I have just posted an announcement of the creation of this ticket to gnuplot-beta. plot with linesmarks I realize that this was already an issue for "plot with linespoints", but the current program design isn't flexible enough to permit specifying both a complete set of line properties and a complete set of point (or in this case mark) properties in a single plot component. Is it really so bad to require "plot <foo> with lines, '' with marks" rather than providing a separate plot style that combines...
First-pass review of gnuplotmarks patch I have only glanced at the code, so these comments are based mostly on reading the pdf description and trying it out. Any or all of these points may be worth discussion; perhaps we should start a thread on the gnuplot-beta mailing list? It's not very active these days but still I think people are more likely to see it there than will see it here on the patch tracker. I think the general goal and approach is worth doing. An unmentioned difference between "with...
new plotting style 'marks', 'linesmarks' and new object 'mark'
No warn for calling of plot (via function block) in invalid context
Update release notes and version timestamp for 6.0.1
zoom_reset_after_error()
Additional restriction on what is legal inside a function block
remove vestiges of the old latex tutorial
"make distclean" should not remove Makefile.in
mp: typo in source caused ./configure --with-metapost to fail
fix order-dependent parsing of 2D plots with "fillstyle solid fillcolor variable"
Enable the use of 'lc rgb variable' in 'splot with polygons'
Additional restriction on what is legal inside a function block
No warn for calling of plot (via function block) in invalid context
Allows input from binary data for 'splot with polygons'
Now in git. The associated documentation section is under "binary general". gnuplot> help binary blank Some plot styles require blank lines to separate groups of data read from a text input file. For instance the vertices of a one polygon in an input text data stream are separated from those of the next polygon by a blank line. Since there is no such thing as a 'blank line' in a binary file, this option allows a special record in a general binary file to be interpreted as if it were a blank line....
remove vestiges of the old latex tutorial
docs: tweak Makefile syntax that was leaving .Po files after "make distclean"
binary blank lines - demo
New binary keyword option "blank=NaN"
"make distclean" should not remove Makefile.in
Does adding "LC_ALL=C" at the start of the sed command have any impact on the conversion process?
mp: typo in source caused ./configure --with-metapost to fail
"binary_blank_if_NaN.patch" has been applied to my environment and confirmed to work. So if the first column contains NaN, the record is considered a blank row. I am satisfied with this behaviour, without the need to modify plot3d_polygons!
Thank you for working on this matter. Your solution with the keyword option blank=NaN is preferable and sufficient for me. I like the fact that this method is only valid when a break line with NaN is required and has no side effects in other cases. Another good part is that it can be applied to cases other than polygons. There are certainly plotting styles that treat NaN as valid data (hsteps), so I agree with the solution by introducing this keyword option.
I was over-thinking this. Yes it's hard for the program to know when it is or isn't appropriate to treat a NaN value as indicating a blank line. But we can leave that up to the user. This patch to datafile.c introduces a new binary file keyword optionblank=NaN. If that option appears in the plot command then a NaN value in the first field of a binary record is treated as a blank line. This only applies to general binary input, and only if the user requests it. No changes are necessary to higher level...
Sorry. Attached is my test script.
test script was not attached. I used a variant of the one for patch #813. First thoughts Rearranging the data inside the plotting routine for 3D polygons strikes me as being way too late. I think it would be better to make the binary input path more parallel to the ascii input path. If there were a condition that the binary input recognized as equivalent to a blank line, df_readbinary() could simply return DF_FIRST_BLANK and everything would "just work". That's tricky, however, because NaN is valid...
Enabling the use of 'lc rgb variable' in 'splot with polygons'
Both the patch and the fix for the bug in plot #4 are now in 6.1 and pending for the next patchlevel release of 6.0
fix order-dependent parsing of 2D plots with "fillstyle solid fillcolor variable"
Enable the use of 'lc rgb variable' in 'splot with polygons'
I agree that plot #4 shows a bug. It is also an example of how the use of fillcolor and linecolor is confusing.
Thank you for the detailed explanation. I now understand the direction regarding colour fills in Version 6. Unrelated to this ticket, but regarding fillcolor, I sometimes encounter a bit strange behaviour in 2D plots: with circles, for example, only the fourth plot in the script below does not work as expected. It seems that the colour filling behaviour may be sensitive to the positional relationship between fillcolor and fillstyle. I have been unable to determine whether this is a bug or a specification,...
