gnuplot development

On Tuesday 29 July 2008, Theo Hopman wrote:
> Hi all. Why does `set format y "%5f"` not work like `set format y 
> "%5g"`? I'm dealing with numbers between 0 and 5, so I would expect both 
> to produce tic labels with right justified floating point numbers in a 
> field of five characters, with no trailing zeros. %5g does, but %5f 
> doesn't: it produces tic labels with six decimal places and trailing zeros.

That is the C standard.
From "man sprintf"

       f,F    The double argument is rounded and converted to decimal notation in the style
              [-]ddd.ddd,  where  the number of digits after the decimal-point character is
              equal to the precision specification.  If the precision  is  missing,  it  is
              taken  as  6; if the precision is explicitly zero, no decimal-point character
              appears.  If a decimal point appears, at least one digit appears before it.





> I've tried 4.2.3 and the latest CVS on Cygwin, and 4.3.0 (dated February 
> 2008) on Linux.
> 
> Minimal example:
> 
> reset
> plot x
> set format y "%5g"
> replot
> set format y "%5f"
> replot
> 
> According to my interpretation of the documentation, the last two plots 
> should be identical.
> 
> THeo
> 
> -------------------------------------------------------------------------
> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
> Build the coolest Linux based applications with Moblin SDK & win great prizes
> Grand prize is a trip for two to an Open Source event anywhere in the world
> http://moblin-contest.org/redirect.php?banner_id=100&url=/
> _______________________________________________
> gnuplot-beta mailing list
> gnuplot-beta@...
> https://lists.sourceforge.net/lists/listinfo/gnuplot-beta
> 



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

On Tue, 29 Jul 2008, Theo Hopman wrote:

> Hi all. Why does `set format y "%5f"` not work like `set format 
> y "%5g"`? I'm dealing with numbers between 0 and 5...

That's the way things are in C (man 3 printf).

waverley:~/cfiles/misc$ cat fg.c
#include <stdio.h>

int main (void)
{
    int i;

    for (i=0; i<3; i++) {
        printf("5f:   %5f\n", (double) i);
        printf("5g:   %5g\n", (double) i);
        printf("5.0f: %5.0f\n", (double) i);
    }

    return 0;
}
waverley:~/cfiles/misc$ gcc -Wall fg.c
waverley:~/cfiles/misc$ ./a.out 
5f:   0.000000
5g:       0
5.0f:     0
5f:   1.000000
5g:       1
5.0f:     1
5f:   2.000000
5g:       2
5.0f:     2

Allin Cottrell

On Saturday 26 July 2008, Georg Brandl wrote:
> I found this in 4.2.3, but it seems to be in the CVS version
> of gnuplot.doc too:
> 
> { formt "formatstring" } should be { format "formatstring" }.

Got it. Thanks.

-- 
Ethan A Merritt

On Thursday 24 July 2008 13:48:52 Philipp K. Janert wrote:
> 
> I think there might be a problem with 
> enhanced mode in Postscript and the 
> overlay/lifting operator ~
> 
> The following commands produce good
> results with the enhanced wxt terminal,
> but not with the enhanced postscript
> terminal:
> 
> set label 1 "~a{.7{/*.7\\~}}" at 0,0
> set label 2 "~a{.7{/*.8=}}" at .5,0
> set label 3 "~a{.6{/*1-}}" at -.5,0
> plot [-1:1][-1:1] 2
> 
> The problem appears to be that the second
> character is not properly lifted as it should be,
> when using Postscript.
> 
> Is this a known issue?

Yes. But it's more fundamental that that.
Basically you cannot nest curly braces inside the overprint
operator.  If you remove the nested braces, everything works
as expected (I think :-).

 set label 1 "~a{.7\\~}" at 0,0
 set label 2 "~a{.7=}" at .5,0
 set label 3 "~a{.6-}" at -.5,0

The overprint operator really does not work very well.
I have a list of problems from various drivers.  If you can,
it is better to use either the @ operator instead, or in a case
like your example simply use the extended character set to
draw the accented character as a single glyph.

For instance, your first label above could be

  set label 1 "@^\\~_a"
or (much better)
  set encoding iso_8859_15
  set label 1 "{\343}"

This all works better in the gd or wxt terminals than in PostScript, however.

