSUOKA wrote:
> > 
> > I have uploaded new binaries for windows, cygwin and djgpp.
> > http://www.tatsuromatsuoka.com/gnuplot/Eng/gp442/
> 
> Thank you very much for your very quick preparation of these binaries.
> 
> I tried to test the binaries for gp442win32
> using wine running on linux.  
> 
> - The wgnuplot.exe binary seems to work well.
> 
> - Unforunately, the gnuplot.exe binary does not respond to <cr>.
> 
> Is this a failure of the wine emulation, or is it a problem on true Windows also?
> Although this appears like the same failure that was present in 4.5 recently,
> I do not understand how that could happen.  The change the caused the error
> in 4.5 was never applied to 4.4.

For me gnuplot.exe (4.4.2) accepts "Enter" correctly.
*********************
        G N U P L O T
        Version 4.4 patchlevel 2
        last modified Wed Sep 22 12:10:34 PDT 2010
        System: MS-Windows 32 bit

        Copyright (C) 1986-1993, 1998, 2004, 2007-2010
        Thomas Williams, Colin Kelley and many others

        gnuplot home:     http://www.gnuplot.info
        faq, bugs, etc:   type "help seeking-assistance"
        immediate help:   type "help"
        plot window:      hit 'h'

Terminal type set to 'wxt'
gnuplot>
gnuplot> plot sin(x)
gnuplot>
*************************:
And correct sin(x) graph appeared in wxt terminal. 

All demo scripts worked correct at the demo directory with some files generated in the make process. 
It was confirmed using  gnuplot.exe.

For DJGPP  I  only did simple test 

gnuplot> plot sin(x) 
It worked correctly.  
For DjGPP, the some demos does work so that I did not carried out all.dem test. 

But I will try .....

Oops! Djgpp version seems not to support NaN

******************** file candlesticks.dem ********************
Hit return to continue
Hit return to continue
Hit return to continue
         "candlesticks.dem", line 32: undefined variable: NaN

The same error  occurred in gnuplot 4.5 cvs.
But this might be the limitation djgpp environments.
At the moment, in my opnion, it is better to ignore this issue and consider the treatment in the cvs
release.

(If I replace 'NaN' by '1/0' ,"candlesticks.dem" goes well.)


> I did not test the gp442djpp binary because it can not run under linux.
> 
> Should I upload all of these to SourceForge, or do we need a fix
> for the win32 gnuplot.exe?

Before uploading them on SourceForge, wgnuplot.hlp should be modified.
As I wrote, my help workshop does not work correctly.
Therefore Petr kindly generated it. However it lacks the explanation of the cairo based terminals.

The gnuplot.rtf generated by make process includes explanation of only linked terminals.
I suppose that Petr's wgnuplot.hlp is genarated with his own environments.  
Perhaps his build environments lacks libraries for the cairo based terminals.

I have uploaded gnuplot.rtf for MS-help workshop.  If it is complied by the correct version of the
MS-help workshop, the help file will be complete.

Regards

Tatsuro  


--------------------------------------
Get the new Internet Explorer 8 optimized for Yahoo! JAPAN
http://pr.mail.yahoo.co.jp/ie8/

Re: Gnuplot 4.4.2 (fixes some problems with 4.4.1)

[Ethan M. <merritt@u.washington.edu>]

On Tuesday 28 September 2010 09:58:57 pm Tatsuro MATSUOKA wrote:
> Oops! Djgpp version seems not to support NaN
> 
> The same error  occurred in gnuplot 4.5 cvs.
> But this might be the limitation djgpp environments.
> At the moment, in my opnion, it is better to ignore this issue and consider the treatment in the cvs
> release.

If you change the configuration file config.dj2 does it build properly?

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
--- config.dj2.old      2010-09-29 10:21:42.000000000 -0700
+++ config.dj2  2010-09-29 10:22:06.000000000 -0700
@@ -142,7 +142,7 @@
 #endif

 /* Define to 1 if you have the `isnan' function. */
-/* #undef HAVE_ISNAN */
+#define HAVE_ISNAN

 /* Define if you use have kpsexpand (TeX). */
 /* #undef HAVE_KPSEXPAND */
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

I think this should work, because I now see that there is a call to isnan()
in datafile.c that is not protected by #ifdef HAVE_ISNAN.  
So if djpp can compile datafile.c, then it must support isnan().

