gnuplot-info Mailing List for gnuplot (Page 4)

Hi!


Let's say I have the following code:

    [...]
    unset grid
    set grid noxtics ytics lc rgb '#c6dbef'
    unset border
    set border 1 lw 2 lc rgb '#000000'
    unset tics
    set ytics axis nomirror offset -1 0.02, 0.02, 0.1 textcolor rgb '#ff0000'
    [...]

...which results in the following fragment of the y-axis: 
https://qu.ax/bsffh.png (or see attached).

Now, the manual for `border` says that:

> Besides the border itself, this line style is used for the tics,
> independent of whether they are plotted on the border or on the axes
> (see set xtics).

...which is probably why I have those little black tics leftovers :)

And therefore my question: can I somehow eliminate those 'dashes'
completely? Would be grateful for a hint :)


Thank you,



-- 
skejg

> 2) The gnuplot development branch supports a new graphical entity called
> a "mark".  Each mark is essentially a graphics subroutine that draws a
> shape consisting of line segments and filled polygons.

Helo Ethan,

  thank you very much for the suggestion. I tried it. It's almost 
perfect, but it does not work correctly with tikz terminal when plotsize 
is set (which I'm using since I'm  plotting a number of plots in cycle 
and I want all graphs to remain the same size, regardless of the legend 
size.).

So, all marks are shrunken horizontally, for example instead of circles 
gnuplot plots agg-like points, please see the attached example. Is there 
any way to fix it?

Thanks!

On Friday, 14 February 2025 11:34:47 PST Dmitry wrote:
> Hello Norwid,
> thank you for your reply. The script I attached is just an example to 
> demonstrate the problem.
> My actual plot is much more complicated, I have attached a .pdf example 
> of what I'm actually plotting.
> Basically I have two datasets: one should be plotted with empty gray 
> points, and the second dataset should be  plotted with filled colored 
> points.
> Since I have a lot of data to plot, gnuplot just does not have enough 
> pointtypes of desired properties, so I decided to try unicode.
> So far I ended up using
> > plot [0:50] "battery.dat" with xyerrorbars pt -1 notitle ,\
> > "battery.dat" using 1:2:("▢") with labels offset 0, character 0.035 
> > notitle
> , where I adjust offset manually for each pointtype (unicode char) by 
> try and error method. That's how I plotted the attached pdf.
> But it's far from ideal, since if decide to change the size of the plot, 
> all offsets are required to be readjusted.
> I think that it would be great if gnuplot could offer more pointtypes, 
> since adding unicode points is a lot of pain because each character has 
> it's own shift.
> 
> Or may be there are and easy way that I don't know about? That would be 
> great.

I can offer two thoughts that may or may not be realistic options for you.

1) gnuplot's postscript terminal provides some huge number of point types,
mostly partially filled squares and circles.
That might be an option if you can deal with importing the figure as *.eps
and if the point types are distinct enough for your purpose.

2) The gnuplot development branch supports a new graphical entity called
a "mark".  Each mark is essentially a graphics subroutine that draws a
shape consisting of line segments and filled polygons.

Once defined, a mark can be used in a "plot with marks" command directly
analogous to "plot with points" except that there is more flexibility with
respect to options for scaling.   Defining new point types is an obvious
example of use, shown here:

  https://gnuplot.sourceforge.net/demo_6.1/extra_points.html
  https://gnuplot.sourceforge.net/demo_6.1/mark_formulas.html

This is an experimental extension added about a year ago.
Until very recently it has been sadly lacking in documentation.
Current documentation, such as it is, here

  http://gnuplot.info/docs_6.1/loc3934.html

It is IMHO rather cumbersome to use in its current state.  I suspect that
as people begin to use it and offer feedback it will evolve into something
easier to use.

