gnuplot-beta — Developers and beta testers of new gnuplot code

Re: The bind command is broken in 6.1

[Ethan A M. <me...@uw...>]

On Thu, Jul 16, 2026 at 9:47 AM Juhász Péter <pet...@gm...>
wrote:

>
>
> executive summary:
> the canonical form of the bind command (e.g. `bind a "replot"`) does
> not work, probably since 2024, because of changeset 96d580c508.
>

Yes.  My fault.  Fixed now.
Thanks for the report.

So here's the story.  I regularly run the full set of demos with saved
output
so that I can compare the plots produced before/after any code changes.
The tests can be run under valgrind or asan to detect memory access
violations. To the extent that the demos actually exercise all program
options
and the resulting code paths, it does a reasonable job of finding parsing
errors, run-time errors, memory leaks, and so on in addition to detecting
incorrectly generated plots.

Unfortunately these tests do not exercise all terminal-specific code, and
do not test mousing or other interactive use of the program.
So they did not, and could not, catch this breakage in the "bind" command.

If anyone has ideas of how to increase the coverage of automated testing,
please speak up.  It would also be helpful if you know of any program
options or plot modes that are not part of the demo collection, and hence
not tested.

 - Ethan




> explanation:
> Gnuplot is one of the few programs I still compile from source,
> recompiling rarely, usually only when I change computers or upgrade the
> OS. That's why I only noticed now that a company-internal analysis
> script that relies heavily on `bind` stopped working after recompiling
> the current gnuplot source.
>
> The simplest form of `bind`, e.g. `bind a "replot"`, as given in the
> manual, does not do anything, it just fails silently.
>
> I believe the breakage was introduced in commit 96d580c508,
> specifically this bit:
>
> @ -1037,8 +1037,7 @@ bind_command()
>         char *first = gp_input_line + token[c_token].start_index;
>         int size = strcspn(first, " \";");
>         lhs = gp_alloc(size + 1, "bind_command->lhs");
> -       strncpy(lhs, first, size);
> -       lhs[size] = '\0';
> +       safe_strncpy(lhs, first, size);
>         FPRINTF((stderr,"Got bind unquoted lhs = \"%s\"\n",lhs));
>         while (gp_input_line + token[c_token].start_index < first+size)
>             c_token++;
>
> `safe_strncpy` is aliased to `strlcpy`, if available, and I seem to
> have it, but in any case, this is not an equivalent change.
>
> In the case of a naked single-character first argument to bind, size
> will be 1 in the code above. `strlcpy` copies at most `size` (so, 1)
> bytes, and sets the last (that is, the only) byte to null.
>
> Then the code that wants to set up the binding receives an empty
> string, which causes it to silently fail.
>
> This incorrect behavior can be also seen with the command `bind RR
> "replot"`, which is nonsensical and should be an error, but instead it
> sets up the binding for "R".
>
> Luckily, the quoted case `bind "R" "replot"` still works.
>
> There may be other instances of this incorrect usage of safe_strncpy, I
> haven't checked.
>
> best regards,
>
> Peter Juhasz
>
>
> _______________________________________________
> gnuplot-beta mailing list
> gnu...@li...
> Membership management via:
> https://urldefense.com/v3/__https://lists.sourceforge.net/lists/listinfo/gnuplot-beta__;!!K-Hz7m0Vt54!hwk5Sr7cG1SlpX05GVACsAMQf-E35prPcaVkLTXV_8BuQRntHCZc_iqZwwDXW9GpogtRy8M-YMpd9o-wp5MdDT0$
>


-- 
Ethan A Merritt
Department of Biochemistry
University of Washington, Seattle

The bind command is broken in 6.1

[Juhász P. <pet...@gm...>]

Hi,

executive summary:
the canonical form of the bind command (e.g. `bind a "replot"`) does
not work, probably since 2024, because of changeset 96d580c508.

explanation:
Gnuplot is one of the few programs I still compile from source,
recompiling rarely, usually only when I change computers or upgrade the
OS. That's why I only noticed now that a company-internal analysis
script that relies heavily on `bind` stopped working after recompiling
the current gnuplot source.

The simplest form of `bind`, e.g. `bind a "replot"`, as given in the
manual, does not do anything, it just fails silently.

I believe the breakage was introduced in commit 96d580c508,
specifically this bit:

@ -1037,8 +1037,7 @@ bind_command()
        char *first = gp_input_line + token[c_token].start_index;
        int size = strcspn(first, " \";");
        lhs = gp_alloc(size + 1, "bind_command->lhs");
-       strncpy(lhs, first, size);
-       lhs[size] = '\0';
+       safe_strncpy(lhs, first, size);
        FPRINTF((stderr,"Got bind unquoted lhs = \"%s\"\n",lhs));
        while (gp_input_line + token[c_token].start_index < first+size)
            c_token++;

`safe_strncpy` is aliased to `strlcpy`, if available, and I seem to
have it, but in any case, this is not an equivalent change.

In the case of a naked single-character first argument to bind, size
will be 1 in the code above. `strlcpy` copies at most `size` (so, 1)
bytes, and sets the last (that is, the only) byte to null.

Then the code that wants to set up the binding receives an empty
string, which causes it to silently fail.

This incorrect behavior can be also seen with the command `bind RR
"replot"`, which is nonsensical and should be an error, but instead it
sets up the binding for "R".

Luckily, the quoted case `bind "R" "replot"` still works.

There may be other instances of this incorrect usage of safe_strncpy, I
haven't checked.

best regards,

Peter Juhasz

Re: Pre-release testing version of gnuplot 6.0.5

[Ethan A M. <me...@uw...>]

On Wed, Jul 15, 2026 at 10:41 AM Bastian Märkisch <bma...@we...> wrote:

> Builds fine on Windows with MSYS2/CLANG64 and MSYS2/CLANGARM64.
>
> There's only a minor type issue in internal.c I noticed only know:  On
> Windows, intgr_t is a "long long int", not a "long int" (which would
> have only 32 bits). Hence the labs() call on line 1174 should really be
> llabs().  Is that fine for other platforms, too? Or should I add an
> ugly  #ifdef?
>
>
That call is already llabs() in version 6.1, so it should be OK for 6.0
also.

But, thinking about it, the type of intgr_t is configuration-dependent.
To assume it is specifically either (long) or (long long) might cause the
equivalent
compiler complaint for someone else.
I think the problem is avoidable by changing this to
        intgr_t tmag = (a.v.int_val < 0) ? -a.v.int_val : a.v.int_val;

