plot_constraint violations(..)

[Peter230655]

I think, you were right, too complicated:

Mabybe better just like this:
The number of axes depends on the number of instance constraints.
If the user gives two axes, the method will tell how many are needed, unless two is correct.

If you plan to merge variable h before you change this, maybe it could be added:

If node_time_interval is a symbol, rather than a float, the times in the instance constraints should be given as
m * node_time_interval, where m is an integer or a float.

[moorepants]

Feel free to improve it in a new PR.

[Peter230655]

Feel free to improve it in a new PR.

The text I see in the API comes directly from the code?
I did not know this.

[moorepants]

I still don't like controlling figsize inside the function. We didn't do that before and the user simply had to make the figsize long enough (by trial and error) and pass that axes into the function.

[Peter230655]

I still don't like controlling figsize inside the function. We didn't do that before and the user simply had to make the figsize long enough (by trial and error) and pass that axes into the function.

I can remove it, of course. My only concern was that if there are many instance constraints (like in the drome example, there are 32), and the user does not give axes, the plots are not nice.
Should I remove in in my 'variable h' PR or make a new PR for the version you merged?

[moorepants]

Should I remove in in my 'variable h' PR or make a new PR for the version you merged?

I don't know what you mean precisely.

[Peter230655]

Sorry! What I meant was this:

I made an PR where I changed the plots, but did not touch the variable h printout. You merged this one. So If I make the change here, I will make a new PR

There is a PR pending, not merged, which in addition takes care of the variable h also. Here I'd simply push the changes.

[Peter230655]

I think, taken care off.