	Ethan




> 
> Also, and I am not sure whether I consider
> this a bug or a feature, when making the
> overlayed character very large, it affects
> the absolute position of the first character
> in the enhanced wxt terminal.
> 
> (My guess is that gnuplot uses the combined
> bounding box of both characters to position
> the entire label. From a user perspective, I
> would have expected that only the non-lifted
> character determines the position.)
> 
> 
> 
> -------------------------------------------------------------------------
> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
> Build the coolest Linux based applications with Moblin SDK & win great prizes
> Grand prize is a trip for two to an Open Source event anywhere in the world
> http://moblin-contest.org/redirect.php?banner_id=100&url=/
> _______________________________________________
> gnuplot-beta mailing list
> gnuplot-beta@...
> https://lists.sourceforge.net/lists/listinfo/gnuplot-beta
> 



-- 
Ethan A Merritt

Cool. Thanks. That's good advice.

Best,

			Ph.


On Thursday 24 July 2008 14:54, you wrote:
> On Thursday 24 July 2008 13:48:52 Philipp K. Janert wrote:
> > I think there might be a problem with
> > enhanced mode in Postscript and the
> > overlay/lifting operator ~
> >
> > The following commands produce good
> > results with the enhanced wxt terminal,
> > but not with the enhanced postscript
> > terminal:
> >
> > set label 1 "~a{.7{/*.7\\~}}" at 0,0
> > set label 2 "~a{.7{/*.8=}}" at .5,0
> > set label 3 "~a{.6{/*1-}}" at -.5,0
> > plot [-1:1][-1:1] 2
> >
> > The problem appears to be that the second
> > character is not properly lifted as it should be,
> > when using Postscript.
> >
> > Is this a known issue?
>
> Yes. But it's more fundamental that that.
> Basically you cannot nest curly braces inside the overprint
> operator.  If you remove the nested braces, everything works
> as expected (I think :-).
>
>  set label 1 "~a{.7\\~}" at 0,0
>  set label 2 "~a{.7=}" at .5,0
>  set label 3 "~a{.6-}" at -.5,0
>
> The overprint operator really does not work very well.
> I have a list of problems from various drivers.  If you can,
> it is better to use either the @ operator instead, or in a case
> like your example simply use the extended character set to
> draw the accented character as a single glyph.
>
> For instance, your first label above could be
>
>   set label 1 "@^\\~_a"
> or (much better)
>   set encoding iso_8859_15
>   set label 1 "{\343}"
>
> This all works better in the gd or wxt terminals than in PostScript,
> however.
>
> 	Ethan
>
> > Also, and I am not sure whether I consider
> > this a bug or a feature, when making the
> > overlayed character very large, it affects
> > the absolute position of the first character
> > in the enhanced wxt terminal.
> >
> > (My guess is that gnuplot uses the combined
> > bounding box of both characters to position
> > the entire label. From a user perspective, I
> > would have expected that only the non-lifted
> > character determines the position.)
> >
> >
> >
> > -------------------------------------------------------------------------
> > This SF.Net email is sponsored by the Moblin Your Move Developer's
> > challenge Build the coolest Linux based applications with Moblin SDK &
> > win great prizes Grand prize is a trip for two to an Open Source event
> > anywhere in the world
> > http://moblin-contest.org/redirect.php?banner_id=100&url=/
> > _______________________________________________
> > gnuplot-beta mailing list
> > gnuplot-beta@...
> > https://lists.sourceforge.net/lists/listinfo/gnuplot-beta

maybe somebody has an idea why wxt terminal does not work with gnuplot
4.2 in my Dell Vostro 1510 with Ubuntu 8.04. It plots first figure and
then freezes, does not respond to anything and I have to force quit from
gnuplot. x11 and other terminals work fine. Maybe I need some libraries,
which are not installed?
thank you very much,
Vitaly.

On Thursday 24 July 2008 12:57:42 Philipp K. Janert wrote:
> 
> According to the manual, the units for the
> 	set ?margin
> commands are "character heights or widths,
> as appropriate", using the default font for the
> current terminal, I assume.
> 
> What does this mean for non-fixed character
> sized fonts, including Helvetica?

Internally, it means the values for width and height stored in the
current terminal's term->h_char and term->v_char.
Each terminal driver is supposed to make a best-effort attempt to
store reasonable numbers for these values, but the accuracy varies.

In the specific case of PostScript, the vertical size is set to the normal
font size in points.  (E.g. for "Helvetica,12" the height is 12 points).
The horizontal size is set to 0.6 of the vertical size.