thanks,
          Ethan


> Best wishes,
>
>    Bastian
>
> Am 10.07.2026 um 05:56 schrieb Ethan A Merritt:
> > I have placed a testing version of the source release
> > package for gnuplot 6.0.4 on the "testing" section of the
> > SourceForge download site.
> >
> >
> https://urldefense.com/v3/__https://sourceforge.net/projects/gnuplot/files/gnuplot/testing/__;!!K-Hz7m0Vt54!hICq3ODdIA1Zl3EhVy-KLGRkCS1rUe6lf7K7i7xJonlbqmlHMTa2K5cdPyoYdchWBtnkavSckVKW1EZ4MA$
> >
> > My tentative plan is to announce the actual release in about
> > two weeks (end of July).  That can be pushed back if you find
> > any problems or want more time to test on other platforms.
> >
> > Brief release notes below, or see
> >       gnuplot.sourceforge.net/ReleaseNotes_6_0_5.html
> >
> > %%%%%%%%%%    6.0.5 Release Notes July 2026   %%%%%%%%%%%%%%%%
> >
> > New features
> >
> > * Full mousing support for multiplots
> >    - all panels report mouse coordinates
> >    - all panels respond to pan/zoom/rotate mouse movement and hotkeys
> >    - "replot" reexecutes multiplot commands starting from current state
> >    - "remultiplot" restores the original state before reexecuting
> multiplot
> >    - interactive mousing remains active after ^C or "reset"
> >
> > * Unicode escape sequences \U+xxxx are accepted in more contexts
> >    - escape sequences are expanded in "noenhanced" strings
> >    - escape sequences are expanded during substring evaluation
> >    - escape sequences are expanded in string functions strstrt() and
> strlen()
> >
> > * New built-in function prod [<var> = <start> : <end>] <expression>
> >    - implements the sequential product operation
> >               prod [i=a:b] f(i)   =   f(a) * f(a+1) * ... * f(b)
> >
> > * Data files in EUC-JP encoding can be processed from other environments
> >    - If running in a EUC-JP locale, use "set encoding locale"
> >    - If running in a different locale, use "set encoding eucjp"
> >    - In either case, this affects only input data.  It does not change
> >      the encoding (usually UTF-8) written to plots on output.
> >
> > * New coloring option for contourfill
> >    - Arbitrary discrete color ranges can be define without using a
> special palette
> >          set contourfill defined [zmin1:zmax1] color1, [zmin2:zmax2]
> color2, ...
> >
> > * Allow "string"[i] as shorthand for "string"[i:i]
> >
> > Changed behaviour
> >
> > * Trailing whitespace on the command line is now ignored on input.
> >    This means that the line-continuation character '\' is effective even
> >    if it is followed by an unintentional blank.
> >
> > * Octal escape sequences must conform to the documented form \ooo
> >    - E.g. \7 or \0106 or \456 or \0 are now rejected
> >
> > * "plot with labels" now accepts the "noenhanced" keyword
> >
> > Bug fixes
> >
> > * qt: Slow font initialization on Windows                       #Bug#
> 2316 2476
> > * gd: Resolved long-standing font handling problems.            #Bug#
> 2819
> >    - requires recent library versions (libgd 2.4.0 libraqm 0.10.3)
> > * Better tic placement on logscale axis spanning two decades    #Bug#
> 2856
> > * Always clear previous plot title before a new plot is started #Bug#
> 2865
> > * Endpoint errors due to floating point precision               #Bug#
> 2768
> > * Repaired axis range parsing in parametric mode broken by change in
> 6.0.4
> > * Reworked logic for extending the command buffer incrementally on input.
> >    Long input lines with backslash continuation could lead to corruption
> >    of the command string if encountered early in program execution.
> >
> > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
> >
> > cheers,
> >       Ethan
> >
> >
> >
> >
> >
> > _______________________________________________
> > gnuplot-beta mailing list
> > gnu...@li...
> > Membership management via:
> https://urldefense.com/v3/__https://lists.sourceforge.net/lists/listinfo/gnuplot-beta__;!!K-Hz7m0Vt54!hICq3ODdIA1Zl3EhVy-KLGRkCS1rUe6lf7K7i7xJonlbqmlHMTa2K5cdPyoYdchWBtnkavSckVLdIDACDQ$
>


-- 
Ethan A Merritt
Department of Biochemistry
University of Washington, Seattle

Re: Pre-release testing version of gnuplot 6.0.5

[Bastian M. <bma...@we...>]

Builds fine on Windows with MSYS2/CLANG64 and MSYS2/CLANGARM64.

There's only a minor type issue in internal.c I noticed only know:  On 
Windows, intgr_t is a "long long int", not a "long int" (which would 
have only 32 bits). Hence the labs() call on line 1174 should really be 
llabs().  Is that fine for other platforms, too? Or should I add an 
ugly  #ifdef?

Best wishes,

   Bastian

Am 10.07.2026 um 05:56 schrieb Ethan A Merritt:
> I have placed a testing version of the source release
> package for gnuplot 6.0.4 on the "testing" section of the
> SourceForge download site.
>
>      https://sourceforge.net/projects/gnuplot/files/gnuplot/testing/
>
> My tentative plan is to announce the actual release in about
> two weeks (end of July).  That can be pushed back if you find
> any problems or want more time to test on other platforms.
>
> Brief release notes below, or see
>       gnuplot.sourceforge.net/ReleaseNotes_6_0_5.html
>
> %%%%%%%%%%    6.0.5 Release Notes July 2026   %%%%%%%%%%%%%%%%
>
> New features
>
> * Full mousing support for multiplots
>    - all panels report mouse coordinates
>    - all panels respond to pan/zoom/rotate mouse movement and hotkeys
>    - "replot" reexecutes multiplot commands starting from current state
>    - "remultiplot" restores the original state before reexecuting multiplot
>    - interactive mousing remains active after ^C or "reset"
>
> * Unicode escape sequences \U+xxxx are accepted in more contexts
>    - escape sequences are expanded in "noenhanced" strings
>    - escape sequences are expanded during substring evaluation
>    - escape sequences are expanded in string functions strstrt() and strlen()
>
> * New built-in function prod [<var> = <start> : <end>] <expression>
>    - implements the sequential product operation
>               prod [i=a:b] f(i)   =   f(a) * f(a+1) * ... * f(b)
>
> * Data files in EUC-JP encoding can be processed from other environments
>    - If running in a EUC-JP locale, use "set encoding locale"
>    - If running in a different locale, use "set encoding eucjp"
>    - In either case, this affects only input data.  It does not change
>      the encoding (usually UTF-8) written to plots on output.
>
> * New coloring option for contourfill
>    - Arbitrary discrete color ranges can be define without using a special palette
>          set contourfill defined [zmin1:zmax1] color1, [zmin2:zmax2] color2, ...
>
> * Allow "string"[i] as shorthand for "string"[i:i]
>
> Changed behaviour
>
> * Trailing whitespace on the command line is now ignored on input.
>    This means that the line-continuation character '\' is effective even
>    if it is followed by an unintentional blank.
>
> * Octal escape sequences must conform to the documented form \ooo
>    - E.g. \7 or \0106 or \456 or \0 are now rejected
>
> * "plot with labels" now accepts the "noenhanced" keyword
>
> Bug fixes
>
> * qt: Slow font initialization on Windows                       #Bug# 2316 2476
> * gd: Resolved long-standing font handling problems.            #Bug# 2819
>    - requires recent library versions (libgd 2.4.0 libraqm 0.10.3)
> * Better tic placement on logscale axis spanning two decades    #Bug# 2856
> * Always clear previous plot title before a new plot is started #Bug# 2865
> * Endpoint errors due to floating point precision               #Bug# 2768
> * Repaired axis range parsing in parametric mode broken by change in 6.0.4
> * Reworked logic for extending the command buffer incrementally on input.
>    Long input lines with backslash continuation could lead to corruption
>    of the command string if encountered early in program execution.
>
> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
>
> cheers,
> 	Ethan
>
>
>
>
>
> _______________________________________________
> gnuplot-beta mailing list
> gnu...@li...
> Membership management via: https://lists.sourceforge.net/lists/listinfo/gnuplot-beta

Re: Pre-release testing version of gnuplot 6.0.5

[ASSI <Str...@ne...>]

Ethan A Merritt writes:
> I have placed a testing version of the source release
> package for gnuplot 6.0.4 on the "testing" section of the
> SourceForge download site.
>
>     https://sourceforge.net/projects/gnuplot/files/gnuplot/testing/
>
> My tentative plan is to announce the actual release in about
> two weeks (end of July).  That can be pushed back if you find
> any problems or want more time to test on other platforms.

Building and testing fine on Cygwin.


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

SD adaptations for Waldorf Q V3.00R3 and Q+ V3.54R2:
http://Synth.Stromeko.net/Downloads.html#WaldorfSDada

Pre-release testing version of gnuplot 6.0.5

[Ethan A M. <me...@uw...>]

I have placed a testing version of the source release
package for gnuplot 6.0.4 on the "testing" section of the
SourceForge download site.

    https://sourceforge.net/projects/gnuplot/files/gnuplot/testing/

My tentative plan is to announce the actual release in about
two weeks (end of July).  That can be pushed back if you find
any problems or want more time to test on other platforms.

