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

Re: cvs(today): splot "with image" y-size; missing data color

[Daniel J S. <dan...@ie...>]

Daniel J Sebald wrote:

> Andreas K. Huettel wrote:
>
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>>
>> Dear developers,
>> I just checked out the CVS version and was amazed that work has been 
>> done on all the "killer features" I've been missing in 4.0 - THANK 
>> YOU!!!
>> (especially "with image", "set ytics offset ...", string variables)
>>
>> Anyway, here's one bugreport and one feature request (for details,
>> see also http://www2.nano.physik.uni-muenchen.de/~huettel/gnuplot/ ):
>>
>> 1) when I use identical data and an identical macro for "with pm3d" 
>> and "with image" (postscript terminal), the image is drawn much too 
>> big and covers the lower plot border. do you have any idea how to fix 
>> this?
>> (for examples see above url)
>>
>
> That is a strange one.  I assume you have an image for which you 
> select a range that is only a portion of the image.  This is supposed 
> to work fine.  I will have a look this weekend.


Actually, now that I look closely, I see there _aren't_ any such demos 
in image.dem.  (I will add such an example.)  I know this works in the X 
term driver.  But I don't think any of us thought to check this out in 
all the terminals (i.e., PostScript, PDF, PNG, etc.).  In fact, OK, I 
just checked a range clipped image example using the PNG terminal 
driver.  It fails.  There are two sets of coordinates sent through the 
terminal image call, the location of the image and the visible area.  
(For systems that can display portions of pixels on the edge.)

I'll go through the terminals one at a time and let you know when they 
are ready.

Thanks for find this,

Dan

Re: "Bind all <key> <command>" patch on SourceForge

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

I suddenly realized that I have been thinking of this in the
wrong way, and misinterpreting what is happening.
Including "set term foo" in a key binding has nothing to do with it.

The command  "gnuplot < myscriptfile"
resets stdin to a piped copy of the file myscriptfile.
But now "pause -1" makes no sense, because it says
to wait for the next input on stdin, and we know that stdin
will keep on feeding characters from myscriptfile until it
runs off the end.  That's why it returns immediately in your
original test case.

OK, so we can't use "pause -1".  The obvious replacement
is "pause mouse key". And this almost works, but not quite.
The glitch is that currently "pause mouse key"
(a) only works in the active window, and
(b) overrides all key bindings

Both of these are fixable.  So I have uploaded a new version
of the patchset that changes gnuplot behaviour as follows:

- "bind all <key> foo" causes that key to be terminate a
  "pause mouse key" even if it is typed in an inactive window

- key bindings are evaluated *before* checking for
  "pause mouse key", so if you terminate the pause by hitting
  a key with a binding, the binding executes first

With this revised version I think you can do everything you
wanted. Here are my test files:

bindall.dem:

  set mouse
  n=1; set term x11 n; plot x**n
  n=2; set term x11 n; plot x**n
  bind all Home 'n=MOUSE_KEY_WINDOW; set term x11 n; plot x**n'
  bind all End 'DONE = 1'
  bind all s 'show variable'
  DONE = 0
  print "Hit <End> in any plot window to continue"
  load "wait_for_End"
  #
  print "Now I am done"

wait_for_End:

  pause mouse key
  if (!DONE) print "Saw key in window ",MOUSE_KEY_WINDOW," but not yet done"
  if (!DONE) reread

These scripts work as intended for either
	gnuplot < bindall.dem
or
	gnuplot
	gnuplot> load 'bindall.dem'

There is still something out of sync if you say
	gnuplot bindall.dem

As before, please let me know if more, or different, functionality
is ideally needed for your full application.

-- 
Ethan A Merritt
Biomolecular Structure Center
University of Washington 98195-7742

Tgif terminal bug (font reset)

[Shigeharu T. <sh...@ie...>]

shige 01/07 2005
----------------

I found a bug of term/tgif.trm. On tgif terminal, write_multiline
use the specified font in the first line, but in the following 
lines the font setting is reset and default font is used. 

