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

Re: more oddities on wxt multiplot and legend toggle

[<pl...@pi...>]

On 02/09/12 18:35, Ethan A Merritt wrote:
> On Thursday, February 09, 2012 06:35:21 am pl...@pi... wrote:
>> Hi,
>>
>> I have some more oddities on the legend toggle feature in multiplot.
>>
>> I have a 2x2 plot with four lines in each, hence four legend entries.
>>
>> First I note that the toggle feature is affecting all graphs at the same
>> time. This is not too bad as it happens for my instance but generally
>> would not make sence because there is not necessarily any link between
>> line1 on plot 1 and line1 on plot 2.
>
> To the best of my knowledge, this was fixed by the following ChangeLog
> entry:
>
> 2011-12-10  Ethan A Merritt<merritt@u.washington.edu>
>          * src/term.c src/wxterminal/wxt_gui.cpp term/svg.trm:
>          Allow toggling individual plots on/off inside a multiplot page.
>          Bugfix.
>
>> The main oddity displays serveral anomalies that I suspect have the same
>> single root cause.
>>
>> a) I have a legend entry that is "period = FFT window" . If I click at
>> the begining of the legend text it does nothing (usual gnuplot not
>> really knowing text extents , I guess).
>
> The size of the "hot box" around the text is only an approximation,
> since most terminals do not provide feedback on the exact bounding box
> of rendered text.
>
>> If I click progessively further along more an more lines coming into the
>> toggle.
>> (on all four plots) . This seems to be a function if whether my
>> click.x.coord is within their legend text (estimated) extent.
>>
>> If I click far enough along to be in everyone's (extimated) text extent
>> it toggles  all four plots  for the line I clicked on AND ANY lines
>> above it in the legend.
>>
>> The active area here extends right across to the end if the
>> corresponding legend line segment in TOP, RIGHT plot. ie end of the 1,2
>> position plot's legend box.
>>
>> ie If I click on second legend entry it affects lines 1&2 on all four
>> plots. If I click on legend item 4 in 1,1 plot I get all four lines
>> toggling in all four plots !!
>>
>> The fun continues if I go below the line. It's almost the opposite. Now
>> it toggles any line LOWER than where I click. The same general rules
>> seem to apply to position with the added twist that the 2,2 postition
>> plot has a top , left key position. I can now see that the active area
>> goes well beyond the end of the key box and almost the end of the plot
>> area.
>>
>> In fact this end seems to match the RHS of where a top,right legend box
>> would be (it may be picking up a value from plot 1,2 legend box.).
>>
>> The bottom line of all this madness it that legend coordinate detection
>> seems totally confused in the case of multiplot. I doubt it's a big
>> issue to fix but it has some crazy (and amuzing) results.
>>
>> regards, Peter.
>
> This all sounds familiar, but I think it was fixed a couple of months ago.
> Could you please tell us the wxt version and the exact gnuplot build date?
> Does it work correctly with svg output?
>
> 	Ethan
>
>

Thanks Ethan.

as I said on an earlier issue, my current build is nov 2011 so that may 
well have been fixed since. Though I don't want to update until my 
current report is finished unless there is something stopping me working.

I will check and report back if I still see oddities next time I update.

Peter.

Re: mouse coord read-out in wxt multiplot

[Ethan A M. <sf...@us...>]

On Thursday, February 09, 2012 09:22:50 am pl...@pi... wrote:
> On 02/09/12 15:45, pl...@pi... wrote:
> > Hi,
> >
> > quick one.
> >
> > coordinate readout in wxt multiplot is on "whole window" coord. system,
> > centred at screen centre.
> >
> > This seems rather unhelpful since it does not relate to ANY of the
> > plots' axes.
> >
> > Wouldn't this be a more useful feature if it translated mouse to 'local'
> > axes of the hovered plot.

This is a very long-standing Feature Request.

The nature of the problem is that the generic mousing code uses internal
variables that are only correct for the "current" plot (actually the 
plot that was just completed).  Since these variables are overwritten with
each new plot in a multiplot, the information is not available except for
the most recent subpanel.

