Should hlines and vlines take part in the style cycle?

[Mc-Zen]

Some plotting functions like plot, scatter, and bar take part in the style cycle which means that consecutive calls to such plotting functions will give them different colors/marks/etc. depending on the value of diagram.cycle.

For other functions like colormesh, cycles do not make any sense. Currently, hlines and vlines are among them. In my opinion this makes sense because these two functions are more intended to mark certain coordinates in the plot rather than for visualizing actual data.

Matplotlib are having the same discussion actually, see matplotlib/matplotlib#14222 (comment).

Not taking part in the style cycle also means that these plot functions do not advance the style counter, i.e., they can be put in between calls to plot, scatter etc. without changing the style assigned to those. Or rather they should not advance the style counter, because currently they do in Lilaq (but I want to change that). This concept is a good feature I think.

[eltos]

In my opinion this makes sense because these two functions are more intended to mark certain coordinates in the plot rather than for visualizing actual data.

Totally agree that these should not be part of style cycling, not advance it, and their default black stroke is perfect. 👌

In many years of matplotlib usage I've never had a use case where them being part of the cycle would have been useful, and I always had to manually change their color to black (or something else) because of matplotlib making them blue by default.

Plus, it's quite frequent that I used one or two such lines in different plots (with different numbers of traces) to give the reader some common reference points. If they were part of the cycle these reference lines would appear in different colours which would be very confusing and contradict the idea of reference marks.

[Mc-Zen]

Thank you for sharing your experience!