The following small patch may fix it and this includes the code 
for t->set_font("").

----- From here -----
*** term/tgif.trm.ORG	Thu Jan  6 12:11:10 2005
--- term/tgif.trm	Fri Jan  7 13:44:48 2005
***************
*** 720,728 ****
--- 720,730 ----
  
      uActNr += 2;
  
+ #if 0
      /* reset font size and font, they might have been changed */
      uActFontSize = uDefaultFontSize;
      strcpy(sActFont, sDefaultFont);
+ #endif
  
  }				/* TGIF_put_text */
  /*}}}  */
***************
*** 1588,1594 ****
  TERM_PUBLIC int
  TGIF_set_font(const char *font)
  {
!     int sep = strcspn(font, ",");
      if (sep > 0) {
  	strncpy(sActFont, font, sep);
  	sActFont[sep] = NUL;
--- 1590,1600 ----
  TERM_PUBLIC int
  TGIF_set_font(const char *font)
  {
!     int sep;
! 
!     if (!font || !(*font))
!         font = sDefaultFont;
!     sep = strcspn(font, ",");
      if (sep > 0) {
  	strncpy(sActFont, font, sep);
  	sActFont[sep] = NUL;
----- To here -----

+========================================================+
 Shigeharu TAKENO     NIigata Institute of Technology
                       kashiwazaki,Niigata 945-1195 JAPAN
 sh...@ie...   TEL(&FAX): +81-257-22-8161
+========================================================+

Re: cvs(today): splot "with image" y-size; missing data color

[Daniel J S. <dan...@ie...>]

Andreas K. Huettel wrote:

>-----BEGIN PGP SIGNED MESSAGE-----
>Hash: SHA1
>
>
>Dear developers, 
>
>I just checked out the CVS version and was amazed that work has been done 
>on all the "killer features" I've been missing in 4.0 - THANK YOU!!!
>(especially "with image", "set ytics offset ...", string variables)
>
>Anyway, here's one bugreport and one feature request (for details,
>see also http://www2.nano.physik.uni-muenchen.de/~huettel/gnuplot/ ):
>
>1) when I use identical data and an identical macro for "with pm3d" and 
>"with image" (postscript terminal), the image is drawn much too big and 
>covers the lower plot border. do you have any idea how to fix this?
>(for examples see above url)
>

That is a strange one.  I assume you have an image for which you select 
a range that is only a portion of the image.  This is supposed to work 
fine.  I will have a look this weekend.  (I'm just back from a few days 
of meetings.)  In the mean time, perhaps you could run the demo 
"image.dem" and cross-reference against the file:

http://acer-access.com/~ds...@ac.../gnuplot/image.pdf

If there are differences, please let me know or send a PDF file or 
compressed PostScript file.  (There are some examples in the demo which 
show only a subregion of an image.

Thanks,

Dan

Re: cvs(today): splot "with image" y-size; missing data color

[Andreas K. H. <And...@Ph...>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


Dear Mr Mikulik, 

#]> 1) when I use identical data and an identical macro for "with pm3d" and
#]> "with image" (postscript terminal), the image is drawn much too big and
#]> covers the lower plot border. do you have any idea how to fix this?
#]> (for examples see above url)
#]
#]In pm3d, you can use "set pm3d clip..." settings.

I'm sorry, I have not expressed myself clearly enough... everything is 
working perfectly with pm3d mode. 

Its just that when I use the same parameters for "with image" mode
instead, then the color plot is scaled incorrectly and does not fit into
the "axes frame".

#]> 2) is there any way to set the "missing data" color for "with image" (and
#]> perhaps "with pm3d") ?
(...)
#]
#]For pm3d, either miss them in the data file, or set a z-range, or use
#]different z-range and cb-range;
#]	set zrange [10:1e8]
#]will invalidate all data out of the above range.

Do I understand this correctly that in pm3d map mode the zrange controls 
whether points are "drawn", i.e. whether for example in a postscript image 
the background is visible or not?