Relatively recently we added terminal-specific code to dump plot layout
information so that mousing could be done by terminals like svg and canvas
where the information was necessarily stored in the output file itself.
As each plot is completed, the summary variables (essentially the ones
reported in GPVAL_TERM_XXX) are written into a data structure.

That provides a template for how we might do the same for the interactive
terminals.  I think it would have to be tackled one terminal at a time,
but it sounds like a relatively straightforward programming task.
If anyone is looking for a project .... :-)

	Ethan

Re: gnuplot-mode and Gnuplot licensing question

[Ethan A M. <sf...@us...>]

> > 
> > I'd say the better method would be to write a quick doc2tooltip.c
> > program to parallel the others.  It would look for the string
> >   Syntax:
> > in gnuplot.doc and extract the following lines into a tooltip
> > collection.
> 
> 
> Indeed, that's a better idea; thanks for the suggestion. Now that I
> look at the docs/ directory I see there is already a `doc2texi.el'
> contributed by Bruce Ravel, the author of gnuplot-mode. It would
> probably make the most sense for me to modify that to output an Elisp
> file of tooltips along with the texinfo file. Any objections?

No objections. Just a warning that the code in doc2texi.el may itself 
be a bit rusty.  I seem to recall some reports of problems with the
*.info file involving infinite loops of redirected links or something
like that.

	Ethan

Re: more oddities on wxt multiplot and legend toggle

[Ethan A M. <sf...@us...>]

On Thursday, February 09, 2012 06:35:21 am pl...@pi... wrote:
> Hi,
> 
> I have some more oddities on the legend toggle feature in multiplot.
> 
> I have a 2x2 plot with four lines in each, hence four legend entries.
> 
> First I note that the toggle feature is affecting all graphs at the same 
> time. This is not too bad as it happens for my instance but generally 
> would not make sence because there is not necessarily any link between 
> line1 on plot 1 and line1 on plot 2.

To the best of my knowledge, this was fixed by the following ChangeLog 
entry:

2011-12-10  Ethan A Merritt  <merritt@u.washington.edu>
        * src/term.c src/wxterminal/wxt_gui.cpp term/svg.trm:
        Allow toggling individual plots on/off inside a multiplot page.
        Bugfix.

> The main oddity displays serveral anomalies that I suspect have the same 
> single root cause.
> 
> a) I have a legend entry that is "period = FFT window" . If I click at 
> the begining of the legend text it does nothing (usual gnuplot not 
> really knowing text extents , I guess).

The size of the "hot box" around the text is only an approximation, 
since most terminals do not provide feedback on the exact bounding box
of rendered text.

> If I click progessively further along more an more lines coming into the 
> toggle.
> (on all four plots) . This seems to be a function if whether my 
> click.x.coord is within their legend text (estimated) extent.
>
> If I click far enough along to be in everyone's (extimated) text extent 
> it toggles  all four plots  for the line I clicked on AND ANY lines 
> above it in the legend.
> 
> The active area here extends right across to the end if the 
> corresponding legend line segment in TOP, RIGHT plot. ie end of the 1,2 
> position plot's legend box.
> 
> ie If I click on second legend entry it affects lines 1 &2 on all four 
> plots. If I click on legend item 4 in 1,1 plot I get all four lines 
> toggling in all four plots !!
> 
> The fun continues if I go below the line. It's almost the opposite. Now 
> it toggles any line LOWER than where I click. The same general rules 
> seem to apply to position with the added twist that the 2,2 postition 
> plot has a top , left key position. I can now see that the active area 
> goes well beyond the end of the key box and almost the end of the plot 
> area.
> 
> In fact this end seems to match the RHS of where a top,right legend box 
> would be (it may be picking up a value from plot 1,2 legend box.).
> 
> The bottom line of all this madness it that legend coordinate detection 
> seems totally confused in the case of multiplot. I doubt it's a big 
> issue to fix but it has some crazy (and amuzing) results.
> 
> regards, Peter.

This all sounds familiar, but I think it was fixed a couple of months ago.
Could you please tell us the wxt version and the exact gnuplot build date?
Does it work correctly with svg output?