Ethan

> Thanks!
> 
> 
> On 14.02.2025 14:16, Norwid Behrnd via gnuplot-info wrote:
> > @Dmitry Preface: I'm not yet convinced a character pointtype by `"\U+25A2"`
> > improves the default tools available if one wants to indicate the points'
> > location by a box.  The difference is gnuplot's box does not use rounded
> > corners -- is this a requirement on your side?  Because depending on font and
> > specific glyph, what is the origin of coordinates of its bounding box (in
> > LaTeX's parlance) gnuplot uses a reference to put the glyph on the canvas --
> > is it in one of the bounding box' corners, or its center (then affected by the
> > glyph's height and width)?[1]  Would using the "rounded square" render the plot
> > much easier to read, than gnuplot's box?  Perhaps the script shared by you used
> > `"\U+25A2"` as a place holder for a different printable object.
> >
> > My suggest is to use `pt 4` within the pdfcairo terminal to write an
> > intermediate .pdf subsequently converted e.g., by David Barton's pdf2svg[2]
> > in lines of `pdf2svg test.pdf out.svg`.  As an illustration, I edited your
> > script's header and last line to
> >
> > ``` test.spt
> > set terminal pdfcairo enhanced font 'DejaVuSans,12'
> > set output "test.pdf"
> >
> > set style data lines
> > set title "error represented by xyerrorbars"
> > set xlabel "Resistance [Ohm]"
> > set ylabel "Power [W]"
> > n(x)=1.53**2*x/(5.67+x)**2
> > NO_ANIMATION = 1
> > unset pointintervalbox
> > #plot [0:50] "battery.dat" t "Power" with xyerrorbars pt "\U+25A1" pointsize
> > 44 lw 2 lc "red", n(x) t "Theory" w lines #plot "battery.dat" t "Power" with
> > points pt "\U+25A1" pointsize 44 lw 2 lc "red"
> >
> > plot [0:50] "battery.dat" with xyerrorbars pt -1 notitle ,\
> > "battery.dat" t "Power" with points pt 4 pointsize 1 lw 2 lc "red"
> > ```
> >
> > which works reasonably well (gnuplot 6.0.2).
> >
> > [1] see for instance `x` and `V` in the accepted answer by `theozh` (May 26,
> > 2019) https://stackoverflow.com/questions/56313642/change-the-color-of-a-character-pointtype-in-gnuplot
> > [2] http://cityinthesky.co.uk/opensource/pdf2svg/
> >
> >
> > _______________________________________________
> > gnuplot-info mailing list
> > gnu...@li...
> > Membership management via: https://lists.sourceforge.net/lists/listinfo/gnuplot-info
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> Membership management via: https://lists.sourceforge.net/lists/listinfo/gnuplot-info
> 

Hello Norwid,
thank you for your reply. The script I attached is just an example to 
demonstrate the problem.
My actual plot is much more complicated, I have attached a .pdf example 
of what I'm actually plotting.
Basically I have two datasets: one should be plotted with empty gray 
points, and the second dataset should be  plotted with filled colored 
points.
Since I have a lot of data to plot, gnuplot just does not have enough 
pointtypes of desired properties, so I decided to try unicode.
So far I ended up using
> plot [0:50] "battery.dat" with xyerrorbars pt -1 notitle ,\
> "battery.dat" using 1:2:("▢") with labels offset 0, character 0.035 
> notitle
, where I adjust offset manually for each pointtype (unicode char) by 
try and error method. That's how I plotted the attached pdf.
But it's far from ideal, since if decide to change the size of the plot, 
all offsets are required to be readjusted.
I think that it would be great if gnuplot could offer more pointtypes, 
since adding unicode points is a lot of pain because each character has 
it's own shift.

Or may be there are and easy way that I don't know about? That would be 
great.
Thanks!


On 14.02.2025 14:16, Norwid Behrnd via gnuplot-info wrote:
> @Dmitry Preface: I'm not yet convinced a character pointtype by `"\U+25A2"`
> improves the default tools available if one wants to indicate the points'
> location by a box.  The difference is gnuplot's box does not use rounded
> corners -- is this a requirement on your side?  Because depending on font and
> specific glyph, what is the origin of coordinates of its bounding box (in
> LaTeX's parlance) gnuplot uses a reference to put the glyph on the canvas --
> is it in one of the bounding box' corners, or its center (then affected by the
> glyph's height and width)?[1]  Would using the "rounded square" render the plot
> much easier to read, than gnuplot's box?  Perhaps the script shared by you used
> `"\U+25A2"` as a place holder for a different printable object.
>
> My suggest is to use `pt 4` within the pdfcairo terminal to write an
> intermediate .pdf subsequently converted e.g., by David Barton's pdf2svg[2]
> in lines of `pdf2svg test.pdf out.svg`.  As an illustration, I edited your
> script's header and last line to
>
> ``` test.spt
> set terminal pdfcairo enhanced font 'DejaVuSans,12'
> set output "test.pdf"
>
> set style data lines
> set title "error represented by xyerrorbars"
> set xlabel "Resistance [Ohm]"
> set ylabel "Power [W]"
> n(x)=1.53**2*x/(5.67+x)**2
> NO_ANIMATION = 1
> unset pointintervalbox
> #plot [0:50] "battery.dat" t "Power" with xyerrorbars pt "\U+25A1" pointsize
> 44 lw 2 lc "red", n(x) t "Theory" w lines #plot "battery.dat" t "Power" with
> points pt "\U+25A1" pointsize 44 lw 2 lc "red"
>
> plot [0:50] "battery.dat" with xyerrorbars pt -1 notitle ,\
> "battery.dat" t "Power" with points pt 4 pointsize 1 lw 2 lc "red"
> ```
>
> which works reasonably well (gnuplot 6.0.2).
>
> [1] see for instance `x` and `V` in the accepted answer by `theozh` (May 26,
> 2019) https://stackoverflow.com/questions/56313642/change-the-color-of-a-character-pointtype-in-gnuplot
> [2] http://cityinthesky.co.uk/opensource/pdf2svg/
>
>
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> Membership management via: https://lists.sourceforge.net/lists/listinfo/gnuplot-info