#]That cannot be done for "with image" for principal reasons -- that's to draw
#]a matrix, so all pts must be valid. Thus, you have to make a palette with
#]one color white and use the corresponding value for coloring your special
#]pixels.

Sure... what I was asking is: is there any way to set this special color
value?

regards, Andreas



- ---------------------------------------------------------------------
Dipl.-Phys. Andreas K. Huettel          tel. +49 89 2180 3349 (univ.)
Sektion Physik der LMU                  fax  +49 89 2180 3182 (univ.)
LS Prof. J.P. Kotthaus                                 hu...@lm...
Geschwister-Scholl-Platz 1                          ma...@ak...
80539 Muenchen                 and...@ph...
Germany                             http://www.akhuettel.de/research/
- ---------------------------------------------------------------------
Privat: Andreas K. Huettel, Enhuberstrasse 5, 80333 Muenchen, Germany
- ---------------------------------------------------------------------
Please use GNUPG or PGP for signed and encrypted email. My public key
can be found at  http://www.akhuettel.de/pgp_key.html




-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)

iD8DBQFB3YRrL+gLs3iH94cRAoXQAKCOtQnPG9WBIRSDGsBnIhGv+G23wgCcDZwH
lGEIGMqQNzQolsJqP8MSStA=
=9F+s
-----END PGP SIGNATURE-----

Re: cvs(today): splot "with image" y-size; missing data color

[Petr M. <mi...@ph...>]

> 1) when I use identical data and an identical macro for "with pm3d" and
> "with image" (postscript terminal), the image is drawn much too big and
> covers the lower plot border. do you have any idea how to fix this?
> (for examples see above url)

In pm3d, you can use "set pm3d clip..." settings.


> 2) is there any way to set the "missing data" color for "with image" (and
> perhaps "with pm3d") ?
> I.e., I am doing many log plots with noisy data, and "low value" means
> "white" in my palette. I'd very much like to be able to set "missing data"
> as "white", since thats what (minimally) negative values end up as....

For pm3d, either miss them in the data file, or set a z-range, or use
different z-range and cb-range;
	set zrange [10:1e8]
will invalidate all data out of the above range.

That cannot be done for "with image" for principal reasons -- that's to draw
a matrix, so all pts must be valid. Thus, you have to make a palette with
one color white and use the corresponding value for coloring your special
pixels.

---
PM

cvs(today): splot "with image" y-size; missing data color

[Andreas K. H. <And...@Ph...>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


Dear developers, 

I just checked out the CVS version and was amazed that work has been done 
on all the "killer features" I've been missing in 4.0 - THANK YOU!!!
(especially "with image", "set ytics offset ...", string variables)

Anyway, here's one bugreport and one feature request (for details,
see also http://www2.nano.physik.uni-muenchen.de/~huettel/gnuplot/ ):

1) when I use identical data and an identical macro for "with pm3d" and 
"with image" (postscript terminal), the image is drawn much too big and 
covers the lower plot border. do you have any idea how to fix this?
(for examples see above url)

2) is there any way to set the "missing data" color for "with image" (and 
perhaps "with pm3d") ? 
I.e., I am doing many log plots with noisy data, and "low value" means
"white" in my palette. I'd very much like to be able to set "missing data" 
as "white", since thats what (minimally) negative values end up as....

Thanks a lot in advance, Andreas


- ---------------------------------------------------------------------
Dipl.-Phys. Andreas K. Huettel          tel. +49 89 2180 3349 (univ.)
Sektion Physik der LMU                  fax  +49 89 2180 3182 (univ.)
LS Prof. J.P. Kotthaus                                 hu...@lm...
Geschwister-Scholl-Platz 1                          ma...@ak...
80539 Muenchen                 and...@ph...
Germany                             http://www.akhuettel.de/research/
- ---------------------------------------------------------------------
Privat: Andreas K. Huettel, Enhuberstrasse 5, 80333 Muenchen, Germany
- ---------------------------------------------------------------------
Please use GNUPG or PGP for signed and encrypted email. My public key
can be found at  http://www.akhuettel.de/pgp_key.html




