Menu
▾
▴
gnuplot-beta
|
From: Ethan A M. <me...@uw...> - 2023-05-18 21:37:53
|
I have placed a pre-release tarball of 5.4.7 on the testing area of SourceForge
https://sourceforge.net/projects/gnuplot/files/gnuplot/testing/gnuplot-5.4.7testing.tar.gz
The number of bug fixes is surprisingly large, given how late this is in the
5.4 series. Many of these bugs date back to earlier versions of gnuplot but
either were not reported or not considered fixable until recently.
I had been thinking that this would be the last patchlevel release for 5.4
but now I am not so sure.
If no problems are reported, I plan to use this for release of 5.4.7
next week.
Changes in 5.4.7
================
* FIX pseudofiles '+','++' must ignore "set datafile columnheader" Bug #2585
* FIX data from an array must ignore "set datafile columnheader" Bug #2585
* FIX various problems with unbounded plot iteration Bug #2589
* FIX allow plot ... smooth acsplines with filledcurves Bug #2592
* FIX "plot title at {end|beginning}" belongs in KEYSAMPLE layer
* FIX handle empty string in last field of csv file
* FIX Handle nonlinear axes during contour generation Bug #2593
* FIX yerrorbars key sample should match bars in the plot Bug #2598
* FIX "plot with table" must skip all smoothing operations Bug #2610
* FIX cairolatex/epslatex: prevent doubled text in opaque textboxes
* FIX wxt: versions prior to 3.0 do not support XDG Bug #2587
* FIX wxt: fontscale and fontsize handling was inconsistent Bug #2373
* FIX tikz epslatex cairolatex: updated to work with TeXLive2023 Bug #2613
* FIX correctly report cardinality of an empty array is zero Bug #2616
* FIX ignore datafile separator when plotting from an array
* FIX combination of pm3d options "at b" or "at t" with "depthorder"
* FIX combination of 'autotitle columnhead' and 'using col("key")' Bug #2620
* FIX do not complain if XDG files are not found because no $HOME Bug #2623
* FIX windows terminal: arrow length corrected for aspect ratio Bug #2565
Ethan
|
|
From: Tatsuro M. <tma...@ya...> - 2023-05-19 01:32:49
|
I have executed the test build on windows (MinGW64). I ended successfully Tatsuro > ----- Original Message ----- > > From: "Ethan A Merritt" <me...@uw...> > To: "gnu...@li..." <gnu...@li...> > Date: 2023/05/19 金 06:38 > Subject: Gnuplot 5.4.7 release tarball available for testing > > > I have placed a pre-release tarball of 5.4.7 on the testing area of SourceForge > > https://sourceforge.net/projects/gnuplot/files/gnuplot/testing/gnuplot-5.4.7testing.tar.gz > > The number of bug fixes is surprisingly large, given how late this is in the > 5.4 series. Many of these bugs date back to earlier versions of gnuplot but > either were not reported or not considered fixable until recently. > I had been thinking that this would be the last patchlevel release for 5.4 > but now I am not so sure. > > If no problems are reported, I plan to use this for release of 5.4.7 > next week. > > > Changes in 5.4.7 > ================ > > * FIX pseudofiles '+','++' must ignore "set datafile columnheader" Bug #2585 > * FIX data from an array must ignore "set datafile columnheader" Bug #2585 > * FIX various problems with unbounded plot iteration Bug #2589 > * FIX allow plot ... smooth acsplines with filledcurves Bug #2592 > * FIX "plot title at {end|beginning}" belongs in KEYSAMPLE layer > * FIX handle empty string in last field of csv file > * FIX Handle nonlinear axes during contour generation Bug #2593 > * FIX yerrorbars key sample should match bars in the plot Bug #2598 > * FIX "plot with table" must skip all smoothing operations Bug #2610 > * FIX cairolatex/epslatex: prevent doubled text in opaque textboxes > * FIX wxt: versions prior to 3.0 do not support XDG Bug #2587 > * FIX wxt: fontscale and fontsize handling was inconsistent Bug #2373 > * FIX tikz epslatex cairolatex: updated to work with TeXLive2023 Bug #2613 > * FIX correctly report cardinality of an empty array is zero Bug #2616 > * FIX ignore datafile separator when plotting from an array > * FIX combination of pm3d options "at b" or "at t" with "depthorder" > * FIX combination of 'autotitle columnhead' and 'using col("key")' Bug #2620 > * FIX do not complain if XDG files are not found because no $HOME Bug #2623 > * FIX windows terminal: arrow length corrected for aspect ratio Bug #2565 > > > Ethan > > > > > _______________________________________________ > gnuplot-beta mailing list > gnu...@li... > Membership management via: https://lists.sourceforge.net/lists/listinfo/gnuplot-beta > |
|
From: Petr M. <mi...@ph...> - 2023-05-19 15:13:42
|
Hello Tatsuro,
> I have executed the test build on windows (MinGW64).
