I am tempted to create a new lexp `AttrTable` which would enable our properties/attributes to be inside our current lexp_context instead of creating a special context for them as it is now. I think it would make everything a lot more simple.
It sounds good, tho I have no idea what this AttrTable would look like. My guess is that you want to add a new constructor
AttrTable (<someargs>, e)
to the `lexp` type, such that `e` will be considered inside a context where the attributes have been modified according to <someargs>? But then I don't see how it'd appear in lexp_context.
So maybe instead you're thinking of
AttrTable (<someargs>)
which is then used inside a `let` and bound to some special variable, somewhat along the lines of the attribute_greater-then_table I suggested earlier?
Stefan
Afficher les réponses par date
Yes, I am thinking of bounding the AttrTable to a variable Making the table behave like a regular variable so we don't have to worry about scoping or modifying the context when it is not allowed since everything will handled automatically.
The special AttrTable will be replaced by a dummy during type erasure. Everything should simply work.
Yes, I am thinking of bounding the AttrTable to a variable Making the table behave like a regular variable so we don't have to worry about scoping or modifying the context when it is not allowed since everything will handled automatically.
Sounds good. So the attr-table would basically be a new built-in type with some new built-in primitives to add things to it and to consult it.
Stefan