However, I am not certain that the variable NaN will be properly initialized.
So please check the output of "print NaN" after building.

	Ethan

-- 
Ethan A Merritt
Biomolecular Structure Center,  K-428 Health Sciences Bldg
University of Washington, Seattle 98195-7742

Re: Gnuplot 4.4.2 (fixes some problems with 4.4.1)

[Ethan M. <merritt@u.washington.edu>]

On Wednesday 29 September 2010 10:32:23 am Ethan Merritt wrote:
> If you change the configuration file config.dj2 does it build properly?
> 
> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
> --- config.dj2.old      2010-09-29 10:21:42.000000000 -0700
> +++ config.dj2  2010-09-29 10:22:06.000000000 -0700
> @@ -142,7 +142,7 @@
>  #endif
> 
>  /* Define to 1 if you have the `isnan' function. */
> -/* #undef HAVE_ISNAN */
> +#define HAVE_ISNAN
> 
>  /* Define if you use have kpsexpand (TeX). */
>  /* #undef HAVE_KPSEXPAND */
> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
> 
> I think this should work, because I now see that there is a call to isnan()
> in datafile.c that is not protected by #ifdef HAVE_ISNAN.  
> So if djpp can compile datafile.c, then it must support isnan().

I made a small mistake.
The unprotected call to isnan() is in graphics.c, not datafile.c.
But still I think it must be true that djpp supports isnan().

	Ethan 

Re: Gnuplot 4.4.2 (fixes some problems with 4.4.1)

[Hans-Bernhard B. <HBB...@t-...>]

On 29.09.2010 19:32, Ethan Merritt wrote:

> If you change the configuration file config.dj2 does it build properly?
>
> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
> --- config.dj2.old      2010-09-29 10:21:42.000000000 -0700
> +++ config.dj2  2010-09-29 10:22:06.000000000 -0700
> @@ -142,7 +142,7 @@
>   #endif
>
>   /* Define to 1 if you have the `isnan' function. */
> -/* #undef HAVE_ISNAN */
> +#define HAVE_ISNAN

FWIW 'configure' doesn't believe DJGPP has isnan().  The autoconf test 
for it yields false because it looks for a function, not a macro, but 
it's done without -lm.  So it finds neither the macro, nor the function 
in libm.a.

I can't currently perform any real tests of my own using DJGPP --- gcc 
4.4.4 consistently crashes on me in term.c with a heap fault.

Re: Gnuplot 4.4.2 (fixes some problems with 4.4.1)

[Tatsuro M. <tma...@ya...>]

Hello

I have tried Ethan's modification to config.dj2.
It seems to work fine on gnuplot 4.4.2 and 4.5.

Before Patch
gnuplot> print NaN
         undefined variable: NaN
After patch
gnuplot> print NaN
0.0

and candlesticks.dem works without error.

Thanks!!

Tatsuro
--- Hans-Bernhard Br将モker  wrote:

> On 29.09.2010 19:32, Ethan Merritt wrote:
> 
> > If you change the configuration file config.dj2 does it build properly?
> >
> > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
> > --- config.dj2.old      2010-09-29 10:21:42.000000000 -0700
> > +++ config.dj2  2010-09-29 10:22:06.000000000 -0700
> > @@ -142,7 +142,7 @@
> >   #endif
> >
> >   /* Define to 1 if you have the `isnan' function. */
> > -/* #undef HAVE_ISNAN */
> > +#define HAVE_ISNAN
> 
> FWIW 'configure' doesn't believe DJGPP has isnan().  The autoconf test 
> for it yields false because it looks for a function, not a macro, but 
> it's done without -lm.  So it finds neither the macro, nor the function 
> in libm.a.
> 
> I can't currently perform any real tests of my own using DJGPP --- gcc 
> 4.4.4 consistently crashes on me in term.c with a heap fault.
> 
> 


--------------------------------------
Get the new Internet Explorer 8 optimized for Yahoo! JAPAN
http://pr.mail.yahoo.co.jp/ie8/

Re: Gnuplot 4.4.2 (fixes some problems with 4.4.1)

[Ethan M. <merritt@u.washington.edu>]

On Thursday 30 September 2010 04:24:36 pm Tatsuro MATSUOKA wrote:
> Hello
> 
> I have tried Ethan's modification to config.dj2.
> It seems to work fine on gnuplot 4.4.2 and 4.5.
> 
> Before Patch
> gnuplot> print NaN
>          undefined variable: NaN
> After patch
> gnuplot> print NaN
> 0.0

Unfortunately, that shows isnan() is found but the initialization of
the user variable NaN is not working correctly.
NaN should print as "NaN" or "nan" rather than 0.0.