Brief release notes below, or see 
     gnuplot.sourceforge.net/ReleaseNotes_6_0_5.html

%%%%%%%%%%    6.0.5 Release Notes July 2026   %%%%%%%%%%%%%%%%

New features

* Full mousing support for multiplots
  - all panels report mouse coordinates
  - all panels respond to pan/zoom/rotate mouse movement and hotkeys
  - "replot" reexecutes multiplot commands starting from current state
  - "remultiplot" restores the original state before reexecuting multiplot
  - interactive mousing remains active after ^C or "reset"

* Unicode escape sequences \U+xxxx are accepted in more contexts
  - escape sequences are expanded in "noenhanced" strings
  - escape sequences are expanded during substring evaluation
  - escape sequences are expanded in string functions strstrt() and strlen()

* New built-in function prod [<var> = <start> : <end>] <expression>
  - implements the sequential product operation
             prod [i=a:b] f(i)   =   f(a) * f(a+1) * ... * f(b)

* Data files in EUC-JP encoding can be processed from other environments
  - If running in a EUC-JP locale, use "set encoding locale"
  - If running in a different locale, use "set encoding eucjp"
  - In either case, this affects only input data.  It does not change
    the encoding (usually UTF-8) written to plots on output.

* New coloring option for contourfill
  - Arbitrary discrete color ranges can be define without using a special palette
        set contourfill defined [zmin1:zmax1] color1, [zmin2:zmax2] color2, ...

* Allow "string"[i] as shorthand for "string"[i:i]

Changed behaviour

* Trailing whitespace on the command line is now ignored on input.
  This means that the line-continuation character '\' is effective even
  if it is followed by an unintentional blank.

* Octal escape sequences must conform to the documented form \ooo
  - E.g. \7 or \0106 or \456 or \0 are now rejected

* "plot with labels" now accepts the "noenhanced" keyword

Bug fixes

* qt: Slow font initialization on Windows                       #Bug# 2316 2476
* gd: Resolved long-standing font handling problems.            #Bug# 2819
  - requires recent library versions (libgd 2.4.0 libraqm 0.10.3)
* Better tic placement on logscale axis spanning two decades    #Bug# 2856
* Always clear previous plot title before a new plot is started #Bug# 2865
* Endpoint errors due to floating point precision               #Bug# 2768
* Repaired axis range parsing in parametric mode broken by change in 6.0.4
* Reworked logic for extending the command buffer incrementally on input.
  Long input lines with backslash continuation could lead to corruption
  of the command string if encountered early in program execution.

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

cheers,
	Ethan

Re: GNUPLOT used in a new Android app

[Alan L. <ala...@jh...>]

Dear Ethan,

thank you much for your response.
I verified the cross-compilation process with the most recent Android 
NDK and the current development version 6.1 snapshot 20 June 2026.
Really, the command.c patch with <sys/wait.h> turned out not to be 
necessary. So I updated my patch for the binaries distributed in the app 
accordingly, and deleted this modification.

Cheers,

Alan


Am 2026-05-30 02:35, schrieb Ethan A Merritt:
> On Monday, 25 May 2026 09:35:02 PDT Alan Liška via gnuplot-beta wrote:
>> Dear gnuplot authors and contributors,
>> 
>> I have published an open-source Android app (ChemKinDroid) which
>> installer contains the gnuplot binaries cross-compiled for Android.
> 
> I see no problem with the attribution or with copyright compliance.
> 
> However I am curious about the need for a patch to source file 
> command.c.
> In general we would like to understand the problem and fix the upstream
> source so that a patch is not needed.
> 
> Github shows
> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
> diff -ur gnuplot-6.0.3/src/command.c gnuplot_arm/src/command.c
> --- gnuplot-6.0.3/src/command.c	2025-05-31 06:36:14.000000000 +0200
> +++ gnuplot_arm/src/command.c	2026-05-12 21:29:08.842121763 +0200
> @@ -56,6 +56,7 @@
>   * input, and load_file() for non-interactive input.
>   */
> 
> +#include <sys/wait.h>
>  #include "command.h"
> 
>  #include "axis.h"
> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
> 
> Normally <sys/wait.h> should already have been included via syscfg.h as 
> a result
> of the autoconfigure test AC_HEADER_SYS_WAIT.
> Does this fail on Android?
> Would the attached alternative patch to syscfg.h work instead?
> If so I will make that change.
> 
> My only other comment is that gnuplot 6.1 is the development version,
> which means it is a moving target.  Unlike the 6.0.x release versions.
> Any bug report against the Adroid build will be more useful if it can
> be traced to, e.g., "version 6.1 snapshot 25 May 2026"
> or "version 6.1 commit  c5fc5c1c22".
> 
> 	cheers,
> 		Ethan
> 
>> The included docs contain these citations:
>> 
>> ...
>>   * GNUPLOT
>>   Authors: Thomas Williams, Colin Kelley
>>   Ref.: Williams, T., Kelley, C., Bersch, C., Bröker, H.B., Campbell, 
>> J.,
>> Cunningham, R., Denholm, D., Elber, G., Fearick, R., Grammes, C. and
>> Hart, L., 2023. gnuplot 6.1.
>>   Source code of the original version:
>> https://urldefense.com/v3/__https://sourceforge.net/projects/gnuplot/__;!!K-Hz7m0Vt54!jfNFRsKf4HsbEnTZC_MQpmC4Z9dgcPfwcgQDacKcw9AHwvgCYOj3LVPCKE1pl9s5JztsVZK9V0ewXbnCf716gzNFloLo6XDpmZ4T$
>>   Source code of the patch used for Android binaries used in this app:
>> https://urldefense.com/v3/__https://github.com/alanliska/gnuplot-patch__;!!K-Hz7m0Vt54!jfNFRsKf4HsbEnTZC_MQpmC4Z9dgcPfwcgQDacKcw9AHwvgCYOj3LVPCKE1pl9s5JztsVZK9V0ewXbnCf716gzNFloLo6eKC6iE-$
>>   License: open source, see the full license text under the Licenses
>> button
>>   Copyright 1986 - 1993, 1998, 2004   Thomas Williams, Colin Kelley
>>   *
>>   * Permission to use, copy, and distribute this software and its
>>   * documentation for any purpose with or without fee is hereby 
>> granted,
>>   * provided that the above copyright notice appear in all copies and
>>   * that both that copyright notice and this permission notice appear
>>   * in supporting documentation.
>>   *
>>   * Permission to modify the software is granted, but not the right to
>>   * distribute the complete modified source code.  Modifications are 
>> to
>>   * be distributed as patches to the released version.  Permission to
>>   * distribute binaries produced by compiling modified sources is
>> granted,
>>   * provided you
>>   *   1. distribute the corresponding source modifications from the
>>   *    released version in the form of a patch file along with the
>> binaries,
>>   *   2. add special version identification to distinguish your 
>> version
>>   *    in addition to the base release version number,
>>   *   3. provide your name and address as the primary contact for the
>>   *    support of your modified version, and
>>   *   4. retain our contact information in regard to use of the base
>>   *    software.
>>   * Permission to distribute the released version of the source code
>> along
>>   * with corresponding source modifications in the form of a patch 
>> file
>> is
>>   * granted with same provisions 2 through 4 for binary distributions.
>>   *
>>   * This software is provided "as is" without express or implied 
>> warranty
>>   * to the extent permitted by applicable law.
>> ...
>> 
>> Sources of the pre-compiled binaries:
>>   * gnuplot: 
>> https://urldefense.com/v3/__https://sourceforge.net/projects/gnuplot/__;!!K-Hz7m0Vt54!jfNFRsKf4HsbEnTZC_MQpmC4Z9dgcPfwcgQDacKcw9AHwvgCYOj3LVPCKE1pl9s5JztsVZK9V0ewXbnCf716gzNFloLo6XDpmZ4T$ 
>>  (official with
>> custom patch)
>>     patch: 
>> https://urldefense.com/v3/__https://github.com/alanliska/gnuplot-patch__;!!K-Hz7m0Vt54!jfNFRsKf4HsbEnTZC_MQpmC4Z9dgcPfwcgQDacKcw9AHwvgCYOj3LVPCKE1pl9s5JztsVZK9V0ewXbnCf716gzNFloLo6eKC6iE-$
>> ...
>> 
>> Is it sufficient?
>> You can check it also in the source code of the app:
>> https://urldefense.com/v3/__https://github.com/alanliska/ChemKinDroid__;!!K-Hz7m0Vt54!jfNFRsKf4HsbEnTZC_MQpmC4Z9dgcPfwcgQDacKcw9AHwvgCYOj3LVPCKE1pl9s5JztsVZK9V0ewXbnCf716gzNFloLo6b_AFNbx$
>> and in the patch itself:
>> https://urldefense.com/v3/__https://github.com/alanliska/gnuplot-patch__;!!K-Hz7m0Vt54!jfNFRsKf4HsbEnTZC_MQpmC4Z9dgcPfwcgQDacKcw9AHwvgCYOj3LVPCKE1pl9s5JztsVZK9V0ewXbnCf716gzNFloLo6eKC6iE-$
>> 
>> Please let me know if you liked to add something more.
>> Thank you in advance.
>> 
>> Best regards,
>> 
>> Alan Liška
>> 
>> 