@Dmitry Preface: I'm not yet convinced a character pointtype by `"\U+25A2"`
improves the default tools available if one wants to indicate the points'
location by a box.  The difference is gnuplot's box does not use rounded
corners -- is this a requirement on your side?  Because depending on font and
specific glyph, what is the origin of coordinates of its bounding box (in
LaTeX's parlance) gnuplot uses a reference to put the glyph on the canvas --
is it in one of the bounding box' corners, or its center (then affected by the
glyph's height and width)?[1]  Would using the "rounded square" render the plot
much easier to read, than gnuplot's box?  Perhaps the script shared by you used
`"\U+25A2"` as a place holder for a different printable object.

My suggest is to use `pt 4` within the pdfcairo terminal to write an
intermediate .pdf subsequently converted e.g., by David Barton's pdf2svg[2]
in lines of `pdf2svg test.pdf out.svg`.  As an illustration, I edited your
script's header and last line to

``` test.spt
set terminal pdfcairo enhanced font 'DejaVuSans,12'
set output "test.pdf"

set style data lines
set title "error represented by xyerrorbars"
set xlabel "Resistance [Ohm]"
set ylabel "Power [W]"
n(x)=1.53**2*x/(5.67+x)**2
NO_ANIMATION = 1
unset pointintervalbox
#plot [0:50] "battery.dat" t "Power" with xyerrorbars pt "\U+25A1" pointsize
44 lw 2 lc "red", n(x) t "Theory" w lines #plot "battery.dat" t "Power" with
points pt "\U+25A1" pointsize 44 lw 2 lc "red"

plot [0:50] "battery.dat" with xyerrorbars pt -1 notitle ,\
"battery.dat" t "Power" with points pt 4 pointsize 1 lw 2 lc "red"
```

which works reasonably well (gnuplot 6.0.2).

[1] see for instance `x` and `V` in the accepted answer by `theozh` (May 26,
2019) https://stackoverflow.com/questions/56313642/change-the-color-of-a-character-pointtype-in-gnuplot
[2] http://cityinthesky.co.uk/opensource/pdf2svg/

Hello,
  I'm plotting errorbars with points, and I need more pointtypes then 
are presented in gnuplot.
I tried adding unicode characters like "with points pt "\U+25A2", it 
works,  but points are shifted in the "y" direction (relative to 
errorbars centers).
I have attached the script and the resulting picture. On the attached 
picture points are shifted down. I'm actually using tikz terminal (I 
need latex support), and with tikz terminal points are shifted upwards.
Is there any way to fix this vertical shift?
Thank you!

P.S. I could shift points manually by changing their "y" coordinates, 
but I'm plotting a lot of graphs with different scales in cycle (using c++),
so manually shifting "y" coordinates is out of question (unless there is 
some option to shift points in absolute units (like milimeters)).

`.set y2label`  was my shorthand for .set("y2label") which is
javascript.  I know nothing about any other syntax, I only use it with
js.

I am not surprised that 6.0.2 works.  I went back to 4.6 and
everything works.   The bug is apparently only in 5.x.


On Thu, Feb 13, 2025 at 5:07 AM Norwid Behrnd via gnuplot-info
<gnu...@li...> wrote:
>
> @Mark While I'm not aware `.set y2label` to be a syntax accepted by gnuplot, I
> found the test script below -- without the leading period -- worked fine from
> the Bash by either `gnuplot ./test.spt`, or `gnuplot-qt ./test.spt`.  The test
> run gnuplot 6.0.2 as currently provided by Linux Debian 13/trixie which for
> now (still) is branch testing.
>
> If an interactive GUI isn't a requirement for you, could the installation of a
> second (i.e., additional) Linux distribution equally be an option for you?  Or
> one which lives in a USB thumbdrive, thanks to e.g.,
> [rufus](https://rufus.ie/)?
>
> According to Debian's tracker,[1] Ubuntu 24.10 and future 25.04 currently
> package version 6.0.0.  However, repology.org[2] mentions for instance Devuan
> (unstable), a Debian relative, equally set to gnuplot 6.0.2.
>
> Norwid
>
> ``` test.spt
> set terminal pngcairo enhanced font 'Verdana,10'
> set output "test.png"
> set title "title"
> set grid
> set key left
>
> set xlabel "primary abscissa [x]"
>
> set x2tics
> set link x2 via x/3.14 inverse x*3.14
> set x2label "secondary abscissa [{/Symbol p}]"
> set format x2 "%.1f"
>
> set ylabel "ordinate [y]"
> set format y "%.2f"
>
> set ytics nomirror
> set y2tics
> set y2label "ordinate [y/8]"
> set link y2 via y/8 inverse y*8
> set format y2 "%.3f"
>
> plot sin(x)
> ```
>
> [1] https://tracker.debian.org/pkg/gnuplot
> [2] https://repology.org/project/gnuplot/packages
>
>
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> Membership management via: https://lists.sourceforge.net/lists/listinfo/gnuplot-info