	Ethan

Re: gnuplot-mode and Gnuplot licensing question

[Jonathan O. <j.j...@gm...>]

> 
> I'd say the better method would be to write a quick doc2tooltip.c
> program to parallel the others.  It would look for the string
>   Syntax:
> in gnuplot.doc and extract the following lines into a tooltip
> collection.


Indeed, that's a better idea; thanks for the suggestion. Now that I
look at the docs/ directory I see there is already a `doc2texi.el'
contributed by Bruce Ravel, the author of gnuplot-mode. It would
probably make the most sense for me to modify that to output an Elisp
file of tooltips along with the texinfo file. Any objections?

I would hope that would also nullify the license conflict problem, but
someone correct me if I'm wrong.

> Probably there are some commands that don't have this keyword
> in the current *.doc file, but it seems reasonable to standardize
> them  if it helps a real-world use.


I agree -- if anyone uses this enough to be annoyed by its limitations in
the future, it can be addressed then.


Jonathan

Re: mouse coord read-out in wxt multiplot

[<pl...@pi...>]

On 02/09/12 15:45, pl...@pi... wrote:
> Hi,
>
> quick one.
>
> coordinate readout in wxt multiplot is on "whole window" coord. system,
> centred at screen centre.
>
> This seems rather unhelpful since it does not relate to ANY of the
> plots' axes.
>
> Wouldn't this be a more useful feature if it translated mouse to 'local'
> axes of the hovered plot.
>
> similarly , if I create an on-screen annotation with a mouse button
> click , I get a 'global' coord that has no meaning on the graph where it
> is applied.
>
> This makes this rather nifty feature pretty much useless in the
> multiplot context. :(
>
> Peter.


PS.  I don't want to give the impression I'm not happy with all this. It 
has a lot of really neat features which is why it would be good to get 
the wrinkles ironed out. It seems some of this has not had thorough , 
real use testing yet.

I hope these reports will identify some rough edges and corner cases 
that have been missed.

Apart from that it's bloody amazing at what I am able do with gnuplot . 
I'm producing some quite complex multiplot output of report quality that 
I can flick between a live wxt terminal and png file output at the flick 
of a variable (do_png=1).

gnuplot is also calling bash scripts that get the data from source, 
process, filter and FFT it .

The whole process in nearly down to one gnuplot load command now. I am 
impressed !


Peter.

more oddities on wxt multiplot and legend toggle

[<pl...@pi...>]

Hi,

I have some more oddities on the legend toggle feature in multiplot.

I have a 2x2 plot with four lines in each, hence four legend entries.

First I note that the toggle feature is affecting all graphs at the same 
time. This is not too bad as it happens for my instance but generally 
would not make sence because there is not necessarily any link between 
line1 on plot 1 and line1 on plot 2.


The main oddity displays serveral anomalies that I suspect have the same 
single root cause.

a) I have a legend entry that is "period = FFT window" . If I click at 
the begining of the legend text it does nothing (usual gnuplot not 
really knowing text extents , I guess).

If I click progessively further along more an more lines coming into the 
toggle.
(on all four plots) . This seems to be a function if whether my 
click.x.coord is within their legend text (estimated) extent.

If I click far enough along to be in everyone's (extimated) text extent 
it toggles  all four plots  for the line I clicked on AND ANY lines 
above it in the legend.

The active area here extends right across to the end if the 
corresponding legend line segment in TOP, RIGHT plot. ie end of the 1,2 
position plot's legend box.

ie If I click on second legend entry it affects lines 1 &2 on all four 
plots. If I click on legend item 4 in 1,1 plot I get all four lines 
toggling in all four plots !!

The fun continues if I go below the line. It's almost the opposite. Now 
it toggles any line LOWER than where I click. The same general rules 
seem to apply to position with the added twist that the 2,2 postition 
plot has a top , left key position. I can now see that the active area 
goes well beyond the end of the key box and almost the end of the plot 
area.

In fact this end seems to match the RHS of where a top,right legend box 
would be (it may be picking up a value from plot 1,2 legend box.).