-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)

iD8DBQFB3WHIL+gLs3iH94cRAoOsAJwNI7SeUJa6nMmEOf9cxSeJ/hszlQCfRtJo
tDyJCSP6kfh+1eSSAr42cg4=
=yp2p
-----END PGP SIGNATURE-----

Re: "Bind all <key> <command>" patch on SourceForge

[Jacques B. <jac...@no...>]

Ethan Merritt wrote:
> It is a bit ugly, however, since it burns a lot of CPU
> cycles in that wait loop (the "pause -1" returns immediately,
> as you already know, and so the loop is very tight). 
> ... I do not understand
> the details of what goes wrong with the "pause" command
> in your example, and without understanding it is hard to
> develop a fix.

Yes, the behavior of your "gnuplot<inputfile.dat" example fits exactly
my need but its heavy CPU use is not acceptable. So I tried it in the
"gnuplot inputfile.dat" way, but then it doesn't work so well: the
application has to be terminated with both End and then Return in the
launch terminal (only one method should be used). So, as you say that
there is no problem with the pause command in that way, I hoped that you
would show me how to modify this example so as to use it with "gnuplot
inputfile.dat", and get the same behavior without the CPU use.

Thanks again for your help.

{column,row}major of multiplot

[Shigeharu T. <sh...@ie...>]

shige 01/06 2005
----------------

In ChangeLog of CVS version: 

  2005-01-04  Ethan Merritt  <merritt@u.washington.edu>
  ...
     * docs/gnuplot.doc: Correct the documentation to say
     'set multiplot layout <rows>,<cols>' (rows/cols were reversed in the
     previous version, and mis-described in the example).

I think it is true. Moreover, I wonder the behavior of options 
{column,row}major are reversed. gnuplot.doc says:

  This grid is filled rows first or columns first depending on
  whether `rowmajor` or `columnmajor` ...

However, in columnmajor (default), first two graphs are placed in
the first "row" from left to right. Is it correct behavior ?

+========================================================+
 Shigeharu TAKENO     NIigata Institute of Technology
                       kashiwazaki,Niigata 945-1195 JAPAN
 sh...@ie...   TEL(&FAX): +81-257-22-8161
+========================================================+

Re: "Bind all <key> <command>" patch on SourceForge

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

On Tuesday 04 January 2005 02:50 pm, Jacques Bouchard wrote:
> > 
> > File wait_for_End would contain two lines:
> >    pause -1
> >    if (!DONE) reread
> 
> That fits exactly my need.

I am glad to hear that it works for your application.
It is a bit ugly, however, since it burns a lot of CPU
cycles in that wait loop (the "pause -1" returns immediately,
as you already know, and so the loop is very tight). 
I hope someone can think of a better way.

> Nice thing to write in a gnuplot HOWTO.

One thing at a time.  I think there is more work to do on
this patchset before it is ready for documentation and 
inclusion in the main source tree.  I do not understand
the details of what goes wrong with the "pause" command
in your example, and without understanding it is hard to
develop a fix.


-- 
Ethan A Merritt       merritt@u.washington.edu
Biomolecular Structure Center
Mailstop 357742
University of Washington, Seattle, WA 98195

Re: "Bind all <key> <command>" patch on SourceForge

[Jacques B. <jac...@no...>]

Ethan Merritt wrote:
> When you say "batch mode" do you mean:
> 	gnuplot inputfile.dat
> or
> 	gnuplot < inputfile.dat
> or
> 	gnuplot
> 	gnuplot> load "inputfile.dat"

I mean the 1st method.

> For me, the 1st and 3d methods work as expected, but the
> 2nd method does not.

Hmmm.

