lelio-brun / obelisk Goto Github PK

Hi,

The way < and > are handled in the generated HTML is a little bit... unusual. Using < and > would be better. What do you think ? This one is easy, so I'll make a PR.

Hi,

On the same file than the one in #10 ; on the line with the nonempty_list, here's what I get in the LaTeX output:

  & \grambar &\MATCH*{} \expr*{} nonempty\_list \zzzzzzEND*{}\\

Which visually gives something like: MATCH <expr> nonempty_list END.

Obviously, something is missing, it should be: MATCH <expr> nonempty_list(<match_case>) END.

The grammar

edn_assoc:
  | BEGIN_ASSOC; lp = list(pair(value, value)); RIGHT_BRACE { lp }
;

will result in

<edn_assoc> ::= BEGIN_ASSOC <value> <value>* RIGHT_BRACE

and I would expect

<edn_assoc> ::= BEGIN_ASSOC (<value> <value>)* RIGHT_BRACE

Please let me know if there something I'm missing!
Also, thank you for the great software 👍

See #2 (comment)

Hi !

I've have a .mly file that looks like:

let expr :=
...
| MATCH; m = expr; l = nonempty_list(match_case); END; { Match (Raw, m, l) }

I then run obelisk -i parser.mly and get:

<expr> ::= ...
         ...
         | MATCH <expr> nonempty_list(<match_case>) END

But I would expect the following:

<expr> ::= ...
         ...
         | MATCH <expr> <match_case>+ END

Am I missing something ? Note that, in the same file, list(X) is correctly recognized and translated to X*.

EDIT: I'm on OCaml 4.09.0, menhir 20200211 and obelisk 0.5.0.

Consider the following production definition:

endif: END | ENDIF { }

In Latex version, it will generate a \newcommand\endif{end\_if}, however, \newcommand definitions are not allowed to have names starting with end, or it will emit an error. The generated code for productions starting with end should be with \def instead.

Hi,

There's a new syntax available in menhir. In practice, it looks like this. I'm wondering if you have any plan to add support for this new syntax in obelisk ?

Cheers,

Unless I'm mistaken, it's only possible to generate a complete, standalone file, as of now. If so, it would be nice to add an option to generate only the grammar part, in order to include it in a larger document.

See #2 (comment)

See #2 (comment)

Production and group bars should be treated differently in printers.

The following file does not produce a very nice output for comma_sep and comma_sep1:

%public %inline comma_sep(X) :
  xs = separated_list(COMMA, X)
    { xs }

%public %inline comma_sep1(X) :
  xs = separated_nonempty_list(COMMA, X)
    { xs }

Would be nice if this tool could output an EBNF accepted by https://www.bottlecaps.de/rr/ui .

Ideally menhir would have an option to output this kind of naked grammar to facilitate the work like I did for bison, byacc, ... here https://github.com/mingodad/lalr-parser-test .

See discussion here stan-dev/stanc3#1142 .