@Mark While I'm not aware `.set y2label` to be a syntax accepted by gnuplot, I
found the test script below -- without the leading period -- worked fine from
the Bash by either `gnuplot ./test.spt`, or `gnuplot-qt ./test.spt`.  The test
run gnuplot 6.0.2 as currently provided by Linux Debian 13/trixie which for
now (still) is branch testing.

If an interactive GUI isn't a requirement for you, could the installation of a
second (i.e., additional) Linux distribution equally be an option for you?  Or
one which lives in a USB thumbdrive, thanks to e.g.,
[rufus](https://rufus.ie/)?

According to Debian's tracker,[1] Ubuntu 24.10 and future 25.04 currently
package version 6.0.0.  However, repology.org[2] mentions for instance Devuan
(unstable), a Debian relative, equally set to gnuplot 6.0.2.

Norwid

``` test.spt
set terminal pngcairo enhanced font 'Verdana,10'
set output "test.png"
set title "title"
set grid
set key left

set xlabel "primary abscissa [x]"

set x2tics
set link x2 via x/3.14 inverse x*3.14 
set x2label "secondary abscissa [{/Symbol p}]"
set format x2 "%.1f"

set ylabel "ordinate [y]"
set format y "%.2f"

set ytics nomirror
set y2tics
set y2label "ordinate [y/8]"
set link y2 via y/8 inverse y*8
set format y2 "%.3f"

plot sin(x)
```

[1] https://tracker.debian.org/pkg/gnuplot
[2] https://repology.org/project/gnuplot/packages

On 12/02/2025 04:19, Dave Horsfall wrote:
> On Tue, 11 Feb 2025, Mark Hahn wrote:
>
>> I've been running gnuplot for years and it worked great.  I just
>> installed the newest version of stable ubuntu and gnuplot broke.  I
>> don't know what version my old gnuplot was but it is now 5.4 patchlevel
>> 2.  I had Y labels on both sides.  Now my right label has disappeared.
>> I haven't changed any gnuplot commands.
> On my MacBook running High Sierra:
>
>      gnuplot> show ver
>
> 	G N U P L O T
> 	Version 6.0 patchlevel 2    last modified 2024-12-19
>
> (I check for software updates every week; doesn't everybody?)

Yes. But the OP is using the latest Ubuntu. Be aware that Gnuplot 6 will 
not build (out of the box) on Ubuntu due to a packaging error with Qt. 
This is on the Ubuntu maintainers' todo list but Ubuntu users are likely 
to be stuck on Gnulplot 5.4 for a long time yet. By the time I reported 
the issue for Gnuplot, this packaging problem had already been on their 
radar for over a year because it had been reported as having broken some 
other (gaming?) app.

P.

>
> -- Dave
>
>
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> Membership management via: https://lists.sourceforge.net/lists/listinfo/gnuplot-info

On Tue, 11 Feb 2025, Mark Hahn wrote:

> I've been running gnuplot for years and it worked great.  I just 
> installed the newest version of stable ubuntu and gnuplot broke.  I 
> don't know what version my old gnuplot was but it is now 5.4 patchlevel 
> 2.  I had Y labels on both sides.  Now my right label has disappeared.  
> I haven't changed any gnuplot commands.

On my MacBook running High Sierra:

    gnuplot> show ver

	G N U P L O T
	Version 6.0 patchlevel 2    last modified 2024-12-19 

(I check for software updates every week; doesn't everybody?)

-- Dave

I've been running gnuplot for years and it worked great.  I just
installed the newest version of stable ubuntu and gnuplot broke.  I
don't know what version my old gnuplot was but it is now 5.4
patchlevel 2.  I had Y labels on both sides.  Now my right label has
disappeared.  I haven't changed any gnuplot commands.