The libgd-based terminals do not assume that the nominal font size
is correct.  Instead they query the actual dimensions of a test string
and set term->h_char and term->v_char on the empirically determined
values.

The accuracy of the values can be judges by inspecting the output
of the "test" command.  If the values are accurate, then the box around
the central test string will fit nicely.  If the values are garbage, the box
will be too big or too small or improperly proportioned.

-- 
Ethan A Merritt

Interesting. Thanks.

On Thursday 24 July 2008 13:25, you wrote:
> On Thursday 24 July 2008 12:57:42 Philipp K. Janert wrote:
> > According to the manual, the units for the
> > 	set ?margin
> > commands are "character heights or widths,
> > as appropriate", using the default font for the
> > current terminal, I assume.
> >
> > What does this mean for non-fixed character
> > sized fonts, including Helvetica?
>
> Internally, it means the values for width and height stored in the
> current terminal's term->h_char and term->v_char.
> Each terminal driver is supposed to make a best-effort attempt to
> store reasonable numbers for these values, but the accuracy varies.
>
> In the specific case of PostScript, the vertical size is set to the normal
> font size in points.  (E.g. for "Helvetica,12" the height is 12 points).
> The horizontal size is set to 0.6 of the vertical size.
>
> The libgd-based terminals do not assume that the nominal font size
> is correct.  Instead they query the actual dimensions of a test string
> and set term->h_char and term->v_char on the empirically determined
> values.
>
> The accuracy of the values can be judges by inspecting the output
> of the "test" command.  If the values are accurate, then the box around
> the central test string will fit nicely.  If the values are garbage, the
> box will be too big or too small or improperly proportioned.

On Monday 21 July 2008 19:43:58 Shigeharu TAKENO wrote:
> shige 07/22 2008
> ----------------
> 
> I will add some explanations for my last report.

Got it.  Thanks.  
I did not understand at first why you proposed to change the word
order in   ... see `foo` ...  for details.  But now I understand.