The bottom line of all this madness it that legend coordinate detection 
seems totally confused in the case of multiplot. I doubt it's a big 
issue to fix but it has some crazy (and amuzing) results.

regards, Peter.

Re: gnuplot-mode and Gnuplot licensing question

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

On Thursday, 09 February 2012, Jonathan Oddie wrote:
> 
> Hi Ethan,
> 
> >> 
> >> Could anyone tell me whether this would be considered acceptable use?
> > 
> > I don't understand the intended mechanism.
> > 
> 
> Sorry I wasn't clear. No, I'm not intending to maintain a separate set
> of documentation. (I use the info viewer inside Emacs for reading
> that, and the code I'm working on creates links to it from the source
> buffer). What I'm thinking of are more like what are called "tooltips"
> in other environments. In other Emacs language modes (C, Lisp, etc.)
> you can have the editor show a one-line message at the bottom of the
> screen listing the arguments to a function as soon as you type it or
> move the cursor over it, and this saves looking up the full
> manual. E.g. when the cursor is after "smooth" Emacs displays
> 
> smooth {unique | frequency | cumulative | kdensity ... }
> 
> or something similar. Or after typing "using" in a plot command with a
> particular plot style it lists the order of columns expected. I find
> this quite useful as I often forget the order of arguments to some
> options.
> 
> I guess this is a maintenance concern to an extent since the
> documentation strings would be in the Elisp source (they can't really
> be extracted automatically from the info page, as far as I can see). I
> think the benefit might outweigh the cost. What do you think?
> 
> Jonathan

Ugh. Not from the info page.  Does anyone really use that?

I'd say the better method would be to write a quick doc2tooltip.c
program to parallel the others.  It would look for the string
   Syntax:
in gnuplot.doc and extract the following lines into a tooltip
collection.

Probably there are some commands that don't have this keyword
in the current *.doc file, but it seems reasonable to standardize
them  if it helps a real-world use.

	Ethan

Re: when did "set decimalsign" come in?

[Lars H. <lhe...@us...>]

Ethan Merritt writes:
> On Wednesday, 08 February 2012, Allin Cottrell wrote:
> > Does anyone happen to know, with which gnuplot release did the "set 
> > decimalsign" option come in? That is, what's the earliest gnuplot 
> > version that supports this "set" variable?
> 
> CVS: 2002-02-13
> official release: 4.0 patchlevel 0  (April 2004)
> 
> I have not kept copies of the 3.8 snapshots, but just going by
> the dates it must have gone in before 3.8i.

 The ChangeLog places addition of the patch between 3.8i and 3.8j, so 3.8j
 was the first snapshot to support it. I think I may have copies of them
 in an archive at home somewhere :)

mouse coord read-out in wxt multiplot

[<pl...@pi...>]

Hi,

quick one.

coordinate readout in wxt multiplot is on "whole window" coord. system, 
centred at screen centre.

This seems rather unhelpful since it does not relate to ANY of the 
plots' axes.

Wouldn't this be a more useful feature if it translated mouse to 'local' 
axes of the hovered plot.

similarly , if I create an on-screen annotation with a mouse button 
click , I get a 'global' coord that has no meaning on the graph where it 
is applied.

This makes this rather nifty feature pretty much useless in the 
multiplot context. :(

Peter.

Re: when did "set decimalsign" come in?

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

On Wed, 8 Feb 2012, Ethan Merritt wrote:

> On Wednesday, 08 February 2012, Allin Cottrell wrote:
>> Does anyone happen to know, with which gnuplot release did the "set
>> decimalsign" option come in? That is, what's the earliest gnuplot
>> version that supports this "set" variable?
>
> CVS: 2002-02-13
> official release: 4.0 patchlevel 0  (April 2004)
>
> I have not kept copies of the 3.8 snapshots, but just going by
> the dates it must have gone in before 3.8i.

Great, thanks.