> If you must use the 2nd method,
> perhaps you could use a loop-until-ready construct like this?
> 
> main script:
>    set mouse
>    n=1; set term x11 n; plot x**n
>    n=2; set term x11 n; plot x**n
>    bind all Home "n=MOUSE_KEY_WINDOW; set term x11 n; plot x**n"
>    bind all End "DONE = 1"
>    DONE = 0
>    print "Hit <End> in any plot window to continue"
>    load "wait_for_End"
> 
> File wait_for_End would contain two lines:
>    pause -1
>    if (!DONE) reread

That fits exactly my need.
Nice thing to write in a gnuplot HOWTO.

And how do you do it with the 1st method?

Re: "Bind all <key> <command>" patch on SourceForge

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

On Tuesday 28 December 2004 02:10 pm, Jacques Bouchard wrote:
> 
> It works well. There is just one thing that could be improved. As I said
> in my last message, in batch mode, each time a binded key is pressed, a
> "pause -1" is eaten if the binded command contains:
> 
> set terminal x11
> 
> Doesn't that prevent you to reactivate a x11 terminal as many times as
> you wish with the "bind all" technique? Is it possible to correct that
> seemingly wrong behavior?

This may be very difficult.
But if I understood better what your application needs to do,
maybe I could suggest or create a way around the problem.
When you say "batch mode" do you mean:
	gnuplot inputfile.dat
or
	gnuplot < inputfile.dat
or
	gnuplot
	gnuplot> load "inputfile.dat"
or something else I haven't thought of?

For me, the 1st and 3d methods work as expected, but the
2nd method does not.   

If you must use the 2nd method,
perhaps you could use a loop-until-ready construct like this?

main script:
   set mouse
   n=1; set term x11 n; plot x**n
   n=2; set term x11 n; plot x**n
   bind all Home "n=MOUSE_KEY_WINDOW; set term x11 n; plot x**n"
   bind all End "DONE = 1"
   DONE = 0
   print "Hit <End> in any plot window to continue"
   load "wait_for_End"

File wait_for_End would contain two lines:
   pause -1
   if (!DONE) reread

Or maybe it would have to be "pause mouse key" or "pause 1"
or I don't know exactly what would fit in with your intended 
use.


 
> A further step forward would be to allow the user to "bind all" mouse
> events. For example, a left button press is unused for 2D graphs, and
> would be a very natural way to reactivate a x11 terminal. But of course
> that is not essential and probably not worth a difficult patch.
> 
> Thank you again.
> 

-- 
Ethan A Merritt       merritt@u.washington.edu
Biomolecular Structure Center
Mailstop 357742
University of Washington, Seattle, WA 98195

Re: "Bind all <key> <command>" patch on SourceForge

[Jacques B. <jac...@no...>]

Ethan Merritt wrote:
> Please try out patchset #1090199 on SourceForge.
> It defines a new command option:
>     bind all <key> "command string"

Thank you very much for your _quick_ answer.
I consider this patch as a _big_ step forward.

> Please let me know how it works for your application, 
> and if you have any ideas for improving it.

It works well. There is just one thing that could be improved. As I said
in my last message, in batch mode, each time a binded key is pressed, a
"pause -1" is eaten if the binded command contains:

set terminal x11

Doesn't that prevent you to reactivate a x11 terminal as many times as
you wish with the "bind all" technique? Is it possible to correct that
seemingly wrong behavior?

A further step forward would be to allow the user to "bind all" mouse
events. For example, a left button press is unused for 2D graphs, and
would be a very natural way to reactivate a x11 terminal. But of course
that is not essential and probably not worth a difficult patch.

Thank you again.

"Bind all <key> <command>" patch on SourceForge (was: x11 terminal activation)

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

On Tuesday 21 December 2004 10:33 am, Jacques Bouchard wrote:
> Ethan Merritt <merritt <at> u.washington.edu> writes:
> > Would it be sufficient to reserve a variable for this returned window
> > id parameter? 

> That would be great.