	Ethan

> I wrote:
> > In docs/gnuplot.doc of current CVS version
> > 
> >  ROS $Id: gnuplot.doc,v 1.518 2008/07/07 16:37:16 sfeam Exp $
> > 
> > I found some points 
> > 
> > 1) that seem to be misprints, 
> 
> They are:
> 
>  line 1 ("ROS" should be "RCS" ?)
>  line 8103 ("nex" should be "next" ?)
>  line 9371 (""set view map"" should be "`set view map`" ?)
>  line 11524 ("`Seeking-assistance`" should be "`seeking-assistance`" ?)
> 
> 
> > 2) for "see `...`" system of doc2tex, or
> 
> doc2tex add "\pageref{}" and "\index{}" for the `...` strings 
> between "[sS]ee" and "[).]". So, for example, the following line 
> (line 1575) include two `...` part between "see" and ".":
> 
>   Please see `plot datafile using` for an explanation of the 
>   `using` syntax.
> 
> so doc2tex expand it to
> 
>   Please see {\bf plot datafile using (p.~\pageref{plot datafile using})
>   \index{using}} for an explanation of the 
>   {\bf using (p.~\pageref{using})\index{using}} syntax.
> 
> If we modify the line to
> 
>   For an explanation of the `using` syntax, please see 
>   `plot datafile using`. 
> 
> doc2tex will expand it to
>   
>   For an explanation of the {\bf using} syntax, 
>   please see {\bf plot datafile using (p.~\pageref{plot datafile using})
>   \index{using}}.
> 
> I think the latter result is better.
> 
> Such lines are:
> 
>  line 1575
>  line 9239-9240
>  line 11104
> 
> 
> > 3) for "  " (double space) at the top of line of doc2tex.
> 
> doc2tex enclose lines starting double spaces by 
> 
>  \preverbatim
>  \begin{verbatim}
> 
> and
> 
>  \end{verbatim}
>  \postverbatim
> 
> But there are some lines which are starting double spaces and 
> they don't seem to be surrounded by *{verbatim}. They are:
> 
>  line 2500
>  line 5506-5510
>  line 8925
>  line 9386
>  line 9493
> 
> 
> > I send the unified diff file for them.
> > 
> > ----- From here -----
> > --- gnuplot.doc.ORG     2008-07-11 20:51:20.000000000 +0900
> > +++ gnuplot.doc 2008-07-11 20:58:14.000000000 +0900
> > @@ -1,4 +1,4 @@
> > -C ROS $Id: gnuplot.doc,v 1.518 2008/07/07 16:37:16 sfeam Exp $
> > +C RCS $Id: gnuplot.doc,v 1.518 2008/07/07 16:37:16 sfeam Exp $
> >  C
> >  C Copyright (C) 1986 - 1993, 1998, 1999, 2000, 2001, 2004   Thomas Williams, Colin Kelley et al.
> >  C
> > @@ -1572,7 +1572,7 @@
> >  
> >         plot 'file' using 1:( $4<0 ? 1/0 : ($2+$3)/2 )
> >  
> > - Please see `plot datafile using` for an explanation of the `using` syntax.
> > + For an explanation of the `using` syntax, please see `plot datafile using`.
> >  
> >  3 Gnuplot-defined variables
> >  ?gnuplot-defined variables
> > @@ -2497,7 +2497,7 @@
> >     # Plot with crossbars on the whiskers, crossbars are 50% of full width
> >     plot 'stat.dat' using 1:3:2:6:5 with candlesticks whiskerbars 0.5
> >  
> > -  See `set boxwidth`, `set bars` and `set style fill`.
> > + See `set boxwidth`, `set bars` and `set style fill`.
> >  2 circles
> >  ?commands set style circles
> >  ?set style circles
> > @@ -5503,11 +5503,11 @@
> >   The optional parameter is ignored for single plot-windows terminal, i.e. `pm`
> >   and `win`.
> >  
> > -  If the window is not raised under X11, then (1) they don't run in the same
> > -  X11 session (telnet or ssh session, for example), or (2) raising is blocked
> > -  by your window manager. On KDE, you may like to go to the KDE Control Center
> > -  => Desktop => Window Behaviour => Advanced and set the "Focus stealing
> > -  prevention level" to None (default is Low).
> > + If the window is not raised under X11, then (1) they don't run in the same
> > + X11 session (telnet or ssh session, for example), or (2) raising is blocked
> > + by your window manager. On KDE, you may like to go to the KDE Control Center
> > + => Desktop => Window Behaviour => Advanced and set the "Focus stealing
> > + prevention level" to None (default is Low).
> >  2 refresh
> >  ?commands refresh
> >  ?refresh
> > @@ -8100,7 +8100,7 @@
> >   the mouse position. Labels are drawn with the current setting of `mouseformat`.
> >   The `labeloptions` string is passed to the `set label` command.  The default is
> >   "point pointstyle 1" which will plot a small plus at the label position.
> > - Temporary labels will disappear at the nex `replot` or mouse zoom operation.
> > + Temporary labels will disappear at the next `replot` or mouse zoom operation.
> >   Persistent labels can be removed by holding the Ctrl-Key down while clicking
> >   Button 2 on the label's point. The threshold for how close you must be to the
> >   label is also determined by the `pointsize`.
> > @@ -8922,7 +8922,7 @@
> >  ^ <a href="http://zimg.sourceforge.net">
> >             http://zimg.sourceforge.net
> >  ^ </a>
> > -  ).
> > + ).
> >  
> >   `set palette gray` switches to a gray only palette. `set palette rgbformulae`,
> >   `set palette defined`, `set palette file` and `set palette functions` switch
> > @@ -9236,8 +9236,8 @@
> >   r = f(t), where t is an angle of rotation.  The trange controls the domain
> >   (the angle) of the function, and the x and y ranges control the range of the
> >   graph in the x and y directions.  Each of these ranges, as well as the
> > - rrange, may be autoscaled or set explicitly.  See `set xrange` for details
> > - of all the `ranges` commands.
> > + rrange, may be autoscaled or set explicitly. For details of all the
> > + `ranges` commands, see `set xrange`.
> >  
> >   Example:
> >         set polar
> > @@ -9368,7 +9368,7 @@
> >   values (1.0).
> >  
> >   `ratio` and `square` have no effect on 3D plots, but do affect 3D projections
> > - created using "set view map".
> > + created using `set view map`.
> >  
> >   Examples:
> >  
> > @@ -9383,7 +9383,7 @@
> >  
> >   See also
> >  ^ <a href="http://www.gnuplot.info/demo/airfoil.html">
> > -  airfoil demo.
> > + airfoil demo.
> >  ^ </a>
> >  3 style
> >  ?set style
> > @@ -9490,7 +9490,7 @@
> >         set style arrow 1 nohead lw 2
> >         set arrow arrowstyle 1
> >  
> > -  See also `set arrow` for further examples.
> > + See also `set arrow` for further examples.
> >  
> >  4 set style data
> >  ?commands set style data
> > @@ -11101,7 +11101,7 @@
> >   `splot` provides only a single x, y, and z axis; there is no equivalent to the
> >   x2 and y2 secondary axes provided by `plot`.
> >  
> > - See `plot` for features common to the `plot` command; only differences are
> > + For features common to the `plot` command, see `plot`; only differences are
> >   discussed in detail here.
> >  
> >   Syntax:
> > @@ -11521,7 +11521,7 @@
> >   Or upload the report to the gnuplot web site on SourceForge.
> >   Please give complete information on the version of gnuplot you are using
> >   and, if possible, a test script that demonstrates the bug.
> > - See `Seeking-assistance`.
> > + See `seeking-assistance`.
> >  
> >   The sections below list problems known to be present in gnuplot version 4.2 at
> >   the time of release.  Some of these are actually bugs in external support
> > ----- To here ----- 
> 
> +========================================================+
>  Shigeharu TAKENO     NIigata Institute of Technology
>                        kashiwazaki,Niigata 945-1195 JAPAN
>  shige@...   TEL(&FAX): +81-257-22-8161
> +========================================================+
> 
> -------------------------------------------------------------------------
> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
> Build the coolest Linux based applications with Moblin SDK & win great prizes
> Grand prize is a trip for two to an Open Source event anywhere in the world
> http://moblin-contest.org/redirect.php?banner_id=100&url=/
> _______________________________________________
> gnuplot-beta mailing list
> gnuplot-beta@...
> https://lists.sourceforge.net/lists/listinfo/gnuplot-beta
> 



