Andreas Abel

Looks like Haskell-CI is currently broken because of - #416. I'll fix a push to `master` and then this PR can be rebased onto `master`.

I pushed the fix for #416 to master just now.

Sorry broke the stack build by accident, but now both builds should work. (Fixed in #417.)

The BNFC philosophy is that `token`s have some content you are interested in, whereas keywords like `"let"` have none. So the latter are not represented in the AST, and consequently...

> But the `pX` functions generated in the `Par` module return `Either String a` with the positions already converted to the `String`. I think the error reporting API also needs...

> One could boost the layout mechanism by allowing also declared tokens as layout keywords, i.e., > > ``` > layout TokLet; > layout stop TokIn; > ``` I am...

> Perhaps a BNFC flag which adds the full range of input tokens parsed to create this node to each CST node? Atm, we do not produce CSTs. We currently...

> It will be better to add a `--case-insensitive` option to `bnfc` besides adding a pragma to a token. I think case-insensitive keywords are rather a property of the language...

I think it would not be hard. However, it is quite easy to roll your own `Bool`, thus it is not so urgent. In contrast, the other literals are very...

I think we could make `Bool` a predefined category and use `True` and `False` for the rule names. The user then writes: ``` True. Bool := "true"; False. Bool :=...