Re: GNUPLOT used in a new Android app

[Ethan A M. <me...@uw...>]

On Monday, 25 May 2026 09:35:02 PDT Alan Liška via gnuplot-beta wrote:
> Dear gnuplot authors and contributors,
> 
> I have published an open-source Android app (ChemKinDroid) which 
> installer contains the gnuplot binaries cross-compiled for Android.

I see no problem with the attribution or with copyright compliance.

However I am curious about the need for a patch to source file command.c.
In general we would like to understand the problem and fix the upstream
source so that a patch is not needed.

Github shows
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
diff -ur gnuplot-6.0.3/src/command.c gnuplot_arm/src/command.c
--- gnuplot-6.0.3/src/command.c	2025-05-31 06:36:14.000000000 +0200
+++ gnuplot_arm/src/command.c	2026-05-12 21:29:08.842121763 +0200
@@ -56,6 +56,7 @@
  * input, and load_file() for non-interactive input.
  */
 
+#include <sys/wait.h>
 #include "command.h"
 
 #include "axis.h"
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

Normally <sys/wait.h> should already have been included via syscfg.h as a result
of the autoconfigure test AC_HEADER_SYS_WAIT.
Does this fail on Android?
Would the attached alternative patch to syscfg.h work instead?
If so I will make that change.

My only other comment is that gnuplot 6.1 is the development version,
which means it is a moving target.  Unlike the 6.0.x release versions.
Any bug report against the Adroid build will be more useful if it can
be traced to, e.g., "version 6.1 snapshot 25 May 2026"
or "version 6.1 commit  c5fc5c1c22".

	cheers,
		Ethan

> The included docs contain these citations:
> 
> ...
>   * GNUPLOT
>   Authors: Thomas Williams, Colin Kelley
>   Ref.: Williams, T., Kelley, C., Bersch, C., Bröker, H.B., Campbell, J., 
> Cunningham, R., Denholm, D., Elber, G., Fearick, R., Grammes, C. and 
> Hart, L., 2023. gnuplot 6.1.
>   Source code of the original version: 
> https://urldefense.com/v3/__https://sourceforge.net/projects/gnuplot/__;!!K-Hz7m0Vt54!jfNFRsKf4HsbEnTZC_MQpmC4Z9dgcPfwcgQDacKcw9AHwvgCYOj3LVPCKE1pl9s5JztsVZK9V0ewXbnCf716gzNFloLo6XDpmZ4T$ 
>   Source code of the patch used for Android binaries used in this app: 
> https://urldefense.com/v3/__https://github.com/alanliska/gnuplot-patch__;!!K-Hz7m0Vt54!jfNFRsKf4HsbEnTZC_MQpmC4Z9dgcPfwcgQDacKcw9AHwvgCYOj3LVPCKE1pl9s5JztsVZK9V0ewXbnCf716gzNFloLo6eKC6iE-$ 
>   License: open source, see the full license text under the Licenses 
> button
>   Copyright 1986 - 1993, 1998, 2004   Thomas Williams, Colin Kelley
>   *
>   * Permission to use, copy, and distribute this software and its
>   * documentation for any purpose with or without fee is hereby granted,
>   * provided that the above copyright notice appear in all copies and
>   * that both that copyright notice and this permission notice appear
>   * in supporting documentation.
>   *
>   * Permission to modify the software is granted, but not the right to
>   * distribute the complete modified source code.  Modifications are to
>   * be distributed as patches to the released version.  Permission to
>   * distribute binaries produced by compiling modified sources is 
> granted,
>   * provided you
>   *   1. distribute the corresponding source modifications from the
>   *    released version in the form of a patch file along with the 
> binaries,
>   *   2. add special version identification to distinguish your version
>   *    in addition to the base release version number,
>   *   3. provide your name and address as the primary contact for the
>   *    support of your modified version, and
>   *   4. retain our contact information in regard to use of the base
>   *    software.
>   * Permission to distribute the released version of the source code 
> along
>   * with corresponding source modifications in the form of a patch file 
> is
>   * granted with same provisions 2 through 4 for binary distributions.
>   *
>   * This software is provided "as is" without express or implied warranty
>   * to the extent permitted by applicable law.
> ...
> 
> Sources of the pre-compiled binaries:
>   * gnuplot: https://urldefense.com/v3/__https://sourceforge.net/projects/gnuplot/__;!!K-Hz7m0Vt54!jfNFRsKf4HsbEnTZC_MQpmC4Z9dgcPfwcgQDacKcw9AHwvgCYOj3LVPCKE1pl9s5JztsVZK9V0ewXbnCf716gzNFloLo6XDpmZ4T$  (official with 
> custom patch)
>     patch: https://urldefense.com/v3/__https://github.com/alanliska/gnuplot-patch__;!!K-Hz7m0Vt54!jfNFRsKf4HsbEnTZC_MQpmC4Z9dgcPfwcgQDacKcw9AHwvgCYOj3LVPCKE1pl9s5JztsVZK9V0ewXbnCf716gzNFloLo6eKC6iE-$ 
> ...
> 
> Is it sufficient?
> You can check it also in the source code of the app:
> https://urldefense.com/v3/__https://github.com/alanliska/ChemKinDroid__;!!K-Hz7m0Vt54!jfNFRsKf4HsbEnTZC_MQpmC4Z9dgcPfwcgQDacKcw9AHwvgCYOj3LVPCKE1pl9s5JztsVZK9V0ewXbnCf716gzNFloLo6b_AFNbx$ 
> and in the patch itself:
> https://urldefense.com/v3/__https://github.com/alanliska/gnuplot-patch__;!!K-Hz7m0Vt54!jfNFRsKf4HsbEnTZC_MQpmC4Z9dgcPfwcgQDacKcw9AHwvgCYOj3LVPCKE1pl9s5JztsVZK9V0ewXbnCf716gzNFloLo6eKC6iE-$ 
> 
> Please let me know if you liked to add something more.
> Thank you in advance.
> 
> Best regards,
> 
> Alan Liška
> 
> 


-- 
Ethan A Merritt
Department of Biochemistry
University of Washington, Seattle

GNUPLOT used in a new Android app

[Alan L. <ala...@jh...>]

Dear gnuplot authors and contributors,

I have published an open-source Android app (ChemKinDroid) which 
installer contains the gnuplot binaries cross-compiled for Android.
The included docs contain these citations:

...
  * GNUPLOT
  Authors: Thomas Williams, Colin Kelley
  Ref.: Williams, T., Kelley, C., Bersch, C., Bröker, H.B., Campbell, J., 
Cunningham, R., Denholm, D., Elber, G., Fearick, R., Grammes, C. and 
Hart, L., 2023. gnuplot 6.1.
  Source code of the original version: 
https://sourceforge.net/projects/gnuplot/
  Source code of the patch used for Android binaries used in this app: 
https://github.com/alanliska/gnuplot-patch
  License: open source, see the full license text under the Licenses 
button
  Copyright 1986 - 1993, 1998, 2004   Thomas Williams, Colin Kelley
  *
  * Permission to use, copy, and distribute this software and its
  * documentation for any purpose with or without fee is hereby granted,
  * provided that the above copyright notice appear in all copies and
  * that both that copyright notice and this permission notice appear
  * in supporting documentation.
  *
  * Permission to modify the software is granted, but not the right to
  * distribute the complete modified source code.  Modifications are to
  * be distributed as patches to the released version.  Permission to
  * distribute binaries produced by compiling modified sources is 
granted,
  * provided you
  *   1. distribute the corresponding source modifications from the
  *    released version in the form of a patch file along with the 
binaries,
  *   2. add special version identification to distinguish your version
  *    in addition to the base release version number,
  *   3. provide your name and address as the primary contact for the
  *    support of your modified version, and
  *   4. retain our contact information in regard to use of the base
  *    software.
  * Permission to distribute the released version of the source code 
along
  * with corresponding source modifications in the form of a patch file 
is
  * granted with same provisions 2 through 4 for binary distributions.
  *
  * This software is provided "as is" without express or implied warranty
  * to the extent permitted by applicable law.
...

