Sun 15 Oct 2017 02:03:23 AM UTC, comment #5:
Correct. I searched old changesets I have for gnuplot on my computer and there's a sizable patch for adding the feature of controlling line color of ticks. So, this was a gnuplot issue, not an Octave bug. I was hoping to resurrect that patch and update it for gnuplot's current repository. However, gnuplot is in a transition period at the moment, and I'm going to wait until things are stable there. My estimate would be a month or two, at which point I'll post something here to make this bug report active once again.
|
Fri 06 Oct 2017 08:54:03 PM UTC, comment #3:
gnuplot chooses the tick color to match the border color. That would be fine, if not for the fact that the plot can have different axes with different colors (and there is only a single border color).
Hmm, I'm starting to have flashbacks on this one. This may be something a created a gnuplot patch for, but the patch never got implemented. Let me investigate the gnuplot bug-report page.
|
Fri 06 Oct 2017 08:23:12 PM UTC, comment #2:
Since this is a regression, I've searched through all the patches I have here (I always copy my patches to a local subdirectory before posting). The closest related patch (not sure it was ever pushed though) seems to be the following from djs2017mar10
That date would be after the July 2016 timeframe. Notice how the re is one fprintf() removed that uses colorspec, but the one that is retained doesn't have the color spec. Again, not sure if this was ever applied, or exactly what happened. But this would be a good starting point for investigating where the issue might be.
I'll look at this and come up with a changeset.
|