OK, I get it. The documentation says "lc rgb variable" should work, but actually the only thing that works is "fc rgb variable". I will apply the patch because accepting "lc" is better than drawing everything in black. But I will also change the documentation to say explicitly that "fillcolor" is the property that should be used. I would ideally prefer that specifying the line color (as opposed to the fill color) should change the border rather than the interior of the polygon. But the current implementation...
OK, I get it. The documentation says "lc rgb variable" should work, but actually it is "fc rgb variable". I will apply the patch because anything is better than drawing everything in black. But I will also change the documentation to say explicitly that "fillcolor" is the property that should be used. I would prefer that specifying the line color (as opposed to the fill color) should change the border rather than the interior of the polygon. But the current implementation does not allow for that...
Enabling the use of 'lc rgb variable' in 'splot with polygons'
I am attaching a sample script. Sorry for the insufficient verification. * lc variable NG 5.4.8 OK 6.0.0 OK develop * fc rgb variable NG 5.4.8 OK 6.0.0 OK develop * lc rgb variable NG 5.4.8 OK 6.0.0 NG develop I didn't notice because I usually work with the develop version, but 'lc rgb variable' worked fine in 6.0.0. On the other hand, 'lc rgb variable' does not work in the develop version. Maybe there was a change somewhere that made it not work. So it might be better to investigate the above changes...
I am having trouble reproducing this. Could you provide an example script that shows a before/after difference? Starting with the polygons.dem example script, I tried splot 'dodecahedron.dat' using 1:2:3:(rgbcolor("gold")) with polygons fillcolor variable which works, and splot 'dodecahedron.dat' using 1:2:3:(rgbcolor("gold")) with polygons linecolor variable which does not work. After applying the patch I see the same behaviour: "fc rgb variable" works and "lc rgb variable" does not. Am I testing...
Allows input from binary data for 'splot with polygons'
Enabling the use of 'lt rgb variable' in 'splot with polygons'
wxt: remove --with-wx-multithreaded configuration option
svg: depth-order pm3d plot component needs its own group
hidden3d disables toggling in interactive terminals
Make sure cbrange is valid when palette(z) is called
draw colorbar tics and mirror tics separately
svg: depth-order pm3d plot component needs its own group
Correct. That is an intrinsic result of the way hidden3d and toggle are implemented. hidden3d affects the way plot elements are ordered prior to sending them as a display list to the output device. toggle is an operation performed by the output device to alter how it processes the display list. Note that gnuplot (the executable program) may be long gone at the time of the toggle operation. For example, a user can toggle the plots on a web page generated elsewhere and at some other time by the gnuplot...
hidden3d disables toggling in interactive terminals
History is not available if -e is specified on command line
Change behaviour of pointintervalbox
Special point types 1001 - 1007 are now in the development branch. They are supported by the postscript, cairo, gd, svg, x11, and qt terminals. I leave it for someone else to look into adding them to the windows terminal also.
Make sure cbrange is valid when palette(z) is called
x11: extra point types 1001 - 1007
cairo, qt, svg: extra pointtype 1007 is a background-filled star
svg: extra point types 1001 - 1006
qt: extra point types 1001 - 1006
cairo: extra point types 1001 - 1006
docs: documentation and demo for extra pointtypes starting at 1001
Reserve point types 1001 - 1020 for special use
Add a mechanism for failed evaluation to return NaN rather than int_error()
docs: typo in equation defining the Lambert W function
stats: Do not delete existing user variables that happen to be named NAME_*
I can add that this advice resolved the fatal error: QtCore file not found error for me on macOS Monterey 12.6.8. I'm building gnuplot-6.0.0 from source. I had to run export PKG_CONFIG_PATH=/usr/local/opt/qt@5/lib/pkgconfig/, which was not working without export. I then configured with ./configure --with-qt=qt5. Thanks for the helpful thread.
O well, You are at a right place. Yes your can create time stamps of the whole year using bamboo plugin link, its the easiest way now a days.