Could you please try this patch to see if the initialization
method used for Microsoft C also works for DJGPP?

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
--- gnuplot-cvs/src/stdfn.c 2010-07-30 12:11:40.000000000 -0700
+++ gnuplot-djgpp/src/stdfn.c     2010-09-30 18:12:55.000000000 -0700
@@ -458,7 +458,7 @@
 double
 not_a_number(void)
 {
-#ifdef __MSC__
+#if defined(__MSC__) || defined(DJGPP) || defined(__DJGPP__)
        unsigned long lnan[2]={0xffffffff, 0x7fffffff};
     return *( double* )lnan;
 #else
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

	thanks,

		Ethan



> 
> and candlesticks.dem works without error.
> 
> Thanks!!
> 
> Tatsuro
> --- Hans-Bernhard Br将モker  wrote:
> 
> > On 29.09.2010 19:32, Ethan Merritt wrote:
> > 
> > > If you change the configuration file config.dj2 does it build properly?
> > >
> > > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
> > > --- config.dj2.old      2010-09-29 10:21:42.000000000 -0700
> > > +++ config.dj2  2010-09-29 10:22:06.000000000 -0700
> > > @@ -142,7 +142,7 @@
> > >   #endif
> > >
> > >   /* Define to 1 if you have the `isnan' function. */
> > > -/* #undef HAVE_ISNAN */
> > > +#define HAVE_ISNAN
> > 
> > FWIW 'configure' doesn't believe DJGPP has isnan().  The autoconf test 
> > for it yields false because it looks for a function, not a macro, but 
> > it's done without -lm.  So it finds neither the macro, nor the function 
> > in libm.a.
> > 
> > I can't currently perform any real tests of my own using DJGPP --- gcc 
> > 4.4.4 consistently crashes on me in term.c with a heap fault.
> > 
> > 
> 
> 
> --------------------------------------
> Get the new Internet Explorer 8 optimized for Yahoo! JAPAN
> http://pr.mail.yahoo.co.jp/ie8/
> 

-- 
Ethan A Merritt
Biomolecular Structure Center,  K-428 Health Sciences Bldg
University of Washington, Seattle 98195-7742

Re: Gnuplot 4.4.2 (fixes some problems with 4.4.1)

[Tatsuro M. <tma...@ya...>]

Hello

--- Ethan Merritt  wrote:

> On Thursday 30 September 2010 04:24:36 pm Tatsuro MATSUOKA wrote:
> > Hello
> > 
> > I have tried Ethan's modification to config.dj2.
> > It seems to work fine on gnuplot 4.4.2 and 4.5.
> > 
> > Before Patch
> > gnuplot> print NaN
> >          undefined variable: NaN
> > After patch
> > gnuplot> print NaN
> > 0.0
> 
> Unfortunately, that shows isnan() is found but the initialization of
> the user variable NaN is not working correctly.
> NaN should print as "NaN" or "nan" rather than 0.0.
> 
> Could you please try this patch to see if the initialization
> method used for Microsoft C also works for DJGPP?
> 
> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
> --- gnuplot-cvs/src/stdfn.c 2010-07-30 12:11:40.000000000 -0700
> +++ gnuplot-djgpp/src/stdfn.c     2010-09-30 18:12:55.000000000 -0700
> @@ -458,7 +458,7 @@
>  double
>  not_a_number(void)
>  {
> -#ifdef __MSC__
> +#if defined(__MSC__) || defined(DJGPP) || defined(__DJGPP__)
>         unsigned long lnan[2]={0xffffffff, 0x7fffffff};
>      return *( double* )lnan;
>  #else
> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

Ooops!

On windows version built by gcc on MinGW also gives
gnuplot> pr NaN
0.0               

I will research NaN related topic using search engine for gcc on MinGW (and) DJGPP.
After that I will consider application of the patch. 