Allin Cottrell

Re: gnuplot-mode and Gnuplot licensing question

[Jonathan O. <j.j...@gm...>]

Hi Ethan,

>> 
>> Could anyone tell me whether this would be considered acceptable use?
> 
> I don't understand the intended mechanism.
> 

Sorry I wasn't clear. No, I'm not intending to maintain a separate set
of documentation. (I use the info viewer inside Emacs for reading
that, and the code I'm working on creates links to it from the source
buffer). What I'm thinking of are more like what are called "tooltips"
in other environments. In other Emacs language modes (C, Lisp, etc.)
you can have the editor show a one-line message at the bottom of the
screen listing the arguments to a function as soon as you type it or
move the cursor over it, and this saves looking up the full
manual. E.g. when the cursor is after "smooth" Emacs displays

smooth {unique | frequency | cumulative | kdensity ... }

or something similar. Or after typing "using" in a plot command with a
particular plot style it lists the order of columns expected. I find
this quite useful as I often forget the order of arguments to some
options.

I guess this is a maintenance concern to an extent since the
documentation strings would be in the Elisp source (they can't really
be extracted automatically from the info page, as far as I can see). I
think the benefit might outweigh the cost. What do you think? If
people aren't convinced I'll leave it out and just have the completion
code + links to the info manual.

Jonathan

> 
> If you're saying that you want to maintain a new, parallel, set of
> documentation, I think that's a bad idea.  It would be far too much
> trouble to keep it in sync.
> 
> 	Ethan
> 
> 
>> 
>> Thanks,
>> Jonathan
>> 
>> 
>> ------------------------------------------------------------------------------
>> Virtualization & Cloud Management Using Capacity Planning
>> Cloud computing makes use of virtualization - but cloud computing 
>> also focuses on allowing computing to be delivered as a service.
>> http://www.accelacomm.com/jaw/sfnl/114/51521223/
>> _______________________________________________
>> gnuplot-beta mailing list
>> gnu...@li...
>> https://lists.sourceforge.net/lists/listinfo/gnuplot-beta
>> 
> 

Re: gnuplot-mode and Gnuplot licensing question

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

On Wednesday, 08 February 2012, Jonathan Oddie wrote:
> Hi Gnuplot developers,
> 
> I've been working on an add-on module for Bruce Ravel's Emacs
> gnuplot-mode, to add context-sensitive completion and documentation
> lookup for Gnuplot script buffers. It does a fairly complete parse of
> the Gnuplot language so it should hopefully be quite useful once I
> finish writing the rest of the grammar rules.
> 
> I'd like to release this sometime soon and my question has to do with
> licensing. I would like my Elisp code to be GPL'ed (in fact it
> probably has to be since it extends gnuplot.el, which is
> GPL'ed). However, it has the ability to show short documentation
> strings ("ElDoc") to display in the mode line depending on what's at
> point, and the easiest way of getting these would be to copy them out
> of the Gnuplot info manual. I know Gnuplot itself is not under the GPL
> and so I'm not sure whether this would be against its license terms.
> 
> Could anyone tell me whether this would be considered acceptable use?

I don't understand the intended mechanism.

Are you saying that this new code would act as a viewer for the
existing documentation?  That would not introduce any licensing issues
that I can think of.  You can use any viewer you like.  I'm sure 
some people use the Acrobat reader to view the PDF documentation,
for instance, and Acrobat certainly isn't under the gnuplot license.

If you're saying that you want to maintain a new, parallel, set of
documentation, I think that's a bad idea.  It would be far too much
trouble to keep it in sync.

	Ethan


> 
> Thanks,
> Jonathan
> 
> 
> ------------------------------------------------------------------------------
> Virtualization & Cloud Management Using Capacity Planning
> Cloud computing makes use of virtualization - but cloud computing 
> also focuses on allowing computing to be delivered as a service.
> http://www.accelacomm.com/jaw/sfnl/114/51521223/
> _______________________________________________
> gnuplot-beta mailing list
> gnu...@li...
> https://lists.sourceforge.net/lists/listinfo/gnuplot-beta
> 

Re: when did "set decimalsign" come in?

[Mojca M. <moj...@gm...>]

On Thu, Feb 9, 2012 at 02:36, Allin Cottrell  wrote:
> Does anyone happen to know, with which gnuplot release did the "set
> decimalsign" option come in? That is, what's the earliest gnuplot
> version that supports this "set" variable?

