(
Jump to the original submission)
Wed 09 Aug 2017 05:54:22 PM UTC, comment #14:
Works for me as well.
WRT #12 we do something like that for
_osmesa_print_ test.
Perhaps we should modify "plotimages.m" to include
assert on file size or similar. Now I think that it is better
to use compressed file size, or an actual Shannon's entropy calculation rather than pixel sum per RGB channel, but that
is a finer detail.
Also, there are used to be dump_plot_demos scripts which could be
made into "make check_images" in similar fashion (with added signature check). Since it takes a while to run those I would
rather make them a special target than to be a part of a standard "make check".
Dmitri.
--
|
Wed 09 Aug 2017 05:07:03 PM UTC, comment #13:
thanks Rik
It works for me.
this can be closed.
|
Wed 09 Aug 2017 04:40:49 PM UTC, comment #12:
I pushed a fix here http://hg.savannah.gnu.org/hgweb/octave/rev/cbea10ff137e. Test and see if it works for you.
What this really points out is that we need some way of developing regression tests that includes the plotting functions.
|
Wed 09 Aug 2017 01:42:54 PM UTC, comment #11:
hg id "3544f88a2bb5" is already broken.
Dmitri.
--
|
Wed 09 Aug 2017 01:40:18 PM UTC, comment #10:
Changeset:
23818 (784347c1b7e7) Avoid unnecessary warning when printing hidden axes (bug #51374). …
is the one that is at fault!
|
Wed 09 Aug 2017 12:56:46 PM UTC, comment #9:
hg id = 5b5b021b1476
still works for me
Dmitri.
--
|
Wed 09 Aug 2017 12:56:08 PM UTC, comment #8:
Ok I see what you see now
hg id 0d5fc6e4a96c
I will bisect
|
Wed 09 Aug 2017 12:31:59 PM UTC, comment #7:
OK, may be I was not clear in the original bug report.
This is used to work (and still works with 4.2.1).
It stopped working with a recent development tip.
Dmitri.
|
Wed 09 Aug 2017 12:30:50 PM UTC, comment #6:
hg id = 5b5b021b1476
|
Wed 09 Aug 2017 12:29:04 PM UTC, comment #5:
I am on
ubbuntu 16.04
octave 4.2.1+ latest stable compiled by me.
qt graphics
|
Wed 09 Aug 2017 12:28:10 PM UTC, comment #4:
What is hg id?
Dmitri.
--
|
Wed 09 Aug 2017 12:26:26 PM UTC, comment #3:
polar(0:0.1:10pi, 0:0.1:10pi)
#polar([1 2 3],[.1 .3 .4])
#plot t1.eps
print t1.eps
file is attached
(file #41460)
|
Wed 09 Aug 2017 12:19:44 PM UTC, comment #2:
It does not work either, but the picture is all white.
t1.eps.gz is attached.
I am not a postscript expert, but it looks to me that all the
actual plot code is missing from the file.
(using gnuplot backend works)
Dmitri.
--
(file #41459)
|
Wed 09 Aug 2017 10:02:42 AM UTC, comment #1:
try
polar(0:0.1:10pi, 0:0.1:10pi)
print t1.eps
|
Wed 09 Aug 2017 04:59:14 AM UTC, original submission:
As of 12203140139f tip @
prints of polar plot (e.g. polar.png or .eps or .pdf
in doc/interpreter) are all black.
The same can be reproduced by simple
(though in that case they are all white?!)
I attached the polar.eps from the build -- perhaps somebody can spot a problem there.
Dmitri.
--
|
(Note: upload size limit is set to 16384 kB, after insertion of the required
escape characters.)
Attach File(s):
Comment:
Depends on the following items: None found
Items that depend on this one: None found
Follow 6 latest changes.