Regards

Tatsuro


--------------------------------------
Get the new Internet Explorer 8 optimized for Yahoo! JAPAN
http://pr.mail.yahoo.co.jp/ie8/

Re: Gnuplot 4.4.2 (fixes some problems with 4.4.1)

[Tatsuro M. <tma...@ya...>]

Hello

--- Tatsuro MATSUOKA wrote:

> Hello
> 
> --- Ethan Merritt  wrote:
> 
> > On Thursday 30 September 2010 04:24:36 pm Tatsuro MATSUOKA wrote:
> > > Hello
> > > 
> > > I have tried Ethan's modification to config.dj2.
> > > It seems to work fine on gnuplot 4.4.2 and 4.5.
> > > 
> > > Before Patch
> > > gnuplot> print NaN
> > >          undefined variable: NaN
> > > After patch
> > > gnuplot> print NaN
> > > 0.0
> > 
> > Unfortunately, that shows isnan() is found but the initialization of
> > the user variable NaN is not working correctly.
> > NaN should print as "NaN" or "nan" rather than 0.0.
> > 
> > Could you please try this patch to see if the initialization
> > method used for Microsoft C also works for DJGPP?
> > 
> > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
> > --- gnuplot-cvs/src/stdfn.c 2010-07-30 12:11:40.000000000 -0700
> > +++ gnuplot-djgpp/src/stdfn.c     2010-09-30 18:12:55.000000000 -0700
> > @@ -458,7 +458,7 @@
> >  double
> >  not_a_number(void)
> >  {
> > -#ifdef __MSC__
> > +#if defined(__MSC__) || defined(DJGPP) || defined(__DJGPP__)
> >         unsigned long lnan[2]={0xffffffff, 0x7fffffff};
> >      return *( double* )lnan;
> >  #else
> > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
> 
> Ooops!
> 
> On windows version built by gcc on MinGW also gives
> gnuplot> pr NaN
> 0.0               
> 
> I will research NaN related topic using search engine for gcc on MinGW (and) DJGPP.
> After that I will consider application of the patch. 
> 
> Regards
> 
> Tatsuro

Hello

I have searched the information of the DJGPP 

For version 2.04
http://www.delorie.com/djgpp/doc/kb/kb_6.html

************************
The macro NAN was added.
The functions strtod, strtof, strtold and _strtold now understand "Inf", "Infinity", "NaN", "NaN()",
"NaN(hex-number)" and any variations of case in the input string.

New functions snprintf and vsnprintf added.
***********************

However, version 2.04 is the beta release.
(Last update is 30-Nov-2003 for 2.04).

Application is still updated on the 2.03.
(The gcc has been updated to 4.4.4!  as Hans suggested.)

Thinking to the above, it is better to apply you patch for DJGPP.

Please wait for a while after I will carry out the test the patch

For the MinGW, the different way seem to be used for NaN.  I need to search further. 

Regards

Tatsuro

--------------------------------------
Get the new Internet Explorer 8 optimized for Yahoo! JAPAN
http://pr.mail.yahoo.co.jp/ie8/

Re: Gnuplot 4.4.2 (fixes some problems with 4.4.1)

[Tatsuro M. <tma...@ya...>]

Hello

--- Tatsuro MATSUOKA  wrote:
> I have searched the information of the DJGPP 
> 
> For version 2.04
> http://www.delorie.com/djgpp/doc/kb/kb_6.html
> 
> ************************
> The macro NAN was added.
> The functions strtod, strtof, strtold and _strtold now understand "Inf", "Infinity", "NaN",
> "NaN()",
> "NaN(hex-number)" and any variations of case in the input string.
> 
> New functions snprintf and vsnprintf added.
> ***********************
> 
> However, version 2.04 is the beta release.
> (Last update is 30-Nov-2003 for 2.04).
> 
> Application is still updated on the 2.03.
> (The gcc has been updated to 4.4.4!  as Hans suggested.)
> 
> Thinking to the above, it is better to apply you patch for DJGPP.
> 
> Please wait for a while after I will carry out the test the patch

