Adding too many cursors at once can cause significant hang.

[athompson673]

I mis-clicked and added 600+ cursors using the column cursor feature, which caused spyder to hang for quite a while (did not crash). Should we pick an upper limit and generate a warning for the user?

[athompson673]

Just did some quick testing..

I realized that I'm merging cursors between adding each one, which probably only needs to be called once at the end.

[ccordoba12]

Should we pick an upper limit and generate a warning for the user?

Yep, I think that's the right strategy. I propose to have a max of 100 cursors because that should be enough for most use cases and doesn't cause too much delay according to your plot.

If you manage to do it in the next week and a half, we could include it in 6.1.0.

[athompson673]

on the other hand... quickly fixing the merge problem speeds things up considerably. Maybe a warning at 1000, which the user can bypass?

a = np.array(
    [[100., 0.12635670002782717],
     [200., 0.8105566999875009],
     [300., 2.5749110000324436],
     [400., 5.758394000004046],
     [500., 10.942007500038017],
     [600., 18.42652859998634],
     [700., 29.01852299994789],
     [800., 43.83678370004054],
     [900., 57.852217599982396],
     [1000., 79.72451609995915]]
)

b = np.array(
    [[500, 0.07561930001247674],
     [1000, 0.2444275000016205],
     [1500, 0.5651400000206195],
     [2000, 0.9679742999724112],
     [2500, 1.4596941000199877],
     [3000, 1.9943268000497483],
     [3500, 3.0266019999980927],
     [4000, 3.8106580000021495],
     [4500, 4.5058330000028946]]
)

plt.plot(*a.T, label="Current")
plt.plot(*b.T, label="Merge once")

[ccordoba12]

Maybe a warning at 1000, which the user can bypass?

Sure, sounds good to me.