I see you upload the Windows binary also as a compressed 7z file. Would not be
the traditional zip file more portable? Recently, I wanted to upgrade an older
gnuplot on a Windows machine without admin privileges, but Windows 10 cannot
unzip 7z natively.
Tech note: the EXAFS software Athena uses gnuplot with wxt terminal for
drawing, but on that Windows 10 machine the cross of the ruler ("r" hotkey) is
invisible (white cross on white background?), for all windows gnuplots
downloadable as zip files. Thus peak distance measurement is difficult. I
wonder whether it was a problem on only that machine or the problem is more
frequent.
Greetings, Petr
|
|
From: Ethan A M. <me...@uw...> - 2023-05-19 16:40:47
|
On Friday, 19 May 2023 07:51:47 PDT Petr Mikulik wrote:
> Hello Tatsuro,
>
> > I have executed the test build on windows (MinGW64).
[snip]
> Tech note: the EXAFS software Athena uses gnuplot with wxt terminal for
> drawing, but on that Windows 10 machine the cross of the ruler ("r" hotkey) is
> invisible (white cross on white background?), for all windows gnuplots
> downloadable as zip files. Thus peak distance measurement is difficult. I
> wonder whether it was a problem on only that machine or the problem is more
> frequent.
This is probably an issue with what version of the gtk library is used.
gtk2 provided a "invert the color" pen that would always draw white on a dark
background or black on a light background.
For some reason gtk3 does not support that option. This was bug #2523
https://sourceforge.net/p/gnuplot/bugs/2523/
It was supposedly fixed in version 5.4.4
commit 10152d7a736da4ab289c12d532cae50683a8306b
Date: Thu May 5 16:56:06 2022 -0700
wxt: Flip color of zoombox / rubber band / ruler on dark background
Bug #2523
Maybe that fix did not work for Windows 10?
Or maybe for some reason on your machine gnuplot thinks the
terminal background is dark even though in fact it is white?
Ethan
--
Ethan A Merritt
Biomolecular Structure Center, K-428 Health Sciences Bldg
MS 357742, University of Washington, Seattle 98195-7742
|
|
From: Achim G. <Str...@ne...> - 2023-05-20 06:15:06
|
Ethan A Merritt writes: > I have placed a pre-release tarball of 5.4.7 on the testing area of SourceForge […] > If no problems are reported, I plan to use this for release of 5.4.7 > next week. The Cygwin build has shown no anomalies. Regards, Achim. -- +<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+ |
|
From: Tatsuro M. <tma...@ya...> - 2023-05-21 04:59:14
|
Hi Peter
Preparation zip compressed binary is easy.
I will prepare installer version, 7zip, and zip packages for 5.4.7 and later
Tatsuro
> ----- Original Message -----
>
> From: "pm" <mi...@ph...>
> To: "Tatsuro MATSUOKA" <tma...@ya...>
> Cc: "beta" <gnu...@li...>
> Date: 2023/05/19 金 23:52
> Subject: Re: Gnuplot 5.4.7 release tarball available for testing
>
>
> Hello Tatsuro,
>
> > I have executed the test build on windows (MinGW64).
>
> I see you upload the Windows binary also as a compressed 7z file. Would not be
> the traditional zip file more portable? Recently, I wanted to upgrade an older
> gnuplot on a Windows machine without admin privileges, but Windows 10 cannot
> unzip 7z natively.
>
> Tech note: the EXAFS software Athena uses gnuplot with wxt terminal for
> drawing, but on that Windows 10 machine the cross of the ruler ("r" hotkey) is
> invisible (white cross on white background?), for all windows gnuplots
> downloadable as zip files. Thus peak distance measurement is difficult. I
> wonder whether it was a problem on only that machine or the problem is more
> frequent.
>
> Greetings, Petr
>
|
|
From: Tatsuro M. <tma...@ya...> - 2023-05-21 22:02:15
|
I have uploaded Windows binary packages (exe, 7z. and zip).
