dundalek / grammkit Goto Github PK

The idea is that it cleans up diagrams by removing all the _ and __ space matchers (and any other rules that start with _).

Would you be interested in a PR?

Currently the PEG and EBNF parser does not show very useful information on syntax error:

We could take inspiration from the Ohm parser which shows more useful mesage:

Just a suggestion, but GrammKit might be more useful if SVGs were generated as separate files from <output>.html. They could still be included in the output file, but then people who wanted to use them differently (say by linking to them, possibly out of order, in a README) could easily do so.

Hi there.

I'm trying to integrate Grammkit with AsciiDoctor, and I have run into a snag or two. Is there a forum where I can ask questions?

Hi! Many thanks for this excellent work!

Bug report and suggested fix:

• I observed an exception when diagramming the following PEG:

          start = num
          num = [0-9]+
          whtspace = ([ ]+)

• Exception observed
  Uncaught TypeError: Cannot read property 'usedBy' of undefined
• Origin
  Unused rule in given grammar: whtspace
• To reproduce
  • Navigate to Grammkit's online editor
  • Paste offending grammar snippet into text box.
  • Exception will appear in the browser's web console (Chromium, in my case).
• Suggested fix: Include typeof guards

  • In app/app.js::87

    // Currently
    {references[rule.name].usedBy.length > 0 && <div>
    
    // Replace by
    {typeof references[rule.name] !== "undefined" && references[rule.name].usedBy.length > 0 && <div>
    

  • In app/app.js::92

    // Currently
    {references[rule.name].references.length > 0 && <div>
    
    // Replace by
    {typeof references[rule.name] !== "undefined" && references[rule.name].references.length > 0 && <div>
    

Hope this helps! Keep up the good work!

Hi again.

I'm trying to run a couple of grammars through Grammkit, and they seem to be failing for some reason.

I've attached one of the file I was testing.

Oddly enough, I can't get it to work here either:

https://dundalek.com/grammkit/

It looks valid enough, so I'm a bit stuck.

tcl-statement ::= begin-transaction
                | set-transaction
                | savepoint
                | rollback-transaction
                | commit-transaction


begin-transaction ::= ( 'BEGIN' | 'START' ) ( 'WORK' | 'TRAN' | 'TRANSACTION' )
                      ( 'ISOLATION' 'LEVEL' 'READ' 'COMMITTED' )?

set-transaction ::= 'SET' 'TRANSACTION' 'ISOLATION' 'LEVEL' 'READ' 'COMMITTED'

savepoint ::= 'SAVEPOINT' savepointname

rollback-transaction ::= 'ROLLBACK' ( 'WORK' | 'TRAN' | 'TRANSACTION' )?
                       ( 'TO' 'SAVEPOINT' savepointname )?


commit-transaction ::= 'COMMIT' ( 'WORK' | 'TRAN' | 'TRANSACTION' )?

savepointname ::= identifier

Hello,

When I enter https://raw.githubusercontent.com/fuhrmanator/GIFT-grammar-PEG.js/master/GIFT.pegjs as the grammar on dundalek.com/GrammKit it gets morphed to https://cdn.rawgit.com/fuhrmanator/GIFT-grammar-PEG.js/master/GIFT.pegjs

Is this normal?

I was looking for a way to link to a RR diagram of my grammar from my readme (ideally even show the diagram there). I know it worked some years ago. Cheers!

Fixing the ohmjs dep is great, except that if I install from GH, I get the same problem 😆

Would you be kind enough to bump and push to NPM please?

Hi, I would like to know how to install it in window 10, show me this mistake:

In lib/rd-optimize-loops.js lodash is required, but package.json does not have the name lodash.

Hi,
I am running into an issue after upgrading from 0.3.1 to 0.6.3.
Most of my grammar still works, but GrammKit chokes on one particular rule.
I modified the example from README.md to reproduce the problem:

var grammkit = require('grammkit');
var parse = require('pegjs/lib/parser').parse;

var grammar = parse('Atom "_atom"\n  = !\'AND\' !\'OR\' !\'NOT\' [^\\t\\n\\r ()]+ { return text().trim();}');
console.log(grammkit.diagram(grammar.rules[0]));

This produces the following error:

…/node_modules/grammkit/lib/rd-optimize-loops.js:4
  if (node.type === 'Sequence') {
           ^

TypeError: Cannot read property 'type' of undefined
    at traverse (…/node_modules/grammkit/lib/rd-optimize-loops.js:4:12)
    at Array.map (<anonymous>)
    at traverse (…/node_modules/grammkit/lib/rd-optimize-loops.js:45:33)
    at Array.map (<anonymous>)
    at traverse (…/node_modules/grammkit/lib/rd-optimize-loops.js:45:33)
    at Array.map (<anonymous>)
    at traverse (…/node_modules/grammkit/lib/rd-optimize-loops.js:5:31)
    at Array.map (<anonymous>)
    at traverse (…/node_modules/grammkit/lib/rd-optimize-loops.js:45:33)
    at diagramRd (…/node_modules/grammkit/lib/diagram.js:6:10)

I'll try to dig a bit deeper, but maybe you already have an idea?

Cheers,
--lu

On your example site if you load the JavaScript grammar we have

ExpressionStatement
  = !("{" / FunctionToken) expression:Expression EOS {
      return {
        type:       "ExpressionStatement",
        expression: expression
      };
    }

and the diagram looks like

As far as I understand, that is an invalid railroad diagram.

Is it possible to specify the grammar file to be loaded using query parameter? It's easier to share links that way.

How about adding an option to diagram the entire syntax in terms of its start and terminals? Sometimes it's helpful to see the entire grammar all at once.

I'm trying to package GrammKit up for NixOS using node2nix, and while I can build GrammKit fine, when I try to run the executable I get this error:

[~/code/mine]traveller@nivian $ ./GrammKit/result/bin/grammkit GrammKit/examples/sparql.ebnf 
fs.js:640
  return binding.open(pathModule._makeLong(path), stringToFlags(flags), mode);
                 ^

Error: ENOENT: no such file or directory, open './lib/parse-ebnf.pegjs'
    at Error (native)
    at Object.fs.openSync (fs.js:640:18)
    at Object.fs.readFileSync (fs.js:508:33)
    at Object.<anonymous> (/nix/store/9srzg40jrjbrsn952j4lc9zcbsn2rxqp-node-grammkit-0.5.0/lib/node_modules/grammkit/cli.js:14:36)
    at Module._compile (module.js:570:32)
    at Object.Module._extensions..js (module.js:579:10)
    at Module.load (module.js:487:32)
    at tryModuleLoad (module.js:446:12)
    at Function.Module._load (module.js:438:3)
    at Module.runMain (module.js:604:10)

Is it possible that ./lib/parse-ebnf.pegjs needs to be mentioned somewhere in the package.json file? I think that's what node2nix uses to generate the Nix package.