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

On Thursday 17 July 2008 14:31, Manuel Schölling wrote:
> Hi guys,
> 
> I often use commands like
> 
> plot "afile" w l <and many more options>, \
> "anotherfile" w l <the same options>, \
> ...
> 
> So I implemented an auto completion feature for this, so you can use it 
> like this:
> 
> plot "" <all your options>
>        ^ place cursor here and press <tab>
> result:
> plot "file1" <all your options>,
>   "file2" <all your options>,
>   "file3" <all your options>


How is this different from the auto-completion that libreadline does already?

In other words, is your patch adding the auto-completion capability to
gnuplot's built-in readline, or is it something added on top of libreadline?



 
> There is an minimal example application attached to this email but in 
> this preprocessor command mix I cannot figure out where to place the 
> rl_bind_key command in the gnuplot source code.
> 
> It would be great if you could tell me where to place this command so I 
> can create a patch for this feature.
> 
> Please note that its feature should not remove the current <tab> feature 
> and will via Alt+C or something.
> 
> Cheers,
> Manuel
> 
> 

-- 
Ethan A Merritt            Courier Deliveries: 1959 NE Pacific
Dept of Biochemistry       Regular Mail: Mailstop 357742
Health Sciences Building
University of Washington - Seattle WA 98195-7742

On Thursday 17 July 2008 14:31:20 Manuel Schölling wrote:
> So I implemented an auto completion feature for this, so you can use it
> like this:
>
> plot "" <all your options>
>        ^ place cursor here and press <tab>
> result:
> plot "file1" <all your options>,
>   "file2" <all your options>,
>   "file3" <all your options>

I have no answer to your question, but your email reminded me of a feature 
that I think would be very nice to have, the support of wildcards in 
filenames. For example, the gnuplot command

  plot "file[1-3]" <all your options>

would expand to what you suggest. This approach would also work in scripts, 
whereas your implementation seems to work only in interactive gnuplot 
sessions.

  Lutz

On Friday 18 July 2008 15:23:34 Lutz Maibaum wrote:
> On Thursday 17 July 2008 14:31:20 Manuel Schölling wrote:
> > So I implemented an auto completion feature for this, so you can use it
> > like this:
> >
> > plot "" <all your options>
> >        ^ place cursor here and press <tab>
> > result:
> > plot "file1" <all your options>,
> >   "file2" <all your options>,
> >   "file3" <all your options>
> 
> I have no answer to your question, but your email reminded me of a feature 
> that I think would be very nice to have, the support of wildcards in 
> filenames. For example, the gnuplot command
> 
>   plot "file[1-3]" <all your options>
> 
> would expand to what you suggest. This approach would also work in scripts, 
> whereas your implementation seems to work only in interactive gnuplot 
> sessions.