Please try out patchset #1090199 on SourceForge.
It defines a new command option:
    bind all <key> "command string"

"bind all" causes this binding to apply to all existing plot windows,
active or not.  When <key> is pressed in any plot window,
the corresponding window ID is returned in MOUSE_KEY_WINDOW and
"command string" is executed.

So your example case would become

   bind all PageUp 'set term x11 MOUSE_KEY_WINDOW; \
                    load "dem.".MOUSE_KEY_WINDOW'

Now PageUp should refresh the specific plot assigned
to the window you typed it in.

Please let me know how it works for your application, 
and if you have any ideas for improving it.


-- 
Ethan A Merritt
Biomolecular Structure Center
University of Washington 98195-7742

Re: Question about event handling in pm terminal

[Petr M. <mi...@ph...>]

> struct gp_event_t {
>     int type;           /* see below */
>     int mx, my;         /* current mouse coordinates */
>     int par1, par2;     /* other parameters, depending on the event type */
>     char text[100];     /* literal command string for type=GE_cmd */
> };
>
>
> As of this morning, there are exactly zero uses of text[] in gnuplot
> code, and no instances that generate event types GE_stderr, GE_stdout,
> or GE_cmd.
>
> I propose to remove all of these

The Presentation Manager terminal has a menu entry, where you can put a
command for gnuplot, and that gets passed into gnuplot and executed. You can
enjoy this feature when piping data to gnuplot, where the interactive
terminal prompt is not available. This command works via the shared memory
and input_from_PM_Terminal, as all the data from gnupmdrv.exe go to
gnuplot.exe.

That field "text[100]" was intended for a similar functionality in X11/Mac
et al terminals. As X11 still lacks any menus and user buttons, it was
never implemented.


> but there is one bit of OS2-specific code that puzzles me.  At the very
> end of src/gpexecute.c is this:
>
> /* OS/2 communication via shared memory; coded according to gp_execute() */
> #ifdef OS2_IPC
>     if (input_from_PM_Terminal == NULL)
>         return;
>     ((char *) input_from_PM_Terminal)[0] = '%';
>     /* copy the command to shared memory */
>     memcpy(((char *) input_from_PM_Terminal) + 1, &ge, sizeof(ge));
>     if (pausing) {              /* no communication during pause */
>         /* DosBeep(440,111); */
>         ((char *) input_from_PM_Terminal)[0] = 0;
>         return;
>     }
>     gp_post_shared_mem();
> #endif
>
> Who is at the other end of this communication channel?

It is in command.c, where the reader runs in a thread (the other thread is
normal readline). That's why there are no such communication nightmares in
OS/2 as there were with the select() and stdin and others under unixes.
Search that file for "input_from_PM_Terminal".


> Does it go to some external program that thinks it knows sizeof(gp_event_t)?
> Is struct gp_event_t (including the 100 char text field) a native
> structure in OS2 that gnuplot tries to be compatible with?

No, it's not necessary.


> Petr:
>
> Could you check whether the attached small patch removing unused
> commands and text array breaks the pm terminal?

The current cvs compiles OK on OS/2.

---
PM

Question about event handling in pm terminal

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

The structure gnuplot uses for event handling is defined in mousecmn.h 

struct gp_event_t {
    int type;           /* see below */
    int mx, my;         /* current mouse coordinates */
    int par1, par2;     /* other parameters, depending on the event type */
    char text[100];     /* literal command string for type=GE_cmd */
};


As of this morning, there are exactly zero uses of text[] in gnuplot
code, and no instances that generate event types GE_stderr, GE_stdout,
or GE_cmd.

I propose to remove all of these, but there is one bit of OS2-specific
code that puzzles me.  At the very end of src/gpexecute.c is this:

/* OS/2 communication via shared memory; coded according to gp_execute() */
#ifdef OS2_IPC                  
    if (input_from_PM_Terminal == NULL)
        return;
    ((char *) input_from_PM_Terminal)[0] = '%';
    /* copy the command to shared memory */
    memcpy(((char *) input_from_PM_Terminal) + 1, &ge, sizeof(ge));
    if (pausing) {              /* no communication during pause */
        /* DosBeep(440,111); */
        ((char *) input_from_PM_Terminal)[0] = 0;
        return;
    }
    gp_post_shared_mem();
#endif

Who is at the other end of this communication channel?
Does it go to some external program that thinks it knows sizeof(gp_event_t)?
Is struct gp_event_t (including the 100 char text field) a native
structure in OS2 that gnuplot tries to be compatible with?

