Hey 
,
I am trying to check if a variable exists - but as far as i could imagine doing it, i don’t see a way to do it.
Say i’d like to detect if “a” is being defined declared - i thought of using something like if a ... which of course fails.
So how actually check if an variable exists?
You can check it yourself, e.g.
#let a
a is declared but not defined. The representation of a is
repr(a) // none
none. So checking for none should be sufficient:
#{a == none} // true
Since if requires a boolean condition to be met, we can check:
#if a == none [Yes] else [No] // Yes
On the other hand. a is not none when it is defined:
#let a = 5 // declared and defined
#{a == none} // false
Dear @gpr - thank you very much for your expressive example! I am so sorry i didn’t make clear that i actually try to detect if a variable is already declared (not defined). 
Sorry, maybe I misunderstood something, but this is how you would check whether a variable is defined:
a != none // checks whether a is defined
Evaluates to true if a is defined, e.g.
#{if a != none [a is defined here continue...]}
As far as I know there is no syntax like if a
This will error if a is not in scope. I think what @Florian is trying to do is to check whether it is in scope.
It’s not possible to check this (at least I’m not aware of any way). There is a small exception: If the bindings come from an imported module, you can convert the whole module to a dictionary with the dictionary constructor and check whether it’s contained. For the standard library definitions that module would be std. That lets you solve a subproblem of your question.
All that said: Even if there was a way, it would be quite unorthodox to perform such a check. What are you actually trying to achieve? Maybe there is a better way.
Thanks a lot @gpr and @laurmaedje.
I am recently thinking of and writing modules for different use cases.
And of course i am aware of a) the immutable nature of the processing pipeline diverging to the best result and b) i usually would recommend to put a lot of the data preparing to a preprocessing outside of Typst.
But i experiment a lot with different ways to use Typst as the interface for more or less complex tasks which leads to a annoyingly enjoyable experience creating user interfaces in the Typst scripting language.
In that regard i thought about using state variables behind the scenes, to allow modules to implicitly use the state of other modules.
Don’t get me wrong i think that’s a bit of an antipattern since it could raise issues but in terms of exploring what one could achieve i was interested to find out if it’s possible.