> Le 28 sept. 2020 à 14:30, Adela Vais <[email protected]> a écrit :
>
> The D skeleton was not properly supporting them.
>
> * data/skeletons/d.m4, data/skeletons/lalr1.d: Fix it.
> * tests/calc.at: Check it.
Good catch, thanks Adela! Installed.
Note that these guys are critical in the case of C and C++ because of macros.
So if you define a token named EOF for instance, you'll get a clash with the
EOF macro. That's why it is important there to be able to "move away" from
dangerous names.
In D and Java, since the symbol and token kinds are "scoped", there's no risk
of collision, so it should not be so important. But it does give us some
consistency.
Cheers!
