Mon 22 Aug 2016 01:07:12 AM UTC, comment #10:
Minor grid is now drawn correctly at major locations. Closing report.
|
Sat 20 Aug 2016 01:29:29 PM UTC, comment #9:
I've pushed the part which draws the major grid if minor grid is on here:
http://hg.savannah.gnu.org/hgweb/octave/rev/9deb86bb5632
But not the demos and tests. I think demos shouldn't contain referenced to bugs and the tests in your patch should go to the commit which fixes bug #48533.
|
Sat 16 Jul 2016 05:17:28 PM UTC, comment #8:
Dan, attached please find the updated changeset that also adds the demos and tests to "grid.m".
I also filed the new bug #48533 to track the inconsistent result of "grid".
(file #37914)
|
Sat 16 Jul 2016 03:58:16 PM UTC, comment #7:
Sorry, Dan, this was a misunderstanding on my side. I thought you wanted some manual tests.
I will file another bug report on the issues in the first verbatim-block of comment #5 and add the demos and some generalized (failing) tests in "grid.m" here as soon as I have it ready.
|
Sat 16 Jul 2016 07:07:43 AM UTC, comment #6:
The demos look good, illustrating the various combinations. I'm having a bit of trouble following them with the related bugs in the mix, but it appears that you placed some assert() tests in there to show that the result is not as expected. So, if these are run during the check phase they'll come up as failed tests and will motivate someone to fix them.
The demos should be placed within the file having the name of the test. So, for "demo grid", look to the end of the file "grid.m" and you should see how to format the demos.
Rollback your changeset, add the demos to grid.m and update you commit comments with something like:
Thanks.
|
Sat 16 Jul 2016 06:37:56 AM UTC, comment #5:
I think "grid" doesn't do the right thing as it also changes the other grid which can lead to inconsistent results. In the following example, the minor grid should not be changed at all. The results should not change after calling "grid on" the second time:
But that is imho another bug on which I want to work soon. This is why I refrain from using "grid" in the examples for now:
Should display minor grid lines also on major ticks.
Should display minor grid lines also on major ticks in y. Major grid in x should display as expected. Hits bug #45710. A patch is posted there.
Should display minor grid lines also on major ticks. The minor grid should adapt to the xticks. Hits bug #45850. A patch is posted there.
Please, let me know if you need more examples.
|
Sat 16 Jul 2016 05:23:55 AM UTC, comment #4:
OK, it sounds maybe more complicated then. Perhaps one or two more demo examples with no major grid, and with custom major grid lines and minor grid on/off would be valuable. Type "demo grid" at the command line.
|
Sat 16 Jul 2016 05:06:15 AM UTC, comment #3:
Thank you, Dan, for reviewing this.
The disappearing minor grid you observe might be due to bug #45850. I posted a changeset there as well. At least, the minor grid doesn't disappear for me. But I also have a bunch of other changes in my test build.
|
Fri 15 Jul 2016 09:05:47 PM UTC, comment #2:
I had noticed this about a week ago but thought maybe that was the way it is supposed to be, but I agree it doesn't seem correct.
I like what you've done, as far as not placing both a major grid line and minor grid line one on top of the other.
I did wonder if there is a scenario in which the minor grid and major grid don't aline, but it doesn't appear that the minor grid is individually controllable. Also, for custom xgrid:
So I think it should be fine.
|
Fri 15 Jul 2016 06:35:54 PM UTC, comment #1:
Attached the changeset which should be applied on top of file #37872 in bug #48429.
Steps to test:
(file #37903)
|
Fri 15 Jul 2016 06:26:45 PM UTC, original submission:
When "xgrid" is "off", but "xminorgid" is "on", minor grid lines are also drawn at "xtick" values in Matlab. The same for y and z.
In Octave, no grid lines are drawn at the major tick values with these settings.
I am going to attach a changeset as soon as I have a bug number.
|