Re: Tests from Windows users and MXE build requested

[Andreas Weber]

Hi Philip, thanks for testing

Am 18.09.2014 23:45, schrieb Philip Nienhuis:
> On my Win7 64bit box:
> Yes, plot looks fine with Octave-3.6.4 MSVC, 3.8.2-64bit, and 3.9.0+/4.1.0+
> 32bit, the latter three all cross-built with MXE on Linux here.
> 
>> - Do you see all objects (axes, blue line, circles) in the generated
>> testplot.eps? Please report any differences.
> 
> Coincidentally I have gs9.07 (32b) both in C:\Programs (d/led & installed
> myself) and for octave-3.9.0+/4.1.0+ in the /bin/ subdirectory (built w.
> mxe-octave)
> 
> 3.6.4: .eps file looks fine. Plot window turns blue for a split second, as
> usual when Ghostscript prints a plot.
> 3.8.2-64b, 3.9.0+, 4.1.0+: No testplot.eps is generated at all. The plot
> windows flashes but that's it.

Interesting. The bug report I linked is, that no "data lines" were
generated but axes, grid and so on are visible.

> When I add gs9.14 to the PATH for Octave64-3.8.2 (Ghostscript isn't built w
> mxe-octave for 64-bit builds), the printed plot is generated fine. The plot
> windows flash lasts a bit longer than with gs9.07.
> Same for 3.9.0+ & 4.1.0+ (added gs9.14/bin before all other PATH entries).
> 
> So it is related to the gs version.

gs versions >=9.09 have removed pswrite
(http://savannah.gnu.org/bugs/?func=detailitem&item_id=41141) so If we
suggest 9.14 this would break PS output.

>> If you see any errors please copy them verbatim. If you see differences
>> please report the exact version of your MXE build and where you've
>> downloaded it. Please also add the vendor and type of your GPU and if
>> you have up to date driver for your GPU installed.
> 
> GPU: ?? it's a Lenovo X230 laptop w. Intel HD graphics 4000
> I haven't updated the driver since I purchased this laptop (1,5 years ago
> IIRC).

I remember that I've chatted on IRC with someone who reported problems
like missing lines when printing. After he updated the drivers for the
graphics card this problem was gone. So this was only a wild guess that
the driver might also be a source for problems. Btw: We also have
display issues with OpenGL in the bugtracker which are only visible with
AMD GPUs, Nvidia and Intel renders fine:
http://savannah.gnu.org/bugs/?func=detailitem&item_id=42580

Thanks again, Andy