My gnuplot only has the ".set y2label" command, and not any other
label command.  I don't know how it worked but labels appeared on both
sides.  It's weird that I only had the right side command and now that
is the broken one.  I've tried adding  ".set ylabel".

Can someone give me an idea of how to fix this?  I'd be happy to go to
an older version like 4.6 but I couldn't find a linux installer for
4.x.  Would I have to build it?  I would hate that as I always have
problems with building,  My building-fu is lacking.

Thank you everyone for help. I ended up using "unset pointintervalbox", 
it worked perfectly.
I suggest that http://gnuplot.info/demo_svg_6.0/errorbars.html should be 
updated, since
the first demo script produces the attached plot, and not the plot 
presented on this page.
> I'm unsure about `pt -1`. Why not use `pt 0`?
> The following
> ```
> "battery.dat" t "Power" with xyerrorbars pt 0 lc "red"
> ``` 

Thanks for the suggestion, I tried it, but it produces visible points 
for my real case, and I want to delete these points.

On 22.11.2024 09:56, Yury wrote:
> I'm unsure about `pt -1`. Why not use `pt 0`?
> The following
> ```
> "battery.dat" t "Power" with xyerrorbars pt 0 lc "red"
> ```
> seems to do what you want/expect.
> Gnuplot is
> `Version 6.0 patchlevel 0    last modified 2023-12-09`
>
> -Yury
>
>
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> Membership management via: 
> https://lists.sourceforge.net/lists/listinfo/gnuplot-info

I'm unsure about `pt -1`. Why not use `pt 0`?
The following
```
"battery.dat" t "Power" with xyerrorbars pt 0 lc 
"red"
```
seems to do what you want/expect.
Gnuplot is
`Version 6.0 patchlevel 0    last modified 
2023-12-09`

-Yury

On Thursday, 21 November 2024 03:15:28 PST Dmitry wrote:
> Hi, this particular version of gnuplot plots white errorbar points with 
> the official demo script.
> 
> Event If I set "pointtype -1", I still get the same white points, please 
> see the attached demo script.
> 
> The previous version of gnuplot worked as expected, no white points.
> 
> I'm using gentoo.
> 
> Is this new feature or some kind of bug?

New feature.
Plot styles with errorbars avoid obscuring the central point by placing
a white circle behind it.  You can turn this off by saying "unset pointintervalbox"


gnuplot> help with xyerrorbars    
The `xyerrorbars` style is only relevant to 2D data plots.  `xyerrorbars` is
like `points`, except that horizontal and vertical error bars are also drawn.
At each point (x,y), lines are drawn from (x,y-ydelta) to (x,y+ydelta) and
from (x-xdelta,y) to (x+xdelta,y) or from (x,ylow) to (x,yhigh) and from
(xlow,y) to (xhigh,y), depending upon the number of data columns provided.
The appearance of the tic mark at the ends of the bar is controlled by
`set errorbars`.  The clearance between the point and the error bars is
controlled by `set pointintervalbox`. To have the error bars pass directly
through the point with no interruption, use `unset pointintervalbox`.
Either 4 or 6 input columns are required.

     4 columns:  x  y  xdelta  ydelta
     6 columns:  x  y  xlow  xhigh  ylow  yhigh

Hi, this particular version of gnuplot plots white errorbar points with 
the official demo script.

Event If I set "pointtype -1", I still get the same white points, please 
see the attached demo script.

The previous version of gnuplot worked as expected, no white points.

I'm using gentoo.

Is this new feature or some kind of bug?

Thanks!

Am 19.11.2024 um 21:34 schrieb Patrick Dupre via gnuplot-info:
> Hello,
> 
> On the attached figure, there is one point off (the lowest one).
> I can only explained this by a wrong calculation.
> ~10% of my samples show such a behavior (see the other plot).

No attachment made it through list mailing list processor.

You're probably better filing a support request or bug report at our 
SourceForge site.

Hello,

On the attached figure, there is one point off (the lowest one).
I can only explained this by a wrong calculation.
~10% of my samples show such a behavior (see the other plot).
Some ideas?