Sources of the pre-compiled binaries:
  * gnuplot: https://sourceforge.net/projects/gnuplot/ (official with 
custom patch)
    patch: https://github.com/alanliska/gnuplot-patch
...

Is it sufficient?
You can check it also in the source code of the app:
https://github.com/alanliska/ChemKinDroid
and in the patch itself:
https://github.com/alanliska/gnuplot-patch

Please let me know if you liked to add something more.
Thank you in advance.

Best regards,

Alan Liška

-- 
RNDr. Alan Liška, Ph.D.
Ústav fyzikální chemie J. Heyrovského AV ČR, v.v.i. / J. Heyrovský 
Institute of Physical Chemistry of the CAS, v.v.i.
Dolejškova 3/2155, 18223 Praha 8
Česká republika / Czech Republic

Re: conditional text color?

[Ethan A M. <me...@uw...>]

Not exactly what you asked for, but one solution is to place the label in
an opaque box with neutral fillcolor:

   set style textbox opaque fc "grey" noborder
   plot $map1 matrix using 1:2:3 with image, \
        $map1 matrix using 1:2:(sprintf("%g",$3)) with labels tc "black"
boxed

Here is a way to make it conditional on the matrix value using an extra
pass.
This requires gnuplot 6 for the "if" syntax

   plot $map1 matrix using 1:2:3 with image, \
       $map1 matrix using 1:2:(sprintf("%g",$3)) with labels tc "black" if
($3 > -0.6), \
       $map1 matrix using 1:2:(sprintf("%g",$3)) with labels tc "white" if
($3 <= -0.6)

cheers,
         Ethan


On Mon, Mar 23, 2026 at 10:02 AM Allin Cottrell via gnuplot-beta <
gnu...@li...> wrote:

> I have a gnuplot routine that shows a correlation matrix as a heat
> map. If the dimension of the matrix is not too great it prints the
> value of the correlation coefficient in each "cell". Mostly this
> works fine, but if the color of the cell is a relatively dark blue
> (indicating a strong negative correlation) the default black text is
> not very legible. So I'd like to make the textcolor conditional.
>
> At present I have this plot statement:
>
> plot $data matrix with image, $data matrix using \
>   1:2:(sprintf("%.1f",$3)) with labels
>
> I'd like to append something like
>
> tc $3 < -0.6 ? rgbcolor "white" : rgbcolor "black"
>
> That doesn't work ("colorspec option not recognized", not very
> surprisingly). But is there a way to do this?
>
> Thanks for any guidance.
>
> --
> Allin Cottrell
> Professor Emeritus
> Wake Forest University

-- 
Ethan A Merritt
Department of Biochemistry
University of Washington, Seattle

Re: conditional text color?

[Cottrell, A. <cot...@wf...>]

Thanks, Ethan. I'm glad to be alerted to the "trailing if" syntax in
gnuplot 6; I hadn't noticed that.

Allin

On Mon, Mar 23, 2026 at 4:41 PM Ethan A Merritt <me...@uw...> wrote:
>
> Not exactly what you asked for, but one solution is to place the label in an opaque box with neutral fillcolor:
>
>    set style textbox opaque fc "grey" noborder
>    plot $map1 matrix using 1:2:3 with image, \
>         $map1 matrix using 1:2:(sprintf("%g",$3)) with labels tc "black" boxed
>
> Here is a way to make it conditional on the matrix value using an extra pass.
> This requires gnuplot 6 for the "if" syntax
>
>    plot $map1 matrix using 1:2:3 with image, \
>        $map1 matrix using 1:2:(sprintf("%g",$3)) with labels tc "black" if ($3 > -0.6), \
>        $map1 matrix using 1:2:(sprintf("%g",$3)) with labels tc "white" if ($3 <= -0.6)
>
> cheers,
>          Ethan
>
> On Mon, Mar 23, 2026 at 10:02 AM Allin Cottrell via gnuplot-beta <gnu...@li...> wrote:
>>
>> I have a gnuplot routine that shows a correlation matrix as a heat
>> map. If the dimension of the matrix is not too great it prints the
>> value of the correlation coefficient in each "cell". Mostly this
>> works fine, but if the color of the cell is a relatively dark blue
>> (indicating a strong negative correlation) the default black text is
>> not very legible. So I'd like to make the textcolor conditional.
>>
>> At present I have this plot statement:
>>
>> plot $data matrix with image, $data matrix using \
>>   1:2:(sprintf("%.1f",$3)) with labels
>>
>> I'd like to append something like
>>
>> tc $3 < -0.6 ? rgbcolor "white" : rgbcolor "black"
>>
>> That doesn't work ("colorspec option not recognized", not very
>> surprisingly). But is there a way to do this?
>>
>> Thanks for any guidance.

Re: conditional text color?

[Cottrell, A. <cot...@wf...>]

On Mon, Mar 23, 2026 at 3:36 PM Juhász Péter <pet...@gm...> wrote:
>
> On Mon, 2026-03-23 at 11:57 -0400, Allin Cottrell via gnuplot-beta
> wrote:
> > I have a gnuplot routine that shows a correlation matrix as a heat
> > map. If the dimension of the matrix is not too great it prints the
> > value of the correlation coefficient in each "cell". Mostly this
> > works fine, but if the color of the cell is a relatively dark blue
> > (indicating a strong negative correlation) the default black text is
> > not very legible. So I'd like to make the textcolor conditional.
> >
> > At present I have this plot statement:
> >
> > plot $data matrix with image, $data matrix using \
> >   1:2:(sprintf("%.1f",$3)) with labels
> >
> > I'd like to append something like
> >
> > tc $3 < -0.6 ? rgbcolor "white" : rgbcolor "black"
> >
> > That doesn't work ("colorspec option not recognized", not very
> > surprisingly). But is there a way to do this?
> >
> > Thanks for any guidance.
>
> Hi,
>
> AFAIK the labels style supports variable color. So you should be able
> to do something like
>
> $d << EOD
> 1 2 foo 0.3
> 2 3 bar 0.7
> EOD
>
> plot $d u 1:2:3:($4 > 0.6 ? 0xe00000 : 0x00e000) w labels tc rgb var
>
> In this case you'd have to specify the colors as integers, which may be
> slightly inconvenient and unmaintainable.

Actually that suits me fine; I don't mind giving the colors as ints.
Thanks for the suggestion!

Allin Cottrell

Re: conditional text color?

[Juhász P. <pet...@gm...>]

On Mon, 2026-03-23 at 11:57 -0400, Allin Cottrell via gnuplot-beta
wrote:
> I have a gnuplot routine that shows a correlation matrix as a heat 
> map. If the dimension of the matrix is not too great it prints the 
> value of the correlation coefficient in each "cell". Mostly this 
> works fine, but if the color of the cell is a relatively dark blue 
> (indicating a strong negative correlation) the default black text is 
> not very legible. So I'd like to make the textcolor conditional.
> 
> At present I have this plot statement:
> 
> plot $data matrix with image, $data matrix using \
>   1:2:(sprintf("%.1f",$3)) with labels
> 
> I'd like to append something like
> 
> tc $3 < -0.6 ? rgbcolor "white" : rgbcolor "black"
> 
> That doesn't work ("colorspec option not recognized", not very 
> surprisingly). But is there a way to do this?
> 
> Thanks for any guidance.