Ticket moved from /p/gnuplot/bugs/2698/ Can't be converted: _milestone: _priority:
It is hard to diagnose the problem with out an actual data set and script to plot it. My guess is that the issue has to do with the number of samples (set sample and set isosample). If these are inappropriate for the true coarseness of the data it can introduce artifacts that break up what should be a single contour line into multiple fragments. The program doesn't know that the fragments are supposed to form a continuous line, so it places a label on each fragment. This is stated in the documentation...
contourplot: contour levels with only one label: set cntrlabel interval -1 doesn't work
Thank you for the quick fix!
That appears to indicate that the Qt support packages are not installed. The text below is taken from gnuplot's INSTALL file for version 6, but may well be useful also for building version 5: macOS + homebrew ---------------- As of the time of writing (Feb 2023) homebrew provides a bottled gnuplot with Qt support. However it was not built with all the optional features or terminals that you might want. This is a quick summary of what is needed to configure and build from source. If you have gnuplot...
make error <QtCORE> not found
Wrong line clipping
Thanks for the explanation of why it is a feature and not a bug.
I may be overlooking an error; is there a specific value that is shown incorrectly, or that loses necessary precision? Note that whatever format is used to create the summary table, the underlying values are available individually by name: gnuplot> stats $dat using 1 ... * COLUMN: Mean: 0.0000 Std Dev: 0.0000 Sample StdDev: 0.0000 Skewness: 0.0000 Kurtosis: 1.7758 Avg Dev: 0.0000 Sum: 0.0000 Sum Sq.: 9.62500e-11 ... gnuplot> show variable STATS_ Variables beginning with STATS_: ... STATS_mean = 2.75e-06...
Wrong line clipping
What you are seeing is the difference between set clip one (the default) and set clip two. gnuplot> help set clip Syntax: set clip {points|one|two|radial} unset clip {points|one|two|radial} show clip Default state: unset clip points set clip one unset clip two unset clip radial Data points whose center lies inside the plot boundaries are normally drawn even if the finite size of the point symbol causes it to extend past a boundary line. `set clip points` causes such points to be clipped (i.e. not...
Wrong line clipping
bad stats output number format
Defining a datablock within an IF or an ELSE statement throws an error
Update Version 6 now supports some mousing operations in a multiplot, including coordinate readout of the most recently plotted component. That's one step closer to being able to retrieve coordinate information for all multiplot components.
gnuplot keeps script file open after loading it
I am just getting back to this now, since version 6 is out and apparently doing well except for the wxt terminal. The strange thing is, this script works for me now. The three data blocks are all defined successfully even though two of them are inside the if {} else {} bracketed clauses. It works in the 6.0.0 release version, current git tip, and also in the version 6 -rc1 which was from before the date on this request. So was this ever failing? Or maybe some other problem was contributing?
docs: typo in equation defining the Lambert W function
Error using stats and arrays with certain names in certain sequence
After looking at the code, I blame myself for committing an incomplete fix about five years ago (83a349a0). Deletion of an explicit list of variable names was added, but the original "delete everything" line elsewhere in the code was not removed. Fixed now.
Error using stats and arrays with certain names in certain sequence
stats: Do not delete existing user variables that happen to be named NAME_*
Sentinel One warning in command.c
x11: improve the vertical placement of rotated text (but it's still bad)
Allow 'using xticlabel(<foo>)' for style 'with boxplot'
docs: typos
My machines do not yet have pango 1.52 so I cannot directly test this. Can you get a stack trace from gnuplot while it is in the hung state? For instance use ps to find the pid for the gnuplot process; connect to it with gdb -p <pid>; type where to generate a stack trace.
The reason they are "terminal dependent" is that the details are different for each terminal. For older or stupider terminals (e.g. dot-matrix printers, the dumb terminal) the only thing the program can do is send a character or maybe a hardware command for "point". True vector devices were a thing back in the day (e.g. pen plotters) but almost all newer devices display pixel images created by a vector graphics library. From gnuplot's perspective all terminals implemented to use the same graphics...
Okay. As it works as long as you choose the appropriate font thats fine. Just for my understanding, how do the pre-defined (terminal dependent) point types work? Are they vector graphs? Thanks!