Petr:

Could you check whether the attached small patch removing unused
commands and text array breaks the pm terminal?


-- 
Ethan A Merritt       merritt@u.washington.edu
Biomolecular Structure Center
Mailstop 357742
University of Washington, Seattle, WA 98195

Re: x11 terminal activation

[Jacques B. <jac...@no...>]

Ethan Merritt <merritt <at> u.washington.edu> writes:
> Would it be sufficient to reserve a variable for this returned window id
> parameter?  Then your bind command would look something like this:
> 
> bind End 'n = sprintf("%d",MOUSE_WINDOW_ID);  set term x11 MOUSE_WINDOW_ID; 
> load "plot.".n' 

That would be great.

BTW this script, executed as a _batch_, exits (should it?) if you press 'Right':

n=1; set term x11 n; plot x**n
n=2; set term x11 n; plot x**n
bind Right "n=3-n; set term x11 n; plot x**n"
pause -1

but this one doesn't:

n=1; plot x**n
bind Right "n=3-n; plot x**n"
pause -1

Re: x11 terminal activation

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

On Monday 20 December 2004 12:49 am, Jacques Bouchard wrote:
> If there were a way to trigger the pause command from a non-active X11
> terminal (and then to know its id), it would be possible to activate it
> (by replotting it).

It would be possible to pass back keypress events from keys set with
the "bind" command, including a parameter that contains the id of the
window in which the key was pressed.  Using that parameter in the
bound command would be a bit tricky, however.  

Hmmmm.
Would it be sufficient to reserve a variable for this returned window id
parameter?  Then your bind command would look something like this:

bind End 'n = sprintf("%d",MOUSE_WINDOW_ID);  set term x11 MOUSE_WINDOW_ID; 
load "plot.".n' 

There are probably a number of things that will cause this to fail in
the current code, but I think they could all be made to work.


-- 
Ethan A Merritt
Biomolecular Structure Center
University of Washington 98195-7742

Re: plotting circles

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

On Sunday 19 December 2004 07:02 am, Carolina Brito wrote:
> Hello!
>=20
> I am trying to draw circles arount points and I
> cannot!

Many terminal types have both "circle" and=20
"circle around a point" symbols available.
=46or example,=20

   set term post solid color
   plot 'foo' using 1:2:3 with points pt 6 ps variable

will plot a point with a circle around it at the (x,y)
coordinates given in columns 1 and 2, with the radius
of the circle given in column 3.  (CVS version only).

If you want a solid circle instead, use pt 7




> I am thinking about parametric functions like=20
> f(x) =3D sqrt(1-x**2) and g(x)=3D-sqrt(1-x**2)
> and then to use "filledcurves  closed" but it's not
> working.
> Do you can help me?
>=20
> Thank you!
> Carolina.
>=20
>=20
> =09
> =09
> 	=09
> _______________________________________________________=20
> Yahoo! Acesso Gr=E1tis - Instale o discador do Yahoo! agora. http://br.ac=
esso.yahoo.com/ - Internet r=E1pida e gr=E1tis
>=20
>=20
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real users.
> Discover which products truly live up to the hype. Start reading now.=20
> http://productguide.itmanagersjournal.com/
> _______________________________________________
> gnuplot-beta mailing list
> gnu...@li...
> https://lists.sourceforge.net/lists/listinfo/gnuplot-beta
>=20

=2D-=20
Ethan A Merritt       merritt@u.washington.edu
Biomolecular Structure Center
Mailstop 357742
University of Washington, Seattle, WA 98195

Re: x11 terminal activation

[Jacques B. <jac...@no...>]

> Not by replotting ... you have to re-run the entire script that generates it.