Thanks


===========================================================================
 Patrick DUPRÉ                                 | | email: pd...@gm...
===========================================================================

On 11/11/24 20:16, Patrick Alken wrote:
> Is there a way to plot colored rectangle objects using spherical 
> coordinates to define them? I made a post on stackoverflow with my 
> dataset and what I'm trying to do:
>
> https://stackoverflow.com/questions/79178991/drawing-rectangles-in-spherical-coordinates-in-gnuplot 
>
>
> I am looking for a command like:
>
> set obj 1 rect from spherical THETA0,PHI0,R to spherical THETA1,PHI1,R
>
> but it doesn't appear this is possible (yet) in gnuplot. Is there 
> anyway to do this?
>
>
Hello,

This is probably not what you are looking for, but in any case, using 
Maxima (maxima.sourceforge.net) I can generate a gnuplot input file with 
something similar to what you want, with the following commands:

d:%pi/10$
f:makelist([5,[th,5*d,6*d],[ph,i*d,i*d+d]],i,0,20)$
f:append(f,makelist([5,[th,4*d,5*d],[ph,i*d,i*d+d]],i,0,20))$
f:append(f,makelist([5,[th,3*d,4*d],[ph,i*d,i*d+d]],i,0,20))$
f:append(f,makelist([5,[th,2*d,3*d],[ph,i*d,i*d+d]],i,0,20))$
plot3d(f,[transform_xy, spherical_to_xyz],[grid,1,1],[z,-5,5],[gnuplot_script_file,"spherical_rectangles.gpt"],nolegend);

The last command will show you the location where the 
"spherical_rectangles.gpt" file was saved.

Regards,

Jaime Villate

Is there a way to plot colored rectangle objects using spherical 
coordinates to define them? I made a post on stackoverflow with my 
dataset and what I'm trying to do:

https://stackoverflow.com/questions/79178991/drawing-rectangles-in-spherical-coordinates-in-gnuplot

I am looking for a command like:

set obj 1 rect from spherical THETA0,PHI0,R to spherical THETA1,PHI1,R

but it doesn't appear this is possible (yet) in gnuplot. Is there anyway 
to do this?

tableA.dat:
     1
     2
     3

tableB.dat:
     2.1
     3.9
     6.2

$ paste tableA.dat tableB.dat > tableAB.dat
$ gnuplot -e "stats 'tableAB.dat'"
  (...)
  Linear Model:       y = 2.05 x - 0.03333
  Slope:              2.05 +- 0.1443
  Intercept:          -0.03333 +- 0.3118
  Correlation:        r = 0.9975
  Sum xy:             28.5

Le 16/10/24 à 16:30, Patrick Dupre via gnuplot-info a écrit :
> Hello,
>
> I am reformulating a previous question.
> I have 2 sets of data that I read as tables.
> I would like to fit a set of data against the other set of data by
> adjusting parameters, like a coefficient amplitude.
> However, I do not know how to do it.
>
> Thank for your help.
>
> ===========================================================================
>   Patrick DUPRÉ                                 | | email:pd...@gm...
> ===========================================================================
>
>
>
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> Membership management via:https://lists.sourceforge.net/lists/listinfo/gnuplot-info

As a workaround, you might use https://gnuplot.sourceforge.net/ instead (and https://gnuplot.sourceforge.net/docs_6.0/gnuplot6.html for documentation).

Gnuplot uses SourceForge. If you try
  gnuplot.sourceforge.net is an alias for projects.sourceforge.net.cdn.cloudflare.net.
  projects.sourceforge.net.cdn.cloudflare.net has address 104.18.37.111
  projects.sourceforge.net.cdn.cloudflare.net has address 172.64.150.145
  projects.sourceforge.net.cdn.cloudflare.net has IPv6 address 2606:4700:4400::6812:256f
  projects.sourceforge.net.cdn.cloudflare.net has IPv6 address 2606:4700:4400::ac40:9691
... those IPv6 addresses might look familiar.

What CloudFlare describes from your link:
> A non-Cloudflare domain cannot CNAME to a Cloudflare domain unless the non-Cloudflare domain is added to a Cloudflare account.
... is (approximately) what's happening, although without exactly using a CNAME. It was working at some point, so I guess CF has now changed it so it no longer does. We don't have the ability to add gnuplot.info to SourceForge's CF account, so we may need to get a hold of Clark again.

People (including me) using IPv4 instead of IPv6 probably won't see this error, because gnuplot.info's IPv4 address points to SourceForge directly (204.68.111.101), not Cloudflare. But if -- hypothetically -- gnuplot.info did point to the projects.sourceforge.net.cdn.cloudflare.net v4 address, it would also generate the 1001 error.

And I get the same 1001 error over IPv6 that Aleksandr sees.



Aleksandr Dinu <al...@hu...> said (on 2024/10/12):
> Hello gnuplot community,
> 
> There's indeed a problem with website availability at gnuplot.info, however it might not be strictly DNS problem, but rather some Cloudflare misconfiguration.
> ...
> <
> * Closing connectionerror code: 1001
> 
> ... https://developers.cloudflare.com/support/troubleshooting/cloudflare-errors/troubleshooting-cloudflare-1xxx-errors/#error-1001-dns-resolution-error.
> ...

