gnuplot-beta

Release 5.2.7

[Ethan M. (UW) <me...@uw...>]

Release 5.2.7 tarball and Windows binaries are ready and can be
downloaded from 

     https://sourceforge.net/projects/gnuplot/files/gnuplot/5.2.7/[1]



		GNUPLOT Version 5.2.7 Release Notes
		===================================

These release notes are for version 5.2 patchlevel 7 (5.2.7).
This release contains bug-fixes, a few changes back-ported from the
development version, and some minor new features.
The largest change is major revision to the fig and pm terminals
bringing them into line with capabilities of other version 5 terminals.

Please see the ChangeLog file for a complete list of changes made during the
course of development from 5.0 to 5.2.

Release Notes date: 29-May-2019

Changes in 5.2.7
================

* NEW "set pm3d noclipcb" will skip drawing quadrangles with color < cbmin
* NEW pm terminal supports utf8, dashed lines, bold/italic text
* NEW imaginary component of array values is available to "using" in column 3
* NEW gnuplot --slow may help use of qt terminal with slow system font server
* CHANGE splot "with points pt 0" acts like "with dots"
* CHANGE (tikz) improved default interpretation of gnuplot arrowhead style
* CHANGE teach hidden3d to handle double-headed vectors (arrowstyle "heads")
* CHANGE "unset title" or "reset" resets all title properties to default
* CHANGE font names passed to enhanced text processing may be in quotes
* CHANGE cairo terminals accept fractional font sizes
* CHANGE "plot with table" does not output trailing field separator
* CHANGE better centering for clustered histograms
* CHANGE clear STATS_* variables before performing analysis
* CHANGE cairo, libgd, and svg terminals handle LT_NODRAW by not drawing at all
* CHANGE svg terminal clickable icons are in-line data rather than external refs
* CHANGE revise space allocated for ylabel; allow text justification for ylabel
* CHANGE reduce overhead, making it feasible to have 10^6 hypertext labels
* FIX "plot with table" unnecessarily limited string column output to 60 chars 
* FIX (windows) make sure graph window is shown even if the program is "hidden"
* FIX hidden3d mode was not correctly dealing with arrowstyle backhead
* FIX sprintf in 'using' spec must not clobber numeric locale
* FIX tikz - request for polygon with 0 vertices could hang terminal
* FIX lua/tikz - do not attempt to ftruncate output being piped through stdout
* FIX tkcanvas - do not attempt to ftruncate output being piped through stdout
* FIX handle NaN values in input stream to "stats FOO matrix"
* FIX prevent extra read past the end of a datablock holding matrix data
* FIX pm - issues in box and polygon fill, (wide) lines, enhanced text, images
* FIX dropped character in piped input stream if plot window is manually closed
* FIX extraneous dots and potential segfault from splot with contour labels
* FIX boxplots - color sequence correctly starts with specified linetype
* FIX incomplete initialization led to segfault from "gnuplot -c foo.gp baz"
* FIX Allow "pm3d depthorder base" to handle 3D quadrangles with logscale z

	happy gnuplotting

		Ethan

--------
[1] https://sourceforge.net/projects/gnuplot/files/gnuplot/5.2.4/gnuplot-5.2.4.tar.gz

Re: Release 5.2.7

[Achim G. <Str...@ne...>]

Ethan Merritt (UW) writes:
> * CHANGE clear STATS_* variables before performing analysis

It took me a while to relaize, but that particular change killed on of
my scripts.  That is using data that doesn't always exist, so I have to
test for that like this:

--8<---------------cut here---------------start------------->8---
STATS_records=0
stats [24*sr:24*sq] data using ($4==10?t($1):1/0):($2) nooutput
TEST=STATS_records>0  
--8<---------------cut here---------------end--------------->8---

That gives you a warning from stats, but you could then skip additional
code using the non-existing data (that would error out) based on the
result in TEST.

That no longer works since you're now getting an error (terminating the
script) at the last line.  Note that you can't do something like
exist(STATS_record) either like one could reasonably expect to work.
Any workarounds?


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

SD adaptations for KORG EX-800 and Poly-800MkII V0.9:
http://Synth.Stromeko.net/Downloads.html#KorgSDada

Re: Release 5.2.7

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

On Thursday, June 6, 2019 11:30:40 AM PDT Achim Gratz wrote:
> Ethan Merritt (UW) writes:
> > * CHANGE clear STATS_* variables before performing analysis
> 
> It took me a while to relaize, but that particular change killed on of
> my scripts.  That is using data that doesn't always exist, so I have to
> test for that like this:
> 
> --8<---------------cut here---------------start------------->8---
> STATS_records=0
> stats [24*sr:24*sq] data using ($4==10?t($1):1/0):($2) nooutput
> TEST=STATS_records>0  
> --8<---------------cut here---------------end--------------->8---
> 
> That gives you a warning from stats, but you could then skip additional
> code using the non-existing data (that would error out) based on the
> result in TEST.
> 
> That no longer works since you're now getting an error (terminating the
> script) at the last line.  Note that you can't do something like
> exist(STATS_record) either like one could reasonably expect to work.
> Any workarounds?
> 
> 
> Regards,
> Achim.

I may not understand the required test.  The following works for me:

datafile = "whatever"

 if (exists("STATS_records")) { print 'got it already' }
else { stats datafile }

I would probably use an explicit name rather than STATS:

if (exists("MYDAT_records")) { print 'already did MYDATA' }
else { stats datafile name "MYDAT" }
show var MYDAT

Is the problem that STATS_records might exist for reasons other
than having run a previous STATS command?   That was exactly
the sort of confusion that the change you mention was intended
to prevent.

	Ethan





	Ethan

Re: Release 5.2.7

[Achim G. <Str...@ne...>]

Achim Gratz writes:
> That no longer works since you're now getting an error (terminating the
> script) at the last line.  Note that you can't do something like
> exist(STATS_record) either like one could reasonably expect to work.
> Any workarounds?

Foundit, thinko on my side.  Correcting the test to

--8<---------------cut here---------------start------------->8---
stats [24*sr:24*sq] data using ($4==10?t($1):1/0):($2) nooutput
TEST=exists("STATS_records")
--8<---------------cut here---------------end--------------->8---

actually works as expected.  I was missing the quotes originally.  Now,
what's still left is to suppress the warning about no valid data points
from the stats invocation.


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

SD adaptation for Waldorf rackAttack V1.04R1:
http://Synth.Stromeko.net/Downloads.html#WaldorfSDada

Re: Release 5.2.7

[Achim G. <Str...@ne...>]

Ethan A Merritt writes:
> I may not understand the required test.

The problem is that I need to filter the data and there may not be any
left after that.  Having no data points unfortunately is fatal to some
of the commands I'm using later, so I need to ensure it doesn't happen
and skip those commands if necessary.


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

SD adaptations for KORG EX-800 and Poly-800MkII V0.9:
http://Synth.Stromeko.net/Downloads.html#KorgSDada