We already have that:

   plot for [i=1:3] "file".i <all your options>




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

On Friday 18 July 2008 15:46:27 Ethan Merritt wrote:
> On Friday 18 July 2008 15:23:34 Lutz Maibaum wrote:
> > I have no answer to your question, but your email reminded me of a
> > feature that I think would be very nice to have, the support of
> > wildcards in filenames.

> >   plot "file[1-3]" <all your options>

> We already have that:
>
>    plot for [i=1:3] "file".i <all your options>

Thanks, Ethan, I didn't know about plot iterations. It seems like they are 
limited to integers, are there any plans to allow iterations over string 
lists? It would be great if one could do something like

  plot for (filename in test*.dat) i <all your options>

or at least

  plot for (filename in `ls test*.dat`) i <all your options>

Right now it seems like I would have to do something like

  list="`ls test*.dat`"
  plot for [i=1:words(list)] word(list,i)

which actually does not work, since calling ls through backticks generates 
a list with only a single item in it:

gnuplot> !ls test*.dat
test1.dat  test2.dat
gnuplot> list="`ls test*.dat`"
gnuplot> print word(list,1)
test1.dattest2.dat

  -- Lutz

On Friday 18 July 2008, Lutz Maibaum wrote:
> On Friday 18 July 2008 15:46:27 Ethan Merritt wrote:
> > On Friday 18 July 2008 15:23:34 Lutz Maibaum wrote:
> > > I have no answer to your question, but your email reminded me of a
> > > feature that I think would be very nice to have, the support of
> > > wildcards in filenames.
> 
> > >   plot "file[1-3]" <all your options>
> 
> > We already have that:
> >
> >    plot for [i=1:3] "file".i <all your options>
> 
> Thanks, Ethan, I didn't know about plot iterations. It seems like they are 
> limited to integers, are there any plans to allow iterations over string 
> lists? It would be great if one could do something like
> 
>   plot for (filename in test*.dat) i <all your options>
> 
> or at least
> 
>   plot for (filename in `ls test*.dat`) i <all your options>
> 
> Right now it seems like I would have to do something like
> 
>   list="`ls test*.dat`"
>   plot for [i=1:words(list)] word(list,i)
> 
> which actually does not work

It works. You just have to use a slight variant:

   list = system("ls test*.dat")
   plot for [i=1:words(list)] word(list,i) title word(list,i)

I have toyed with the idea of having an explicit string iterator that
would do this automatically, something like

   list = system("ls *.dat")
   plot for [FILE in list] FILE 

but it turns out to require more code than I would have guessed, and
it didn't seem worth it since the capability is there already.
But if someone has a clever idea for implementation perhaps there is
an easier way than I found at the time.

-- 
Ethan A Merritt

On Sat, 19 Jul 2008 03:41:02 +0200, Lutz Maibaum <lutz.maibaum@...>  
wrote:

>   list="`ls test*.dat`"
>   plot for [i=1:words(list)] word(list,i)
> which actually does not work, since calling ls through backticks  
> generates
> a list with only a single item in it:
> gnuplot> !ls test*.dat
> test1.dat  test2.dat
> gnuplot> list="`ls test*.dat`"
> gnuplot> print word(list,1)
> test1.dattest2.dat
>  -- Lutz


even though Ethan has given a way to do this with system(), doesn't this  
output indicate a bug in the way several lines are concatenated into one  
useless string.

Not sure what the right output from this should be. First string, last  
string, space separated concatenation or concat containing line break  
ASCII 13 ??

One thing seems sure the current output "test1.dattest2.dat" is not a  
valid solution and can never be useful.

However,
    plot for [i=1:words(list)] word(list,i) title word(list,i)
looks highly useful.


/regards.

