Add Mod.apply_unit

[craff]

Solves #610. Just modified ast_helperlite.mli? And tested on my code.

[patricoferris]

Hi @craff, thanks for this adding this.

Is there any reason you are not using the functions supplied by Ast_builder, in this case I think it would bepmod_apply_unit?

[craff]

Patrick Ferris ***@***.***> writes:

* patricoferris left a comment (ocaml-ppx/ppxlib#611) Hi @craff, thanks for this adding this. Is there any reason you are not using the functions supplied by Ast_builder, in this case I think it would bepmod_apply_unit?

Hello, It is what I use to define Mod.apply_unit. I think le ast_helper_lite is very nice as it evolves less than the parsetree itself. Say this, I think an alternative would be to have Mod.apply checking if the structure is empty ? If you think it is better I could modify my PR. Cheers, Christophe

…

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.

-- Christophe Raffalli web: https://raffalli.eu

[patricoferris]

Thanks for the response @craff -- the reason I ask is we have just announced a roadmap that includes the (slow) deprecation for Ast_helper and it would be great to hear your thoughts on that.

[craff]

Patrick Ferris ***@***.***> writes:

* patricoferris left a comment (ocaml-ppx/ppxlib#611) Thanks for the response @craff -- the reason I ask is we have just announced a roadmap that includes the (slow) deprecation for Ast_helper and it would be great to hear your thoughts on that.

I think Ast_helper is nice, because it offers a way to make your code easier to maintain through default value for parameter or keeping old constructor. If there way a similar approach to destructors it would be nice. As far as I am concern I tend to do as much as I can using metaquot and ast_helper and almost never need modification of my code. If Mod.apply where modified to call Pmod_apply_unit when neede, I would not have has to modify pacomb. On the other hand if metaquot were complete (I mean a syntax entry point for every type in the ast both for expressions and pattern). Then I think I would only use that, and this would be the ultimate solution. Though I have not look recently if metaquot was extended. Cheers, Christophe

…

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.

-- Christophe Raffalli web: https://raffalli.eu