Comment author: @lpw25

This is intended behavior. Those things live in different namespaces.

Comment author: @lpw25

(See http://caml.inria.fr/pub/docs/manual-ocaml/names.html)

Comment author: @bobzhang

is this not discussible? I am not saying this is a bug..

Comment author: @lpw25

Well it would be a hugely backwards incompatible change, and the only benefit put forward is that it would make your code-gen easier to write, so it doesn't seem like it is a particularly viable change.

Comment author: @bobzhang

We can fix it on our side(either simply reject such programs or do the name mangling). But I am very surprised to read such programs:

let v  = new M.c
let a = M.c + 3

I think spit out a warning is reasonable feature request

Comment author: @lpw25

I think spit out a warning is reasonable feature request

I agree. I'll edit the description and mark this issue unresolved.

Comment author: @lpw25

(Although I'm not personally in favour of the warning)

Comment author: @alainfrisch

It's rather common to have values and types with the same name. Hongbo: Would you include this case in the warning? I understand that this case does not impact Bucklescript (no runtime value corresponding to types), but it would be rather incoherent to warn in other cases but not this one.

Comment author: @alainfrisch

either simply reject such programs or do the name mangling

You could also group extensible constructors and classes in sub-objects of the generated Javascript module (they are much less likely to be required by JS code than regular values and modules).

Comment author: @bobzhang

currently we take the easy route: spit out a compiler error when detecting two runtime components exported with the same name having the same name(this happens very rare in practice and probably should not pass review).
Yes, it is common to have value and type as same name, but as you said, it does not impact BuckleScript

Comment author: @bobzhang

if we have a warning, it could provide better error message(we can turn it into an error on our side)

This issue has been open one year with no activity. Consequently, it is being marked with the "stale" label. What this means is that the issue will be automatically closed in 30 days unless more comments are added or the "stale" label is removed. Comments that provide new information on the issue are especially welcome: is it still reproducible? did it appear in other contexts? how critical is it? etc.

I don't think there is an interest towards this outside of the bucklescript compiler team. I suggest that we close this ticket for now, and you can make a PR implementing a (disabled by default) warning when you feel ready.