Lutz Maibaum wrote:
> On Friday 18 July 2008 15:46:27 Ethan Merritt wrote:
>> On Friday 18 July 2008 15:23:34 Lutz Maibaum wrote:
>>> I have no answer to your question, but your email reminded me of a
>>> feature that I think would be very nice to have, the support of
>>> wildcards in filenames.
> 
>>>   plot "file[1-3]" <all your options>
> 
>> We already have that:
>>
>>    plot for [i=1:3] "file".i <all your options>
> 
> Thanks, Ethan, I didn't know about plot iterations. It seems like they are 
> limited to integers, are there any plans to allow iterations over string 
> lists? It would be great if one could do something like
> 
>   plot for (filename in test*.dat) i <all your options>
> 
> or at least
> 
>   plot for (filename in `ls test*.dat`) i <all your options>
> 
> Right now it seems like I would have to do something like
> 
>   list="`ls test*.dat`"
>   plot for [i=1:words(list)] word(list,i)
> 
> which actually does not work, since calling ls through backticks generates 
> a list with only a single item in it:
> 
> gnuplot> !ls test*.dat
> test1.dat  test2.dat
> gnuplot> list="`ls test*.dat`"
> gnuplot> print word(list,1)
> test1.dattest2.dat

Using list="`echo test*.dat`" works fine here. Filenames are separated 
by spaces, as expected for use with word().

THeo

On Saturday 19 July 2008, Manuel Schölling wrote:

> Ethan Merritt wrote:
>  > We already have that:
>  >
>  >   plot for [i=1:3] "file".i <all your options>
> 
> This is strange. Both commands do not work here:

It is a feature in the development version (4.3).

> But nevertheless, if you use these commands you've written you don't 
> have the possibility to add further plotting options like a different 
> title for each data file.

Sure you do.  The iteration variable can be used multiple times.
Here's a trivial example:

    plot for [i=i,N] "file".i with lines title "file".i

Or, if you want custom titles:

    name = "Apples Oranges Bananas"
    plot for [i=1,3] "file".i with lines title word(name,i) 

-- 
Ethan A Merritt

On Friday 18 July 2008 21:10:12 Ethan A Merritt wrote:
> On Friday 18 July 2008, Lutz Maibaum wrote:
> > It would be great if one could do something like
> > 
> >   plot for (filename in test*.dat) i <all your options>
> > 
> 
> I have toyed with the idea of having an explicit string iterator that
> would do this automatically, something like
> 
>    list = system("ls *.dat")
>    plot for [FILE in list] FILE 
> 
> but it turns out to require more code than I would have guessed, and
> it didn't seem worth it since the capability is there already.

I had another look over the weekend.  Either I missed the obvious before,
or we have better support routines in place now.  In any event, the extra
code to implement this was quite small.  So I have added to CVS support
for the variant:

   plot for [ <variablename> in "string of words" ] ....

-- 
Ethan A Merritt

Lutz Mainbaum wrote:
> I have no answer to your question, but your email reminded me of a feature 
> that I think would be very nice to have, the support of wildcards in 
> filenames. For example, the gnuplot command
> 
>   plot "file[1-3]" <all your options>
> 
> would expand to what you suggest. This approach would also work in scripts, 
> whereas your implementation seems to work only in interactive gnuplot 
> sessions.

Ethan Merritt wrote:
 > We already have that:
 >
 >   plot for [i=1:3] "file".i <all your options>

This is strange. Both commands do not work here:

I have the data files /tmp/file[1-3] and this is the output of gnuplot:
         G N U P L O T
         Version 4.2 patchlevel 3
         last modified Mar 2008
         System: Linux 2.6.25-tuxonice-r1

         Copyright (C) 1986 - 1993, 1998, 2004, 2007, 2008
         Thomas Williams, Colin Kelley and many others

         Type `help` to access the on-line reference manual.
         The gnuplot FAQ is available from http://www.gnuplot.info/faq/

         Send bug reports and suggestions to 
<http://sourceforge.net/projects/gnuplot>


Terminal type set to 'x11'
gnuplot> plot for [i=1:3] "/tmp/file".i w l
                     ^
          ':' expected

gnuplot> plot "/tmp/file[1-3]" w l
          warning: Skipping unreadable file "/tmp/file[1-3]"
          No data in plot


I'm using Gentoo Linux with these USE flags
[I] sci-visualization/gnuplot
      Available versions:  4.2.2-r1 4.2.3-r2 {X doc emacs gd ggi latex 
pdf plotutils readline svga wxwindows xemacs}
      Installed versions:  4.2.3-r2(22:12:41 15.06.2008)(X emacs gd ggi 
latex pdf plotutils readline -doc -svga -wxwindows -xemacs)
      Homepage:            http://www.gnuplot.info/
      Description:         Command-line driven interactive plotting program


If you have a clue what's wrong here, please send me a mail.

But nevertheless, if you use these commands you've written you don't 
have the possibility to add further plotting options like a different 
title for each data file.


Cheers,
Manuel