Documented that 'smithchart mirrored' produces unexpected results and generated warning
Fix broken tikz@do@circle and missing tikz@do@ellipse hook
Merge branch 'master' of ssh://git.code.sf.net/p/pgfplots/code
fixed table test
test files
updated ignore file
Compatibilty mode for tikz@do@circle
For the record: To circumvent this problem use the syntax: circle [radius=<radius>]. Relevant questions on TeX.SX are: - https://tex.stackexchange.com/q/524076 - https://tex.stackexchange.com/q/524114 - https://tex.stackexchange.com/q/528207
Related Q&As: https://topanswers.xyz/tex?q=503#question https://tex.stackexchange.com/questions/524114
Origin of this topic is my answer at https://tex.stackexchange.com/questions/526852.
Visual Artefact When Using `axis line style = {rounded corners}`
[manual] `/pgfplots/boxplot/data filter/.code` not documented
done in commit https://sourceforge.net/p/pgfplots/code/ci/d962376b5ab77d084f730aedd230c386363032ac/
[manual] `/pgfplots/boxplot/data filter/.code` not documented
- added key `/pgfplots/boxplot/data filter/.code` to the manual
Related Q&As: https://topanswers.xyz/tex?q=503#question (currently still in private beta, so only visible for users) https://tex.stackexchange.com/questions/524114
For the record: To circumvent this problem use the syntax: circle [radius=<radius>]. Relevant questions on TeX.SX are: - https://tex.stackexchange.com/q/524076 - https://tex.stackexchange.com/q/524114
Number of arguments for circle changed in tikz
(sorry for messing up the code formatting in the post above)
Number of arguments for circle changed in tikz
I fully agree with this feature request. I would add to the "outer" values of legend pos the following value as well: outer west (could make sense if the user prefers axis y line*=right) outer east outer south west outer south east outer north west outer north east
minor grid below major grid (?)
bug when trying to store `\pgfplotstablegetrowsof` with `\pgfmathsetmacro`
bug when trying to store `\pgfplotstablegetrowsof` with `\pgfmathsetmacro`
I have modified the above MWE a bit to exclude stuff that is not relevant. I can confirm that this bug occurs (only), when the first coordinate is NaN (or filtered away).
\pgfplots@stacked@diff undefined
ConTeXt: \AtBeginDocument is defined as {#1}
I was too hasty. This is actually a problem in PGF. However, I can't seem to close the issue, there is no button. Please close this for me.
ConTeXt: \AtBeginDocument is defined as {#1}
Use MATHEMATICA as external program
\draw has problems with declared functions
[bug] position of sub-figure
Incorrect bar position when `bar width` is defined in axis unit with `bar shift auto = 0`
Incorrect bar position when 'bar width' is defined in axis unit with 'bar shift auto = 0'
[Feature request] Render nan-only plots without compilation error
On TikZ v3.1.4b and PGFPlots v1.16 I get the desired result. On which versions are you? Could you add an image/screenshot to show the odd behaviour?
[manual] layer `pre main` is missing in the layer definitions
fixed in https://sourceforge.net/p/pgfplots/code/ci/2581cafe7f2504fa7191f9e8b0d55cf84919217c/
- added layer `pre main` to the documentation part of `layers/standard/.define layer set` and `layers/axis on top/.define layer set` in `pgfplots.reference.layers.tex`
Merge branch 'master' of https://git.code.sf.net/p/pgfplots/code
- minor stuff
[manual] layer `pre main` is missing in the layer definitions
Expected behaviour with pdflatex looks like this:
XeLaTex with pgfplots displays wrong color?
Sounds great. Unfortunately I don't have a clue about (PGFPlots) code. But I already contacted Christian who hopefully will have a look in the near future ...
A short update is there. Arbitrary numbers of different x/y data are now possible. At least one row of x values have to be present at the moment. Table delimiters are customizable (but I am not really sure about the lua pattern-matching syntax. Somehow, it is different for semicolons and whitespaces). I am not sure how to handle a separate matrix for point-meta values correctly since I have no idea of a suitable format. Maybe this point is not that important. Can you give me a hint where to look...
The solution on stackexchange is more suited for the particular problem there. Things to consider in a more general case (please add your suggestions): - there may (or may not) be a first row with y values - there may be zero, one or several columns with x values - without x or y values, there is just a plain z matrix - not only a z matrix, but a second one for point-meta values - custom comment chars - custom table delimiters - extracting mesh cols and mesh rows from the matrix I will try to implement...
add support for "matrix-like" input for 3D data
Faekynn provided some solutions of the problems e.g. https://tex.stackexchange.com/a/502680/95441 using Lua code. Maybe this can be generalized and implemented?
Axis break feature
the '\theta=20,\phi=20' sub-figure is shown on the upper-left and outside the figure
[bug] position of sub-figure
fixed test issues
renamed helper scripts
adapt unit test expectations
Fixed tests.
pgfmathprintnumber: no unity mantissa
[dateplot][groupplots] colorbar causes an error when using dates in groupplots
exponent in color specs
Thanks for the request. This might be possible to fix, although probably not in pgfplots but in pgf/tikz (or in xcolor?). Depending on how many colors you have to export, you might be better off if you generate a sequence of \definecolor invocations (see the xcolor manual or the pgfmanual.pdf for details). This saves lots of time for the TeX translation. Example invocation in the RGB space (range 0-255): \pgfutil@definecolor{BuGn-A}{RGB}{247,252,253} I believe the space 'rgb' (lower case) accepts...
Fixed regressions
used FPU for tick foreach
added test
Merge branch 'master' of ssh://git.code.sf.net/p/pgfplots/code
\pgfmathprintnumber: implemented PGF 'retain unit mantissa=true|false' (feature #92)
fixed displayDiff script
fixed SIMPLE scientific notation in tick arguments
normalized reference revisions
fixed lots of broken tests (same issue as previous commits)
improved test coverage
fixed lots of tests.
reset reference rev
exponent in color specs
[layers] don't seem to work for `title style`
[nodes near coords] are under grid lines when `axis on top` is active
- consistently used `samples y=1` for line plots in a 3d diagram (`\addplot3`)