Hi,

AFAIK the labels style supports variable color. So you should be able
to do something like

$d << EOD
1 2 foo 0.3
2 3 bar 0.7    
EOD

plot $d u 1:2:3:($4 > 0.6 ? 0xe00000 : 0x00e000) w labels tc rgb var

In this case you'd have to specify the colors as integers, which may be
slightly inconvenient and unmaintainable. 
Other forms work too, like `tc variable` which would set the colors
based on the numbered color styles (same as used for functions/data by
the plot command), or `tc palette`, which draws colors according to a
color palette (either the default or the one you can set up before the
plot). This palette mode also draws a colorbar by default which you
have to disable if you don't want it.

So you have many options, neither of which is exactly what you want.

best regards,
Peter

conditional text color?

[Allin C. <cot...@wf...>]

I have a gnuplot routine that shows a correlation matrix as a heat 
map. If the dimension of the matrix is not too great it prints the 
value of the correlation coefficient in each "cell". Mostly this 
works fine, but if the color of the cell is a relatively dark blue 
(indicating a strong negative correlation) the default black text is 
not very legible. So I'd like to make the textcolor conditional.

At present I have this plot statement:

plot $data matrix with image, $data matrix using \
  1:2:(sprintf("%.1f",$3)) with labels

I'd like to append something like

tc $3 < -0.6 ? rgbcolor "white" : rgbcolor "black"

That doesn't work ("colorspec option not recognized", not very 
surprisingly). But is there a way to do this?

Thanks for any guidance.

-- 
Allin Cottrell
Professor Emeritus
Wake Forest University

Gnuplot Release 6.0.4

[Ethan A M. <me...@uw...>]

The release package for gnuplot 6.0 patchlevel 4 source code is now
available on SourceForge.
    https://sourceforge.net/projects/gnuplot/files/gnuplot/6.0.4/gnuplot-6.0.4.tar.gz

Binary versions, including one for Windows on ARM64, will be available
from that same download folder on SourceForge.

Many thanks to everyone who reported problems or contributed to testing,
and especially to everyone who contributed fixes or new code via the
development branch.  Your involvement, no matter how small, helps to
keep gnuplot a powerful multi-platform tool as it enters its 40th year of
distribution -- gnuplot version 1.0 was announced in 1986. 
 
A few of those new features are included in the 6.0.4 stable release.
Please see the release notes for a summary of bug fixes and other minor
changes.

Release Notes
   Gnuplot 6.0.4 <https://gnuplot.sourceforge.net/ReleaseNotes_6_0_4.html>

New Features

  • Support for Windows on ARM 64-bit architectures

  • New gprintf format specifiers for complex values
     %C  formats a complex value as {a, b}
          This is the format used by the gnuplot "print" command
     %Ci formats a complex value as a + bi
          Both the real and imaginary parts are printed using libc format %g
          with optional width and precision

  • Pre-defined variable Inf is set to floating point INFINITY

  • The "sharpen" filter handles vertical edges in a step function

  • 3D polygon objects rendered by "splot" can be assigned individual
     colors and fill styles

have fun plotting

    Ethan

gnuplot release 6.0.4 tarball

[Ethan A M. <me...@uw...>]

I have created a release directory on SourceForge for version 6.0.4

   https://sourceforge.net/projects/gnuplot/files/gnuplot/6.0.4/

The directory is "staged" for three days, which means only the developer
group can access it.  On 19 December it will become publically accessible
and I will send out a release announcement.

The release package includes Bastian's updates to the various Windows
configuration directories and build script (added since the testing versions),
and two final bug fixes.

The tarball content matches the state of git commit 3726262627

cheers,
        Ethan

Re: Backport request for 6.0.4 release version

[sfgphm <sf...@gm...>]

Dear Ethan,

I can confirm that all four issues in the test script have been resolved
with commit [4d3398fdd] to branch-6.0-stable.
Thank you very much for taking the time to backport this non-trivial fix.

Best regards,
H. Motoyoshi

On Thu, Dec 11, 2025 at 8:17 AM Ethan A Merritt <me...@uw...> wrote:

> On Tuesday, 9 December 2025 15:20:28 PST sfhm wrote:
> > Dear Ethan,
> >
> > I would like to request a backport to the 6.0.4 release.
> >
> > I have observed differences in polygon color rendering between the
> > development version and previous release versions (including
> > 6.0.4-testing). The development version renders polygons with the
> intended
> > colors, while the release versions render them with different colors.
>
> [snip example]
>
> > If the fix for this issue in the development version is sufficiently
> > stable, I would greatly appreciate its backport to the 6.0.4 release,
> even
> > if only for some of the cases rather than all of them.
>
> Unfortunately, that is not a trivial change.
>
> The code that deals with how fill color and border color are specified,
> stored, and eventually applied to graphical elements in the plot has
> diverged between 6.0 and the development branch.
>
> A partial list of relevant commits in 6.1 that do not apply cleanly to 6.0
>         c92721 fc9818
>
> Nevertheless, it is possible that the specific cases of "splot with
> polygons"
> in your test script could be handled either by some subset of the changes
> currently in the development branch or by some ad hoc fixes for 6.0.
> The trick is not to break anything else.
>
> I will look into it, but no promises about 6.0.4.
>
>  Ethan
>
>
>

Re: Pre-release testing version of gnuplot 6.0.4

[Ethan A M. <me...@uw...>]

On Friday, 12 December 2025 03:10:36 PST Bastian Märkisch via gnuplot-beta wrote:
> I have uploaded testing binaries for Windows x64 (clang), Windows on ARM 
> (clang), OS/2 (emx), DOS (djgpp) to the testing folder on SF.  They all 
> still needed some minor tweaks to the build process.
> 
> I also noted that neither the djgpp nor woa makefiles make it into the 
> source tarball at the moment.

The woa files are new, right?  
If they are to be included in the distribution tarball they need to be added
to the EXTRA_DIST list in config/Makefile.in.

So far as I can see, only two files related to djgpp are in that list
   config/config.dj2   config/djconfig.sh

Are these still needed?  Have they been superseded by
   config/djgpp/config.h  config/djgpp/Makefile

>    Bastian
> 
> Am 06.12.2025 um 03:13 schrieb Ethan A Merritt:
> > I have placed a testing version of the source release
> > package for gnuplot 6.0.4 on the "testing" section of the
> > SourceForge download site.
> > ...
> > My tentative plan is to announce the actual release in about
> > two weeks (the week of 15 December).  That can be pushed back
> > if you find any problems or want to suggest inclusion
> > of a bug-fix from the development version that was not yet
> > backported.

-- 
   Ethan

Re: Pre-release testing version of gnuplot 6.0.4

[Bastian M. <bma...@we...>]

I have uploaded testing binaries for Windows x64 (clang), Windows on ARM 
(clang), OS/2 (emx), DOS (djgpp) to the testing folder on SF.  They all 
still needed some minor tweaks to the build process.

I also noted that neither the djgpp nor woa makefiles make it into the 
source tarball at the moment.