Yes. When I said "replotting", I didn't mean using the command "replot".

Re: plotting circles

[Petr M. <mi...@ph...>]

> I am trying to draw circles arount points and I
> cannot!
> I am thinking about parametric functions like
> f(x) = sqrt(1-x**2) and g(x)=-sqrt(1-x**2)
> and then to use "filledcurves  closed" but it's not
> working.
> Do you can help me?

These are not parametric curves, but this works:
  set samples 2000
  plot f(x) with filledcurves y1=0, g(x) with filledcurv y1=0

---
PM

Re: x11 terminal activation

[Hans-Bernhard B. <br...@ph...>]

On Mon, 20 Dec 2004, Jacques Bouchard wrote:

> Hello Ethan,
> Thanks for your answer.
> 
> So if I understand well, several X11 terminals may be displayed but only 
>   the last one is really _alive_, ie will react properly to mouse events 
> (pause command triggering, zooming, etc.). With the other X11 terminals, 
> you can only:
> - resize the window
> - have the current mouse coordinates displayed (but not with a 
> configurable format)
> 
> If there were a way to trigger the pause command from a non-active X11 
> terminal (and then to know its id), it would be possible to activate it 
> (by replotting it).

Not by replotting.  The entire information needed to do that replot is
lost after you make some other plot.  The bottleneck is not the window
state, or its connection to gnuplot, but the fact that the gnuplot core
engine itself doesn't have a way of controlling more than one plot at a
time.  I.e. you can't regenerate any other plot than the most recent one
by 'replot' --- you have to re-run the entire script that generates it.

-- 
Hans-Bernhard Broeker (br...@ph...)
Even if all the snow were burnt, ashes would remain.

Re: plotting circles

[Hans-Bernhard B. <br...@ph...>]

On Sun, 19 Dec 2004, Carolina Brito wrote:

> Hello!
> 
> I am trying to draw circles arount points and I
> cannot!
> I am thinking about parametric functions like 
> f(x) = sqrt(1-x**2) and g(x)=-sqrt(1-x**2)

Those aren't actually "parametric" functions.  In particular, they're not
closed curves.  I think you'ld have more luck with an actual closed, 
parametric curve:

	set parametric
	set trange [0:2*pi]
	plot cos(t),sin(t) with filledcurves closed

-- 
Hans-Bernhard Broeker (br...@ph...)
Even if all the snow were burnt, ashes would remain.

Re: PalmOne

[Hans-Bernhard B. <br...@ph...>]

On Thu, 16 Dec 2004, helena wrote:

> Can I run gnuplot on my PalmOne?

Not to our best knowledge.  gnuplot is designed to be extremely portable,
but there's limits to even that.

-- 
Hans-Bernhard Broeker (br...@ph...)
Even if all the snow were burnt, ashes would remain.

Re: x11 terminal activation

[Jacques B. <jac...@no...>]

Hello Ethan,
Thanks for your answer.

So if I understand well, several X11 terminals may be displayed but only 
  the last one is really _alive_, ie will react properly to mouse events 
(pause command triggering, zooming, etc.). With the other X11 terminals, 
you can only:
- resize the window
- have the current mouse coordinates displayed (but not with a 
configurable format)

If there were a way to trigger the pause command from a non-active X11 
terminal (and then to know its id), it would be possible to activate it 
(by replotting it).

So right now, I can't display several graphs each in an _interactive_ 
X11 terminal. I use a single X11 terminal to view several graphs one by 
one, and I switch between them with the keyboard:

n=1
nplots=...
load dem.n
bind Home     "n=1; load dem.n"
bind End      "n=nplots; load dem.n"
bind PageDown "n=n+1; if(n>nplots)n=1; load dem.n; else load dem.n"
bind PageUp   "n=n-1; if(n<1)n=nplots; load dem.n; else load dem.n"
pause -1