Tatsuro
> ----- Original Message -----
>
> From: "Tatsuro MATSUOKA via gnuplot-beta" <gnu...@li...>
> To: "pm" <mi...@ph...>
> Cc: "beta" <gnu...@li...>
> Date: 2023/05/21 日 13:59
> Subject: Re: Re: Gnuplot 5.4.7 release tarball available for testing
>
>
> Hi Peter
>
> Preparation zip compressed binary is easy.
> I will prepare installer version, 7zip, and zip packages for 5.4.7 and later
>
> Tatsuro
>
> > ----- Original Message -----
> >
> > From: "pm" <mi...@ph...>
> > To: "Tatsuro MATSUOKA" <tma...@ya...>
> > Cc: "beta" <gnu...@li...>
> > Date: 2023/05/19 金 23:52
> > Subject: Re: Gnuplot 5.4.7 release tarball available for testing
> >
> >
> > Hello Tatsuro,
> >
> > > I have executed the test build on windows (MinGW64).
> >
> > I see you upload the Windows binary also as a compressed 7z file. Would not be
> > the traditional zip file more portable? Recently, I wanted to upgrade an older
> > gnuplot on a Windows machine without admin privileges, but Windows 10 cannot
> > unzip 7z natively.
> >
> > Tech note: the EXAFS software Athena uses gnuplot with wxt terminal for
> > drawing, but on that Windows 10 machine the cross of the ruler ("r" hotkey) is
> > invisible (white cross on white background?), for all windows gnuplots
> > downloadable as zip files. Thus peak distance measurement is difficult. I
> > wonder whether it was a problem on only that machine or the problem is more
> > frequent.
> >
> > Greetings, Petr
> >
>
>
>
> _______________________________________________
> gnuplot-beta mailing list
> gnu...@li...
> Membership management via: https://lists.sourceforge.net/lists/listinfo/gnuplot-beta
>
|
|
From: Jun T <tak...@kb...> - 2023-06-01 05:27:09
|
> 2023/05/22 7:02、Tatsuro MATSUOKA via gnuplot-beta <gnu...@li...>のメール:
>
> I have uploaded Windows binary packages (exe, 7z. and zip).
If I install by using gp547-win64-mingw.exe,
"gnuplot 5.4 patchlevel 7alpha" is used for the name of
the shortcuts on the Desktop and in the Start Menu.
The same name ("alpha") is also used at the top of the
help document (and maybe in other places).
It works fine, but users may think that it is not
the release version but an "alpha" version.
|
|
From: Ethan A M. <me...@uw...> - 2023-06-01 18:09:26
|
Ugh. I wonder how that slipped through? My fault in any case.
I am not sure what is best
- repackage 5.4.7 with the version number changed?
- release a version 5.4.8 that is identical to 5.4.7 except that it
correctly reports its version number?
Either way may lead to some confusion.
The first option would change the check-sums for the tarball named
gnuplot-5.4.7.tar.gz, which has led to concern in the past.
The second seems a bit odd if there are no actual changes or fixes
between 7 and 8, but I guess it would not hurt anything.
Ethan
On Wed, May 31, 2023 at 10:27 PM Jun T <tak...@kb...> wrote:
> > 2023/05/22 7:02、Tatsuro MATSUOKA via gnuplot-beta <gnu...@li...>のメール:
> >
> > I have uploaded Windows binary packages (exe, 7z. and zip).
>
> If I install by using gp547-win64-mingw.exe,
> "gnuplot 5.4 patchlevel 7alpha" is used for the name of
> the shortcuts on the Desktop and in the Start Menu.
>
> The same name ("alpha") is also used at the top of the
> help document (and maybe in other places).
>
> It works fine, but users may think that it is not
> the release version but an "alpha" version.
>
>
>
> _______________________________________________
> gnuplot-beta mailing list
> gnu...@li...
|
|
From: Jun T <tak...@kb...> - 2023-06-06 09:24:05
|
> 2023/06/02 3:08、Ethan A Merritt <me...@uw...>のメール: > > - repackage 5.4.7 with the version number changed? > > - release a version 5.4.8 that is identical to 5.4.7 except that it > correctly reports its version number? > > Either way may lead to some confusion. How to fix this is up to you, but: > The first option would change the check-sums for the tarball named > gnuplot-5.4.7.tar.gz, which has led to concern in the past. The release was just two weeks ago, but, yes, there is a *small* possibility that a packager etc. has already created a spec/recipe with the current check-sum. > The second seems a bit odd if there are no actual changes or fixes > between 7 and 8, but I guess it would not hurt anything. I think most of users do not care if .7 is "skipped". |
Thanks for helping keep SourceForge clean.
X