With the two patches,
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
--- config.dj2.old      2010-09-29 10:21:42.000000000 -0700
+++ config.dj2  2010-09-29 10:22:06.000000000 -0700
@@ -142,7 +142,7 @@
  #endif
   /* Define to 1 if you have the `isnan' function. */
-/* #undef HAVE_ISNAN */
+#define HAVE_ISNAN
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
--- gnuplot-cvs/src/stdfn.c 2010-07-30 12:11:40.000000000 -0700
+++ gnuplot-djgpp/src/stdfn.c     2010-09-30 18:12:55.000000000 -0700
@@ -458,7 +458,7 @@
 double
 not_a_number(void)
 {
-#ifdef __MSC__
+#if defined(__MSC__) || defined(DJGPP) || defined(__DJGPP__)
        unsigned long lnan[2]={0xffffffff, 0x7fffffff};
     return *( double* )lnan;
 #else
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
NaN issue for DJGPP seems to be solved.


Terminal type set to 'svga'
gnuplot> pr NaN
NaN

gnuplot> cd '../demo'
gnuplot> load 'candlesticks.dem'
Hit return to continue
Hit return to continue
Hit return to continue
Hit return to continue
Hit return to continue
Hit return to continue

candlesticks.dem went well.

Thanks!!

Regards

Tatsuro

--------------------------------------
Get the new Internet Explorer 8 optimized for Yahoo! JAPAN
http://pr.mail.yahoo.co.jp/ie8/

Re: Gnuplot 4.4.2 (fixes some problems with 4.4.1)

[Tatsuro M. <tma...@ya...>]

Hello 

The test shown in the previous mail was done by cvs source at 2010-09-30.

Regards

Tatsuro


--- Tatsuro MATSUOKA  wrote:

> Hello
> 
> --- Tatsuro MATSUOKA  wrote:
> > I have searched the information of the DJGPP 
> > 
> > For version 2.04
> > http://www.delorie.com/djgpp/doc/kb/kb_6.html
> > 
> > ************************
> > The macro NAN was added.
> > The functions strtod, strtof, strtold and _strtold now understand "Inf", "Infinity", "NaN",
> > "NaN()",
> > "NaN(hex-number)" and any variations of case in the input string.
> > 
> > New functions snprintf and vsnprintf added.
> > ***********************
> > 
> > However, version 2.04 is the beta release.
> > (Last update is 30-Nov-2003 for 2.04).
> > 
> > Application is still updated on the 2.03.
> > (The gcc has been updated to 4.4.4!  as Hans suggested.)
> > 
> > Thinking to the above, it is better to apply you patch for DJGPP.
> > 
> > Please wait for a while after I will carry out the test the patch
> 
> With the two patches,
> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
> --- config.dj2.old      2010-09-29 10:21:42.000000000 -0700
> +++ config.dj2  2010-09-29 10:22:06.000000000 -0700
> @@ -142,7 +142,7 @@
>   #endif
>    /* Define to 1 if you have the `isnan' function. */
> -/* #undef HAVE_ISNAN */
> +#define HAVE_ISNAN
> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
> --- gnuplot-cvs/src/stdfn.c 2010-07-30 12:11:40.000000000 -0700
> +++ gnuplot-djgpp/src/stdfn.c     2010-09-30 18:12:55.000000000 -0700
> @@ -458,7 +458,7 @@
>  double
>  not_a_number(void)
>  {
> -#ifdef __MSC__
> +#if defined(__MSC__) || defined(DJGPP) || defined(__DJGPP__)
>         unsigned long lnan[2]={0xffffffff, 0x7fffffff};
>      return *( double* )lnan;
>  #else
> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
> NaN issue for DJGPP seems to be solved.
> 
> 
> Terminal type set to 'svga'
> gnuplot> pr NaN
> NaN
> 
> gnuplot> cd '../demo'
> gnuplot> load 'candlesticks.dem'
> Hit return to continue
> Hit return to continue
> Hit return to continue
> Hit return to continue
> Hit return to continue
> Hit return to continue
> 
> candlesticks.dem went well.
> 
> Thanks!!
> 
> Regards
> 
> Tatsuro
> 
> --------------------------------------
> Get the new Internet Explorer 8 optimized for Yahoo! JAPAN
> http://pr.mail.yahoo.co.jp/ie8/
> 


--------------------------------------
Get the new Internet Explorer 8 optimized for Yahoo! JAPAN
http://pr.mail.yahoo.co.jp/ie8/