Autocomplete hinders definition of userfunctions

Helder Correia repo owner

Not specific to user functions. I know about this annoyance, but I'm also not sure what the best behavior would be. Open for feedback ( @Teyut @fk @thadrien ). I'm not sure, but there may be a separate issue for the user function argument completion already.

2016-01-20T12:23:05+00:00

Pol Welter reporter

Can you expand on which other features provoke this behaviour? I was only aware of user function arguments.

(Also note that my PR includes a solution to this; I only posted this issue now that I wrote a complete write-up of what we did in the PR.)

2016-01-20T12:44:15+00:00

Helder Correia repo owner

I meant that if you simpy write "5 * x<ENTER>" it will have the same behavior, not specific to user functions. Unless I misunderstood what your point. Are you referring to pull request #18 ?

2016-01-21T16:23:14+00:00

Pol Welter reporter

Your example is not really an issue. If the variable x is not defined, then summoning the autocompleter will indeed spit out xor. If x is properly defined though, it will suggest that one instead.

Yes, I am referring to PR #18. I put a note about it on the bottom on the description.

2016-01-21T16:37:35+00:00

Tey'

If x is defined as a user variable or a user function, typing 5 * x<ENTER> will have the correct behavior (x not being replaced if it's a user variable, or else, x being replaced by the user function). Otherwise, the auto-completion will replace x by xor() which is also the expected behavior as x does not exist anyway.

I believe the only use case when this is not the correct behavior is the one given by Pol, when defining a user function and x is one of the argument of that function. Pol fixed it at this line so that the autocompletion algorithm is aware of the user function argument names when defining a user function. I can't check it work yet but the code looks correct.