I have a script which I usually run on 1 minute bars. Suppose I make a change to the script that causes compilation to fail, then I press Test, it tells me of the error and resets the bar period slider to 60 and the other sliders to their default values. So I fix the code and then I have to adjust the sliders again before pressing Test.

Does no one else find this exasperating?

A compilation failure should not reset the sliders, or am I doing it all wrong?

Sliders are not reset by a compilation failure. They are reset when you modify the code.

Nope. Code modification usually does not cause slider reset. Given how my teeth grind together each time the sliders are reset, I think I would have noticed. Besides, I have just tested adding a printf line to my script - the sliders did not move.

This test was realised using the latest beta version 1.62.1, downloaded today.

If not by code modifying, then your sliders were probably reset by something else, like selecting a different script? I don't think that sliders know about compilation errors.

My mistake. I need to be more specific. Compilation errors don't cause the sliders to reset, but some runtime errors do.

I'll report back when I find an example.

Here is a weird one.
I have t6 file containing data starting on the 7th September this year. I have a script that compiles ok, but when run gives the following error 055 binEURUSD 2016 (binEURUSD_2016.bar) history missing.

I move the bar period slider to 1, and press test. Same error, and the slider jumps back to 60.

I add BarPeriod = 1; to the script and press test. Hey presto, no error, the script works fine.

Question: why did it try running the script with period 60 when I had explicitly set the period slider to 1?

Similar question: why do I need to run a script once before I can set any custom sliders? Until I have successfully run the script the sliders are all called "SliderX" where X is 1, 2 or 3.