The patch was added into CVS on February 13, 2002, but I didn't check
which version of gnuplot included it first.

Mojca

Re: when did "set decimalsign" come in?

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

On Wednesday, 08 February 2012, Allin Cottrell wrote:
> Does anyone happen to know, with which gnuplot release did the "set 
> decimalsign" option come in? That is, what's the earliest gnuplot 
> version that supports this "set" variable?

CVS: 2002-02-13
official release: 4.0 patchlevel 0  (April 2004)

I have not kept copies of the 3.8 snapshots, but just going by
the dates it must have gone in before 3.8i.

	Ethan

when did "set decimalsign" come in?

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

Does anyone happen to know, with which gnuplot release did the "set 
decimalsign" option come in? That is, what's the earliest gnuplot 
version that supports this "set" variable?

Thanks.

--
Allin Cottrell
Department of Economics
Wake Forest University

gnuplot-mode and Gnuplot licensing question

[Jonathan O. <j.j...@gm...>]

Hi Gnuplot developers,

I've been working on an add-on module for Bruce Ravel's Emacs
gnuplot-mode, to add context-sensitive completion and documentation
lookup for Gnuplot script buffers. It does a fairly complete parse of
the Gnuplot language so it should hopefully be quite useful once I
finish writing the rest of the grammar rules.

I'd like to release this sometime soon and my question has to do with
licensing. I would like my Elisp code to be GPL'ed (in fact it
probably has to be since it extends gnuplot.el, which is
GPL'ed). However, it has the ability to show short documentation
strings ("ElDoc") to display in the mode line depending on what's at
point, and the easiest way of getting these would be to copy them out
of the Gnuplot info manual. I know Gnuplot itself is not under the GPL
and so I'm not sure whether this would be against its license terms.

Could anyone tell me whether this would be considered acceptable use?

Thanks,
Jonathan

Re: exit not working as stated in help

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

On Wednesday, February 08, 2012 10:32:28 am pl...@pi... wrote:
> HI,
> 
> unless I'm reading it wrongly or it simply unclear , it seems that exit 
> is not doing what it says on the box.

Known bug, although I don't see a tracker item for it.
Feel free to add one.
As I recall, it behaves unreliably inside nested "load" commands
as well.  I think the problem was introduced when the lf_push()/
lf_pop() code was revised, but I'm not sure whether the interaction
with block constucts is a consequence or a separate bug.

	Ethan

 
> I have a test file:
> 
> test_val=0
> if (test_val != 1) {print "*** an error was found ; exit without 
> plotting sine ***" ; exit; print "end of if"}
> plot sin(x);
> print "no errors, sine was plotted";
> 
> 
> when I load this from gnuplot  DO get the plot of sine and the following 
> output :
> 
> gnuplot> load "test.gnu"
> *** an error was found ; exit without plotting sine ***
> no errors, sine was plotted
> gnuplot>
> 
> 
> help exit :
> 
>  >>
>   The commands `exit` and `quit`, as well as the END-OF-FILE character 
> (usually
>   Ctrl-D) terminate input from the current input stream: terminal 
> session, pipe,
>   and file input (pipe).
> 
>   If input streams are nested (inherited `load` scripts), then reading will
>   continue in the parent stream.
>  >>
> 
> 
>  From this I was expecting exit to exit from the loaded script and 
> return me to the gnuplot prompt.
> 
> Instead only seems to have exited from the if block.
> 
> Does the doc need clarifying or is this an unintended consequence of the 
> new blocked if structures?
> 
> regards, Peter.
> 
> 
> 
> 
> 
> 
> ------------------------------------------------------------------------------
> Keep Your Developer Skills Current with LearnDevNow!
> The most comprehensive online learning library for Microsoft developers
> is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
> Metro Style Apps, more. Free future releases when you subscribe now!
> http://p.sf.net/sfu/learndevnow-d2d
> _______________________________________________
> gnuplot-beta mailing list
> gnu...@li...
> https://lists.sourceforge.net/lists/listinfo/gnuplot-beta
> 

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

exit not working as stated in help

[<pl...@pi...>]

HI,

unless I'm reading it wrongly or it simply unclear , it seems that exit 
is not doing what it says on the box.

