Is "type" keyword accepted as identifier ?

[mingodad]

Trying to write an alternative parser for tl I found a possible bug on

tl/tl.tl

Line 46 in ef4fcda

typ: Type

we have as field of a record but on

tl/tl.tl

Line 54 in ef4fcda

type: Type

and

tl/tl.tl

Line 1201 in ef4fcda

type: Type

we have type as field name for another record.

This makes difficult to parse because type seems to be a keyword that also is valid inside a record for type declaration like here

tl/tl.tl

Line 10 in ef4fcda

type LoadFunction = function(...:any): any...

inside local record tl.

Is this intended or a bug ?

[hishamhm]

Yes, it is intended and not a bug. type is not a hard keyword, it is a soft keyword (i.e., context-sensitive). All Teal keyword additions are intended as soft keywords, to minimize compatibility issues with existing Lua libraries.

[mingodad]

Thanks for clarify it !
If so why not be consistent and change here

tl/tl.tl

Line 46 in ef4fcda

typ: Type

[hishamhm]

There's no need to, and sometimes the "inconsistency" makes it easier for me to spot which table is which. Also, I use typ in lots of places as locals (where I can't use type) and it's nicer to match these, but Result is more of a "public" interface.

Ultimately it's a matter of taste, of course, but I'm happy with the code as is.

[mingodad]

Thanks for reply !
I've got a kind of alternative parser for Teal and with it I did a crude initial conversion to what could be the LJS (https://github.com/mingodad/ljs) version of it and anyone interested can see it here https://gist.github.com/mingodad/1a5b0ed6c25232cb29db9e3fa7de61b1 , I also successfully converted tl.lua to LJS with the lua2ljs tool (https://github.com/mingodad/ljs/tree/master/lua2ljs) , I'm not sure if I will go forward with making a Tealjs version of Teal.