Hello,

I am reformulating a previous question.
I have 2 sets of data that I read as tables.
I would like to fit a set of data against the other set of data by
adjusting parameters, like a coefficient amplitude.
However, I do not know how to do it.

Thank for your help.

===========================================================================
 Patrick DUPRÉ                                 | | email: pd...@gm...
===========================================================================

Works fine here.


On 10/15/24 13:00, Aleksandr Dinu via gnuplot-info wrote:
> Hello gnuplot community,
>
> There's indeed a problem with website availability at gnuplot.info, however it might not be strictly DNS problem, but rather some Cloudflare misconfiguration.
>
> The domain name resolves normally and points to some IP:
>
>> dig +short gnuplot.info.
> 204.68.111.101
>
> and in case of IPv6 specific AAAA records lookup:
>
>> dig +short AAAA gnuplot.info.
> 2606:4700:4400::ac40:96912606:4700:4400::6812:256f
>
> However when I try to reach it via HTTP request I get the following response:
>
>> curl --verbosehttp://gnuplot.info
> * Host gnuplot.info:80 was resolved.
> * IPv6: 2606:4700:4400::6812:256f, 2606:4700:4400::ac40:9691
> * IPv4: 204.68.111.101
> * Trying [2606:4700:4400::6812:256f]:80...
> * Connected to gnuplot.info (2606:4700:4400::6812:256f) port 80
>> GET / HTTP/1.1
>> Host: gnuplot.info
>> User-Agent: curl/8.7.1
>> Accept: */*
>>
> * Request completely sent off
> < HTTP/1.1 409 Conflict
> < Date: Sat, 12 Oct 2024 18:46:41 GMT
> < Content-Type: text/plain; charset=UTF-8
> < Content-Length: 16
> < Connection: close
> < X-Frame-Options: SAMEORIGIN
> < Referrer-Policy: same-origin
> < Cache-Control: private, max-age=0, no-store, no-cache, must-revalidate, post-check=0, pre-check=0
> < Expires: Thu, 01 Jan 1970 00:00:01 GMT
> < Server: cloudflare
> < CF-RAY: 8d19410bcc6d1af9-AMS
> <
> * Closing connectionerror code: 1001
>
> Seems like the request goes to one of the servers that are configured behind AAAA records, the response comes from Cloudflare and contains the error code 1001. Quick googling brought me to this documentation page with "Resolution" section -https://developers.cloudflare.com/support/troubleshooting/cloudflare-errors/troubleshooting-cloudflare-1xxx-errors/#error-1001-dns-resolution-error.
>
> I suspect there's some misconfiguration in case you use Cloudflare as your load balancer/DNS zone hosting. There's indeed a workaround - you can access the same gnuplot website by just accessing the site bywww.gnuplot.info domain (which, by the way, doesn't have any AAAA records).
>
> I see from the mailing list archives that I'm not the only person facing, so I'd like to ask people who support the website to take a closer look at this issue.
>
> I think it's an important issue, because one of the links that points to the domain with 'www.' part is the one that is on the documentation page (http://www.gnuplot.info/documentation.html) and it leads to the gnuplot online documentation (http://gnuplot.info/docs_5.5/gnuplot5.html). It means that newcomers (like myself) won't be able to access the documentation and consequently would not enjoy gnuplot's features.
>
> You could potentially fix the issue by altering DNS zone file to not include AAAA records or to put there actual IPv6 address of the server that serves the site.
> Thanks in advance!
>
> --
> Best regards,
> Aleksandr Dinu
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> Membership management via:https://lists.sourceforge.net/lists/listinfo/gnuplot-info

Hello gnuplot community,

There's indeed a problem with website availability at gnuplot.info, however it might not be strictly DNS problem, but rather some Cloudflare misconfiguration.

The domain name resolves normally and points to some IP:

> dig +short gnuplot.info.
204.68.111.101

and in case of IPv6 specific AAAA records lookup:

> dig +short AAAA gnuplot.info.
2606:4700:4400::ac40:96912606:4700:4400::6812:256f

However when I try to reach it via HTTP request I get the following response:

> curl --verbose http://gnuplot.info
* Host gnuplot.info:80 was resolved.
* IPv6: 2606:4700:4400::6812:256f, 2606:4700:4400::ac40:9691
* IPv4: 204.68.111.101
* Trying [2606:4700:4400::6812:256f]:80...
* Connected to gnuplot.info (2606:4700:4400::6812:256f) port 80
> GET / HTTP/1.1
> Host: gnuplot.info
> User-Agent: curl/8.7.1
> Accept: */*
>
* Request completely sent off
< HTTP/1.1 409 Conflict
< Date: Sat, 12 Oct 2024 18:46:41 GMT
< Content-Type: text/plain; charset=UTF-8
< Content-Length: 16
< Connection: close
< X-Frame-Options: SAMEORIGIN
< Referrer-Policy: same-origin
< Cache-Control: private, max-age=0, no-store, no-cache, must-revalidate, post-check=0, pre-check=0
< Expires: Thu, 01 Jan 1970 00:00:01 GMT
< Server: cloudflare
< CF-RAY: 8d19410bcc6d1af9-AMS
<
* Closing connectionerror code: 1001

