Interpreter initialisation is slow when symtable/usersyms contains long lists about asteval HOT 4 CLOSED

@damamaty That seems okay to me. Does "cpl mins" mean "couple of minutes"? Anyway, I don't quite see this as a serious problem.

from asteval.

Yes, sorry, "couple minutes". It can take longer depending on the list size.
Reported it because it was a reason of bad performance in our case. Can be avoided by e.g. using wrappers around list/dict with redefined repr, but from my point of view it is unexpected that the library does that.

from asteval.

@damamaty Huh? What does __repr__ have to do with it? Should I know what you expected to happen? As it turns out, I don't.

Why would anyone create a list with a billion elements and then expose that in a sandboxed interpreter?
Performance is just not a high priority here, but you may recall that there is this thing called numpy.

from asteval.

Closing this, I cannot reproduce. FWIW

>>> x = [1]*1_000_000_000     # takes a couple of seconds for me
>>> import asteval
>>> aeval1 = asteval.Interpreter({'x': x})  # much less than 1 sec
>>> aeval2 =  asteval.Interpreter()  # much less than 1 sec
>>> aeval2.symtable['x'] = x  # much less than 1 sec
>>> aeval2('len(x)')                 # much less than 1 sec   
1000000000

from asteval.