I have a test file:

test_val=0
if (test_val != 1) {print "*** an error was found ; exit without 
plotting sine ***" ; exit; print "end of if"}
plot sin(x);
print "no errors, sine was plotted";


when I load this from gnuplot  DO get the plot of sine and the following 
output :

gnuplot> load "test.gnu"
*** an error was found ; exit without plotting sine ***
no errors, sine was plotted
gnuplot>


help exit :

 >>
  The commands `exit` and `quit`, as well as the END-OF-FILE character 
(usually
  Ctrl-D) terminate input from the current input stream: terminal 
session, pipe,
  and file input (pipe).

  If input streams are nested (inherited `load` scripts), then reading will
  continue in the parent stream.
 >>


 From this I was expecting exit to exit from the loaded script and 
return me to the gnuplot prompt.

Instead only seems to have exited from the if block.

Does the doc need clarifying or is this an unintended consequence of the 
new blocked if structures?

regards, Peter.

Re: should gnuplot handle NaN in input data?

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

On Monday, February 06, 2012 09:02:27 am pl...@pi... wrote:
> Hi,
> 
> it appears that gnuplot can handle most things it cannot parse in input 
> data as unplottable data. All that is  _except_ IEEE NaN.

I'm pretty sure that is not true in general.
For instance, the demo "imageNaN.dem" in the standard demo set
illustrates handling NaN in the input data stream.

Here's another simple test case:

gnuplot> set xr [0:5]; set yr [0:5]
gnuplot> plot '-' using 1:2 with lp pt 7
input data ('e' ends) > 1 1
input data ('e' ends) > 2 NaN
input data ('e' ends) > 3 3
input data ('e' ends) > 4 4
input data ('e' ends) > e


Now it is true that some support libraries do not recognize "NaN"
as legal input (including, I believe, MSVC). But that's a different
failure mode.

What exactly was your plot command?

	Ethan

> 
> Of course any specific value can be defined by 'set datafile missing' 
> but there is no default for this.
> 
> This situation has the paradoxical effect mentioned above that if I have 
> a data line that reads " ugh 0.0 " , it will be skipped as unreadable 
> but if I have " NaN 0.0 " it will cause the plot to fail with 
> "unreadable" data.
> 
> It seems it is precisely because gnuplot IS able to parse NaN that it is 
> causing an error. If it could not understand it , it would ignore the line.
> 
> Since gnuplot would effectively regard this as a NaN if it could not 
> parse it , it seems a bit illogical that having parsed "NaN" as meaning 
> NaN it refuses to act on it being NaN and triggers an error.
> 
> Is there some problem with this value, when it is parsed in input, being 
> used for what it is unless another "missing" value has been explicitly 
> defined with 'set datafile missing' ?
> 
> 
> Thanks.


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

Re: second minor bug in wxt legend toggle

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

On Monday, February 06, 2012 11:36:12 am pl...@pi... wrote:
> Hi,
> 
> another oddity.
> 
> If I put annotations on a graph using mouse button to mark coords, these 
> seem in some way linked to the last plotted line.
> 
> If I toggle visibility on the last plotted line the annotations get 
> hidden too. Not the case if I toggle any other plot lines using legend 
> toggle feature.
> 
> since the annotation could be related to any line on the graph this 
> seems accidental and unhelpful.
> 
> Also , once the last plot line is hidden, no annotations can be made 
> (they are not even made "hidden" to reappear later.)
> 
> there seems to be an unexpected linkage between the last plot line and 
> the annotations.

I can't reproduce either this or your earlier report in the current
CVS version using the wxt terminal.  I do see a vaguely similar bug, 
though not quite the same, in the qt terminal, however.
Can you confirm that your error case was using wxt rather than qt?

	Ethan

second minor bug in wxt legend toggle

[<pl...@pi...>]

Hi,

another oddity.

If I put annotations on a graph using mouse button to mark coords, these 
seem in some way linked to the last plotted line.