   Bastian

Am 06.12.2025 um 03:13 schrieb Ethan A Merritt:
> I have placed a testing version of the source release
> package for gnuplot 6.0.4 on the "testing" section of the
> SourceForge download site.
>
>      https://sourceforge.net/projects/gnuplot/files/gnuplot/testing/
>
> Nothing major to look for in this release, but it does include a number
> of recent bug fixes.　The configure script and various header files
> have been modified to allow use with C compilers that enforce c23
> syntax by default. This should not affect use with older compilers.
>
> My tentative plan is to announce the actual release in about
> two weeks (the week of 15 December).  That can be pushed back
> if you find any problems or want to suggest inclusion
> of a bug-fix from the development version that was not yet
> backported.
>
> Brief release notes below, or see
>       gnuplot.sourceforge.net/ReleaseNotes_6_0_4.html
>
>
>
> %%%%%%%%%%    6.0.4 Release Notes 01-December-2025    %%%%%%%%%%%%%%%%
>
> NEW (back-ported from development version)
> ------------------------------------------
> * variable Inf is pre-set to floating point INFINITY
> * "sharpen" filter handles vertical edges in a step function
> * gprintf format specifiers %C and %Ci
>    - %C  formats a complex value as {a, b}
>        This is the format used by the gnuplot "print" command
>    - %Ci formats a complex value as a + bi
>        Both the real and imaginary parts are printed using libc format %g
>        with optional width and precision
> * Support for Windows on ARM 64-bit architectures
>    - mingw: optionally use the MSYS2/CLANG64 environment
>    - Cross-compilation for Windows on ARM is possible using MSYS2/CLANG64/CLANGARM64
>
> CHANGES
> -------
> * better adherence to c23 standard
> * splot with pm3d fillcolor <x> generates a matching key sample
> * gprintf %H format is distinct from %h in utf8 context
>    - %H uses dot operator between mantissa and exponent
> * kitty, sixel: default size to 100% width, 75% height of terminal window
> * Partially deprecate the "sample" keyword for plot commands.
>    The plot command accepts axis ranges as the first thing after "plot".
>    If present, they update the primary axis ranges x y x2 y2.
>    However if this is preced by the keyword "sample" it refers to a sampling range
>    rather than an axis range.  If a second colon is found in the range,
>    as in [t=min:max:increment], this is unambiguously a sample range so the
>    "sample" keyword is not required.
> * Do not break contour lines into 100-segment fragments.
>    - This fixes a bug when saved contour lines are used as polygons.
>    - To maintain previous placement of contour labels along long contours, use
>    "set cntrlabl interval 100".
>
> FIXES
> -----
> * sixel: improved support for a transparent background
> * kitty: support animation on a transparent background
> * webp:  support animation on a transparent background
> * backport fixes for color assignment to pm3d surfaces
>    - "set pm3d implicit" + "with lines" treated the same as "with pm3d"
>    - adds support for "fc background"
>    - pm3d interpolate > 1 inherits rgb color assignments from the parent tile
> * fill style of colorbox should match fill style of pm3d surface
> * svg: fillstyle solid <frac> was incorrectly treated as transparent
> * transparent fill color in 3D boxes and polygons
> * alpha channel colors on ARM platforms
> * resolve ambiguous syntax in "set dashtype i (n,m)"
> * svg: font size changes within a text fragment could be lost
> * "reset session" must terminate multiplot mode
> * determination of above/below in polar mode filledcurves
> * always flush cached lines from hidden3d if any are present
> * incorrect evaluation of a**b for integer a, integer b < 0
>    in the case that overflow handling has been set to "NaN" or "undefined"
> * loss of precision in some ranges for asin acos asinh acosh
>
>

Re: Backport request for 6.0.4 release version

[Ethan A M. <me...@uw...>]

On Tuesday, 9 December 2025 15:20:28 PST sfhm wrote:
> Dear Ethan,
> 
> I would like to request a backport to the 6.0.4 release.
> 
> I have observed differences in polygon color rendering between the
> development version and previous release versions (including
> 6.0.4-testing). The development version renders polygons with the intended
> colors, while the release versions render them with different colors.

[snip example] 

> If the fix for this issue in the development version is sufficiently
> stable, I would greatly appreciate its backport to the 6.0.4 release, even
> if only for some of the cases rather than all of them.

Unfortunately, that is not a trivial change.  

The code that deals with how fill color and border color are specified,
stored, and eventually applied to graphical elements in the plot has
diverged between 6.0 and the development branch.

A partial list of relevant commits in 6.1 that do not apply cleanly to 6.0
	c92721 fc9818 

Nevertheless, it is possible that the specific cases of "splot with polygons"
in your test script could be handled either by some subset of the changes
currently in the development branch or by some ad hoc fixes for 6.0.
The trick is not to break anything else.

I will look into it, but no promises about 6.0.4.  

 Ethan

Re: New warning about inline data and multiplots

[Dima K. <gn...@di...>]

Good info. Thanks for writing that up! Hopefully I'll get some time to
fix this on my end soon.

Backport request for 6.0.4 release version

[sfhm <sf...@gm...>]

Dear Ethan,

I would like to request a backport to the 6.0.4 release.

I have observed differences in polygon color rendering between the
development version and previous release versions (including
6.0.4-testing). The development version renders polygons with the intended
colors, while the release versions render them with different colors.

The following script demonstrates the issue:
```
set xrange [-2:2]
set yrange [-2:2]
set zrange [-2:2]

$contour <<EOD
-1 -1 0
-1  1 0
 1  1 0
 1 -1 0
-1 -1 0
EOD

set style line 3 lc "#d000ff00"
set key noautotitle
set isotropic
set xyplane 0
set view ,,1.6


set multiplot layout 2,2

  # plot 1 : filled in "Red" with "Blue" border
  set title "Plot 1"
  splot $contour using 1:2:3 with polygons fillstyle border lc "blue" fc
rgb "red"

  # plot 2 : filled in pallete color with default border
  set title "Plot 2"
  splot $contour using 1:2:3 with polygons fillstyle fc palette

  # plot 3 : filled in transparent default color with "Black" border
  set title "Plot 3"
  splot $contour using 1:2:3 with polygons fillstyle transparent border lc
black

  # plot 4 : filled in transparent "Green" color with default border
  set title "Plot 4"
  splot $contour using 1:2:3 with polygons ls 3

unset multiplot

pause -1
```

If the fix for this issue in the development version is sufficiently
stable, I would greatly appreciate its backport to the 6.0.4 release, even
if only for some of the cases rather than all of them.

Thank you for your consideration.

Best regards,
H. Motoyoshi

Re: Pre-release testing version of gnuplot 6.0.4

[Erik L. <eri...@gm...>]

Dear Ethan,

This compiles properly on macOS. I placed a universal binary here (all
libraries statically linked):
https://csml.northwestern.edu/Download/Gnuplot/gnuplot-6.0.4-qt5-universal.pkg,
so that others can test it as well.

I did not compile without mouse support, so I cannot speak to the issue
Achim identified.

Regards,

Erik


On Fri, Dec 5, 2025 at 8:14 PM Ethan A Merritt <me...@uw...> wrote:

> I have placed a testing version of the source release
> package for gnuplot 6.0.4 on the "testing" section of the
> SourceForge download site.
>
>     https://sourceforge.net/projects/gnuplot/files/gnuplot/testing/
>
> Nothing major to look for in this release, but it does include a number
> of recent bug fixes. The configure script and various header files
> have been modified to allow use with C compilers that enforce c23
> syntax by default. This should not affect use with older compilers.
>
> My tentative plan is to announce the actual release in about
> two weeks (the week of 15 December).  That can be pushed back
> if you find any problems or want to suggest inclusion
> of a bug-fix from the development version that was not yet
> backported.
>
> Brief release notes below, or see
>      gnuplot.sourceforge.net/ReleaseNotes_6_0_4.html
>
>
>
> %%%%%%%%%%    6.0.4 Release Notes 01-December-2025    %%%%%%%%%%%%%%%%
>
> NEW (back-ported from development version)
> ------------------------------------------
> * variable Inf is pre-set to floating point INFINITY
> * "sharpen" filter handles vertical edges in a step function
> * gprintf format specifiers %C and %Ci
>   - %C  formats a complex value as {a, b}
>       This is the format used by the gnuplot "print" command
>   - %Ci formats a complex value as a + bi
>       Both the real and imaginary parts are printed using libc format %g
>       with optional width and precision
> * Support for Windows on ARM 64-bit architectures
>   - mingw: optionally use the MSYS2/CLANG64 environment
>   - Cross-compilation for Windows on ARM is possible using
> MSYS2/CLANG64/CLANGARM64
>
> CHANGES
> -------
> * better adherence to c23 standard
> * splot with pm3d fillcolor <x> generates a matching key sample
> * gprintf %H format is distinct from %h in utf8 context
>   - %H uses dot operator between mantissa and exponent
> * kitty, sixel: default size to 100% width, 75% height of terminal window
> * Partially deprecate the "sample" keyword for plot commands.
>   The plot command accepts axis ranges as the first thing after "plot".
>   If present, they update the primary axis ranges x y x2 y2.
>   However if this is preced by the keyword "sample" it refers to a
> sampling range
>   rather than an axis range.  If a second colon is found in the range,
>   as in [t=min:max:increment], this is unambiguously a sample range so the
>   "sample" keyword is not required.
> * Do not break contour lines into 100-segment fragments.
>   - This fixes a bug when saved contour lines are used as polygons.
>   - To maintain previous placement of contour labels along long contours,
> use
>   "set cntrlabl interval 100".
>
> FIXES
> -----
> * sixel: improved support for a transparent background
> * kitty: support animation on a transparent background
> * webp:  support animation on a transparent background
> * backport fixes for color assignment to pm3d surfaces
>   - "set pm3d implicit" + "with lines" treated the same as "with pm3d"
>   - adds support for "fc background"
>   - pm3d interpolate > 1 inherits rgb color assignments from the parent
> tile
> * fill style of colorbox should match fill style of pm3d surface
> * svg: fillstyle solid <frac> was incorrectly treated as transparent
> * transparent fill color in 3D boxes and polygons
> * alpha channel colors on ARM platforms
> * resolve ambiguous syntax in "set dashtype i (n,m)"
> * svg: font size changes within a text fragment could be lost
> * "reset session" must terminate multiplot mode
> * determination of above/below in polar mode filledcurves
> * always flush cached lines from hidden3d if any are present
> * incorrect evaluation of a**b for integer a, integer b < 0
>   in the case that overflow handling has been set to "NaN" or "undefined"
> * loss of precision in some ranges for asin acos asinh acosh
>
>
> --
> Ethan A Merritt
> Department of Biochemistry
> University of Washington, Seattle
>
>
>
>
>
> _______________________________________________
> gnuplot-beta mailing list
> gnu...@li...
> Membership management via:
> https://lists.sourceforge.net/lists/listinfo/gnuplot-beta
>

Re: Pre-release testing version of gnuplot 6.0.4

[ASSI <Str...@ne...>]

Ethan A Merritt writes:
> I have placed a testing version of the source release
> package for gnuplot 6.0.4 on the "testing" section of the
> SourceForge download site.

There's a missing conditional in unset.c which prevents successful
compilation when mouse support is not enabled.

--8<---------------cut here---------------start------------->8---
--- origsrc/gnuplot-6.0.4/src/unset.c
+++ src/gnuplot-6.0.4/src/unset.c
@@ -2068,8 +2068,10 @@
 	    multiplot_end();
 	init_constants();
 	init_session();
+#ifdef USE_MOUSE
 	reset_mouse();
 	reset_since_last_plot = TRUE;
+#endif
 	return;
     }

@@ -2116,7 +2118,9 @@
      * suppress some of the commentary output by the individual
      * unset_...() routines. */
     interactive = FALSE;
+#ifdef USE_MOUSE
     reset_since_last_plot = TRUE;
+#endif
 
     unset_samples();
     unset_isosamples();

--- origsrc/gnuplot-6.0.4/src/plot2d.c
+++ src/gnuplot-6.0.4/src/plot2d.c
@@ -225,7 +225,9 @@
 	int_error(c_token, "use 'set term' to set terminal type first");
 
     is_3d_plot = FALSE;
+#ifdef USE_MOUSE
     reset_since_last_plot = FALSE;
+#endif
 
     if (parametric && strcmp(set_dummy_var[0], "u") == 0)
 	strcpy(set_dummy_var[0], "t");

--- origsrc/gnuplot-6.0.4/src/plot3d.c
+++ src/gnuplot-6.0.4/src/plot3d.c
@@ -312,7 +312,9 @@
     AXIS_INDEX axis, u_axis, v_axis;
 
     is_3d_plot = TRUE;
+#ifdef USE_MOUSE
     reset_since_last_plot = FALSE;
+#endif
 
     if (parametric && strcmp(set_dummy_var[0], "t") == 0) {
 	strcpy(set_dummy_var[0], "u");

--8<---------------cut here---------------end--------------->8---


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

SD adaptation for Waldorf Blofeld V1.15B11:
http://Synth.Stromeko.net/Downloads.html#WaldorfSDada

Pre-release testing version of gnuplot 6.0.4

[Ethan A M. <me...@uw...>]

I have placed a testing version of the source release
package for gnuplot 6.0.4 on the "testing" section of the
SourceForge download site.

    https://sourceforge.net/projects/gnuplot/files/gnuplot/testing/

Nothing major to look for in this release, but it does include a number
of recent bug fixes.　The configure script and various header files
have been modified to allow use with C compilers that enforce c23
syntax by default. This should not affect use with older compilers.

My tentative plan is to announce the actual release in about
two weeks (the week of 15 December).  That can be pushed back
if you find any problems or want to suggest inclusion
of a bug-fix from the development version that was not yet
backported.

Brief release notes below, or see 
     gnuplot.sourceforge.net/ReleaseNotes_6_0_4.html



%%%%%%%%%%    6.0.4 Release Notes 01-December-2025    %%%%%%%%%%%%%%%%

NEW (back-ported from development version)
------------------------------------------
* variable Inf is pre-set to floating point INFINITY
* "sharpen" filter handles vertical edges in a step function
* gprintf format specifiers %C and %Ci
  - %C  formats a complex value as {a, b}
      This is the format used by the gnuplot "print" command
  - %Ci formats a complex value as a + bi
      Both the real and imaginary parts are printed using libc format %g
      with optional width and precision
* Support for Windows on ARM 64-bit architectures
  - mingw: optionally use the MSYS2/CLANG64 environment
  - Cross-compilation for Windows on ARM is possible using MSYS2/CLANG64/CLANGARM64

CHANGES
-------
* better adherence to c23 standard
* splot with pm3d fillcolor <x> generates a matching key sample
* gprintf %H format is distinct from %h in utf8 context
  - %H uses dot operator between mantissa and exponent
* kitty, sixel: default size to 100% width, 75% height of terminal window
* Partially deprecate the "sample" keyword for plot commands.
  The plot command accepts axis ranges as the first thing after "plot".
  If present, they update the primary axis ranges x y x2 y2.
  However if this is preced by the keyword "sample" it refers to a sampling range
  rather than an axis range.  If a second colon is found in the range,
  as in [t=min:max:increment], this is unambiguously a sample range so the
  "sample" keyword is not required.
* Do not break contour lines into 100-segment fragments.
  - This fixes a bug when saved contour lines are used as polygons.
  - To maintain previous placement of contour labels along long contours, use
  "set cntrlabl interval 100".

FIXES
-----
* sixel: improved support for a transparent background
* kitty: support animation on a transparent background
* webp:  support animation on a transparent background
* backport fixes for color assignment to pm3d surfaces
  - "set pm3d implicit" + "with lines" treated the same as "with pm3d"
  - adds support for "fc background"
  - pm3d interpolate > 1 inherits rgb color assignments from the parent tile
* fill style of colorbox should match fill style of pm3d surface
* svg: fillstyle solid <frac> was incorrectly treated as transparent
* transparent fill color in 3D boxes and polygons
* alpha channel colors on ARM platforms
* resolve ambiguous syntax in "set dashtype i (n,m)"
* svg: font size changes within a text fragment could be lost
* "reset session" must terminate multiplot mode
* determination of above/below in polar mode filledcurves
* always flush cached lines from hidden3d if any are present
* incorrect evaluation of a**b for integer a, integer b < 0
  in the case that overflow handling has been set to "NaN" or "undefined"
* loss of precision in some ranges for asin acos asinh acosh


-- 
Ethan A Merritt
Department of Biochemistry
University of Washington, Seattle