Seems like the request goes to one of the servers that are configured behind AAAA records, the response comes from Cloudflare and contains the error code 1001. Quick googling brought me to this documentation page with "Resolution" section - https://developers.cloudflare.com/support/troubleshooting/cloudflare-errors/troubleshooting-cloudflare-1xxx-errors/#error-1001-dns-resolution-error.

I suspect there's some misconfiguration in case you use Cloudflare as your load balancer/DNS zone hosting. There's indeed a workaround - you can access the same gnuplot website by just accessing the site by www.gnuplot.info domain (which, by the way, doesn't have any AAAA records).

I see from the mailing list archives that I'm not the only person facing, so I'd like to ask people who support the website to take a closer look at this issue.

I think it's an important issue, because one of the links that points to the domain with 'www.' part is the one that is on the documentation page (http://www.gnuplot.info/documentation.html) and it leads to the gnuplot online documentation (http://gnuplot.info/docs_5.5/gnuplot5.html). It means that newcomers (like myself) won't be able to access the documentation and consequently would not enjoy gnuplot's features.

You could potentially fix the issue by altering DNS zone file to not include AAAA records or to put there actual IPv6 address of the server that serves the site.
Thanks in advance!

--
Best regards,
Aleksandr Dinu

Hello gnuplot community,

There's indeed a problem with website availability at gnuplot.info, however it might not be strictly DNS problem, but rather some Cloudflare misconfiguration.

The domain name resolves normally and points to some IP:

> dig +short gnuplot.info.
204.68.111.101

and in case of IPv6 specific AAAA records lookup:

> dig +short AAAA gnuplot.info.
2606:4700:4400::ac40:96912606:4700:4400::6812:256f

However when I try to reach it via HTTP request I get the following response:

> curl --verbose http://gnuplot.info
* Host gnuplot.info:80 was resolved.
* IPv6: 2606:4700:4400::6812:256f, 2606:4700:4400::ac40:9691
* IPv4: 204.68.111.101
* Trying [2606:4700:4400::6812:256f]:80...
* Connected to gnuplot.info (2606:4700:4400::6812:256f) port 80
> GET / HTTP/1.1
> Host: gnuplot.info
> User-Agent: curl/8.7.1
> Accept: */*
>
* Request completely sent off
< HTTP/1.1 409 Conflict
< Date: Sat, 12 Oct 2024 18:46:41 GMT
< Content-Type: text/plain; charset=UTF-8
< Content-Length: 16
< Connection: close
< X-Frame-Options: SAMEORIGIN
< Referrer-Policy: same-origin
< Cache-Control: private, max-age=0, no-store, no-cache, must-revalidate, post-check=0, pre-check=0
< Expires: Thu, 01 Jan 1970 00:00:01 GMT
< Server: cloudflare
< CF-RAY: 8d19410bcc6d1af9-AMS
<
* Closing connectionerror code: 1001

Seems like the request goes to one of the servers that are configured behind AAAA records, the response comes from Cloudflare and contains the error code 1001. Quick googling brought me to this documentation page with "Resolution" section - https://developers.cloudflare.com/support/troubleshooting/cloudflare-errors/troubleshooting-cloudflare-1xxx-errors/#error-1001-dns-resolution-error.

I suspect there's some misconfiguration in case you use Cloudflare as your load balancer/DNS zone hosting. There's indeed a workaround - you can access the same gnuplot website by just accessing the site by www.gnuplot.info domain (which, by the way, doesn't have any AAAA records).

I see from the mailing list archives that I'm not the only person facing, so I'd like to ask people who support the website to take a closer look at this issue.

I think it's an important issue, because one of the links that points to the domain with 'www.' part is the one that is on the documentation page (http://www.gnuplot.info/documentation.html) and it leads to the gnuplot online documentation (http://gnuplot.info/docs_5.5/gnuplot5.html). It means that newcomers (like myself) won't be able to access the documentation and consequently would not enjoy gnuplot's features.

You could potentially fix the issue by altering DNS zone file to not include AAAA records or to put there actual IPv6 address of the server that serves the site.

Thanks in advance!

--
Best regards,
Aleksandr Dinu