Fri 26 Dec 2014 03:33:13 AM UTC, comment #14:
I still think the line connecting the outer edge of the surface to the x-y plane is unnecessary, but this is clearly a gnuplot decision rather than anything Octave is feeding it.
The coloring for patches doesn't seem to be correct. Even when commenting out the lines you suggest I don't get entirely the correct results. Worse, with those lines commented out I get very bad results for other sorts of color displays.
I'm going to close this report and open a new one which is specifically about patch coloring.
|
Tue 23 Dec 2014 09:23:21 PM UTC, comment #13:
Hi,
the coloring difference is becuase of colormap change
(lines 1633-174) in _go_draw_axes_.m
if you comment them out, you will get the same coloring as in fltk.
The misplaced vertical line is actually not misplaced, but a
vertical line gnuplot add in 3D image. For examples:
gnuplot draws the frontal horizontal line, from
(-1,-1,0) to (-1,-1,exp(-0.02)).
with fltk you will not see this line.
Avinoam
|
Mon 22 Dec 2014 03:53:56 AM UTC, comment #12:
I applied the patch for trisurf here (http://hg.savannah.gnu.org/hgweb/octave/rev/c029983a0389). I've re-titled the bug report to reflect the remaining issue.
|
Mon 15 Dec 2014 06:57:44 PM UTC, comment #11:
The misplaced line, and the incorrect coloring, are both outstanding issues.
|
Mon 15 Dec 2014 06:05:44 PM UTC, comment #10:
Hi Rik,
thanks, I prefer to remain anonymous.
But, please do not close this bug,
because I am stil truing to debug the misplaced line.
Avinoam
|
Mon 15 Dec 2014 04:06:06 PM UTC, comment #9:
@Avinoam: I am going to apply your patch for trisurf. Traditionally we credit the contributor with "firstname lastname <e-mail address>" in the commit log. Is this good for you, or do you prefer to remain anonymous?
|
Sun 07 Dec 2014 12:26:04 AM UTC, comment #8:
I applied the patch and I no longer get the error messages. However, the resulting plots are colored quite differently between FLTK and gnuplot. Is this a different error being exposed or is the fix here not quite right for the color data? See the attached files.
(file #32613, file #32614)
|
Mon 24 Nov 2014 08:15:16 PM UTC, comment #7:
Hi,
I have uploaded a patch for this bug.
I do not know if it's the best way to fix the bug, or if
this patch don't cause other problems.
After applying this patch, gnuplot results are almost the
same as fltk results (except from one misplaced line in gnuplot
graph).
Please check this.
Thanks,
Avinoam
(file #32536)
|
Tue 21 Oct 2014 10:35:25 PM UTC, comment #6:
Confirmed. The bug is still present on the development branch as well (10/21/14). There are many reasons why Octave developers are trying to move away from gnuplot to an OpenGL based plotting engine and this just adds one more.
|
Tue 21 Oct 2014 02:26:32 PM UTC, comment #5:
Here is the content of the file tetra.m
Here is the Octave output :
|
Tue 21 Oct 2014 02:23:40 PM UTC, comment #4:
Here is the Octave output :
|
Tue 21 Oct 2014 07:27:56 AM UTC, comment #3:
Hm, this is strange...
(after closing verbatim block)
|
Tue 21 Oct 2014 07:26:28 AM UTC, comment #2:
@Sigurd: You should use verbatim blocks to guard the code. I repost it for you. I also can confirm that the error below is raised in Octave 3.8.2 only with graphics_toolkit gnuplot. FLTK works as expected.
|
Tue 21 Oct 2014 07:20:57 AM UTC, comment #1:
Hi,
Thanks for your bug report. There must have been a problem with verbatim markup. Can you repost your example, and the error message you get?
|
Mon 20 Oct 2014 06:15:55 PM UTC, original submission:
I tried demo("trisurf",1) and this works
The following script which I wrote produces an error message. I don't think it should...
|