Tue 10 Jan 2017 06:15:06 PM UTC, comment #10:
I pushed the patch here with minor modifications (http://hg.savannah.gnu.org/hgweb/octave/rev/29ea25593dd7). Thanks for the work on this.
|
Tue 10 Jan 2017 11:47:31 AM UTC, comment #9:
Thank you Rik for reviewing the patch.
The attached one should apply and compile cleanly on a current tip. It also fixes an issue with the previous patch when the ticks are spaced unequally.
(file #39425)
|
Mon 09 Jan 2017 11:32:50 PM UTC, comment #8:
@Markus: Could you refresh your patch from July? It's not applying cleanly against the dev branch. As for the question in comment #7, no, there isn't a particular shortcut to do concatenation in C++. The for loops that you are using are pretty standard. If you were trying for ultimate performance you could use Matrix.fortran_vec () to get a pointer to the underlying array and then use std::memcpy, but that kind of performance probably isn't required here.
|
Thu 14 Jul 2016 07:53:06 PM UTC, comment #7:
Attached, please, find a patch that adds support for minor grid lines outside the major grid range. This only fixes the error for linear plots like the example in comment #1.
I did not find where this was broken since 4.0.0 and I don't know whether this is the right approach to fix it.
Btw. is there a way for concatenating matrices or vectors in C++ instead of using loops? E.g. like doing [A,B] for row vectors in Octave.
(file #37876)
|
Tue 28 Jun 2016 12:14:20 PM UTC, comment #6:
Just to note that we should only look at fixing the behavior of the example in comment #1. In comment #4 we actually expect that there is a major grid line along x = 200 and y = 200 (and this is also what happened in 4.0).
|
Sun 23 Aug 2015 07:21:28 PM UTC, comment #5:
Confirmed. This is a regression from 4.0.0.
|
Fri 07 Aug 2015 04:36:01 PM UTC, comment #4:
One more bit of data: varying the plot
range produces other errors in the plot.
For example:
produces extraneous lines at the top and right borders.
So, axis tight needs more work, or, in effect, returning
it to how it worked in 4.0.0.
I cannot find where its behavior really gets determined.
|
Fri 07 Aug 2015 04:09:52 PM UTC, comment #3:
I tried replacing the scripts in:
scripts/plot (and its subfolders) with the
scripts from 4.0.0. This had no effect!
Since there were lots of changes I thought
that something would break, but it did not.
So, the problem is in code that is not in ../scripts/plot
Any ideas?
|
Fri 07 Aug 2015 10:45:41 AM UTC, comment #2:
The minor grid lines are correct in 4.0.0 release.
See attached badgrid_ok_4-0-0.pdf.
Right now, this is the only fully correct rendering.
So, it is appropriate to consider the current default
system as a regression from 4.0.0 release.
(file #34585)
|
Fri 07 Aug 2015 09:20:25 AM UTC, comment #1:
Here is a simple example of the problem:
This is better to use for diagnosis than
the complicated plot where I found the problem.
|
Fri 07 Aug 2015 01:09:52 AM UTC, original submission:
Until recently, use of grid minor and grid tight
produced the plot attached as goodgrid.pdf
Now, this combination produces the attached file badplot.pdf
Which shows that the minor grid lines for the origin
to first major grid line are missing.
If grid tight is not used the plot is correct.
Note that the previous plot was actually not correct either.
Only the xaxis minor grid lines were drawn. I had not
noticed this previously.
I will look at this in more detail when I have a chance.
But, if someone who has made recent changes to the plotting
functions can find what change has caused this it would help.
|