If I toggle visibility on the last plotted line the annotations get 
hidden too. Not the case if I toggle any other plot lines using legend 
toggle feature.

since the annotation could be related to any line on the graph this 
seems accidental and unhelpful.

Also , once the last plot line is hidden, no annotations can be made 
(they are not even made "hidden" to reappear later.)

there seems to be an unexpected linkage between the last plot line and 
the annotations.

best regards, Peter.

should gnuplot handle NaN in input data?

[<pl...@pi...>]

Hi,

it appears that gnuplot can handle most things it cannot parse in input 
data as unplottable data. All that is  _except_ IEEE NaN.

Of course any specific value can be defined by 'set datafile missing' 
but there is no default for this.

This situation has the paradoxical effect mentioned above that if I have 
a data line that reads " ugh 0.0 " , it will be skipped as unreadable 
but if I have " NaN 0.0 " it will cause the plot to fail with 
"unreadable" data.

It seems it is precisely because gnuplot IS able to parse NaN that it is 
causing an error. If it could not understand it , it would ignore the line.

Since gnuplot would effectively regard this as a NaN if it could not 
parse it , it seems a bit illogical that having parsed "NaN" as meaning 
NaN it refuses to act on it being NaN and triggers an error.

Is there some problem with this value, when it is parsed in input, being 
used for what it is unless another "missing" value has been explicitly 
defined with 'set datafile missing' ?


Thanks.

minor bug in wxt legen toggle

[<pl...@pi...>]

Hi,

I just noticed that if I select a zoom rectangle in wxt where the final 
virtex happens to land on a legend entry I get the expanded window but 
without the plot line. Rather dis-routing.

It seems that the mouse click that servers to terminate the rect 
selection is also getting processed as a click on the legend and 
triggering the toggle plot line visibility feature.

I expect this is unintentional. It's certainly unexpected.

I would have thought that one mouse click should produce one event , not 
two.

best regards, Peter.

Re: way to find out if variable has been defined

[Tait <gnu...@t4...>]

Juhász Péter <peter.juhasz83_gmail.com> said (on 2012/01/31):
> On Tue, 2012-01-31 at 18:18 +0100, plotter_piments.com wrote:
> > I have a number of gnuplot scripts that have various options set. I 
> > would like to set a default value in the script and optionally determine 
> > the value from the calling script or command.
> > 
> > This does not seem to be possible.
> > 
> > eg.
> > 
> > if (doPNG) { set terminal png; set output pngfile}
> > else { set terminal wxt }
> > 
> > plot ....
> > 
> > called as :
> > 
> > (echo 'doPNG=1;'  && cat cos_fit.gnu) | gnuplot
> > 
> > Some kind of  isdefined() function would be a solution or simply if the 
> > interpreter could return some kind of NaN value that could then be 
> > tested instead of triggering an error and breaking out.
> > 
> > Alternatively kind of error trapping mechanism but that may be going 
> > beyond what is needed for this simple test.
> > 
> > Do you think this would be a useful feature?
> > 
> > Regards,. Peter.
> 
> It's such a useful feature that it has been implemented already:
> 
> if (exists("doPNG")) { # whatever...
> 
> Look at "help exists". You'll see that, well, help exists. (Sorry, I
> couldn't resist...)
> 
> There is also "defined", but that's deprecated.
> 
> Péter Juhász


For your use case, it might make more sense to use environment
variables. For handling the output terminal, GNUTERM can already be
set to the desired terminal type.

In the more general case...

!#/bin/sh
GP_MYOUTPUT=png
export GP_MYOUTPUT
echo << __GP_EOF__ | gnuplot
print "This part is gnuplot..."
set xrange [-10:10]
set yrange [-10:10]
# etc etc
myoutput=`echo $GP_MYOUTPUT`
# do something with the gnuplot variable, myoutput
print myoutput
plot x, x**2
print "Leaving gnuplot now... bye"
__GP_EOF__

Above is an example of a shell script setting an environment variable,
GP_MYOUTPUT, that is imported into a gnuplot variable (myoutput) via a
system call (`echo $GP_MYOUTPUT`). As a gnuplot string variable, you
can do things like substr(), word(), value(), and so on.