On Tue, 23 Jun 2020, Sven Schreiber wrote:
> Am 22.06.2020 um 21:00 schrieb Artur Tarassow:
>> Am 22.06.20 um 20:24 schrieb Artur Tarassow:
>>> Am 22.06.20 um 20:16 schrieb Sven Schreiber:
>
>>>> Thanks, Artur. It seems to work fine here on Windows 10, however; no
>>>> overlap occurs.
>>>>
>>>> Cristián, what is the gretl version you are using, please? (And
>>>> are you
>>>> actually on Windows?)
>>>
>>> Actually, I can confirm that weird behavior here using WIndows 10 and
>>> gretl 2020b (build data 2020-04-11). Also, gnuplot version 5.2.
>>> patchlevel 4 is used. The created gnuplot file looks fine to me but
>>> the plot is broken.
>> A follow up on this: Also latest snapshot does not work on Win10.
>
> Hmm, I tested on a self-compiled version from 10 days ago, so that's
> interesting.
> I believe that in such a setup the gnuplot program is taken from
>
http://ricardo.ecn.wfu.edu/pub/gretl/winbuild/gp524w64.tar.gz, but when
> I fire up the wgnuplot.exe from the installation location
> C:/programme/gretl it says Version 5.2 patchlevel 6. (where the filename
> would suggest patchlevel 4 instead?)
>
> But I just cross-checked on another Windows machine with an "ordinary"
> snapshot, and there I'm seeing the same problem that you showed us.
> There gnuplot reports itself as 5.2 patchlevel 4.
>
> Not sure whether the patchlevel is supposed to make a difference, but
> that would be my working hypothesis right now. Maybe they fixed
> something relevant that is Windows-specific in some way?
Looks like it. I'll see if I can build gnuplot 5.2.8 (December 2019)
for our Windows packages.