Other useful debugging output about pegjs

[dmajda]

I have thought about this too. If you like, you can try to implement it yourself -- modifying the code generator shouldn't be hard. Otherwise, I'll implement that myself soon.

[celer]

Ok, I may give this a shot, thanks.

[wolever]

This issue hasn't seen much love… But it's something I'd like to see.

For a start, I might propose (and would submit a patch for):

• Adding an options argument to buildParser:

      buildParser: function(grammar, options)

• If options.debug is true while code is being generated, options.log (default "console") is used to print the result of each successful match (untested, but my best guess for the approximate implementation):

      function parse_${name}() {
          ...
          if (${resultVal} !== null)
              ${options.log}.debug("matcher ${name} succeeded:", ${resultVal});
          return ${resultVal};
      }

(of course, that debug code would only be included if the the code generator was asked to generate debug code)

[knewter]

+1 on this. Having more debug information would be very helpful, especially for someone inexperienced at this sort of thing (this guy). Thanks for an awesome project btw.

[CMCDragonkai]

One thing I would like more information on, is whether the parser totally failed to parse (failed the first match) or semi-failed to parse. At this point, it seems that a parser can parse up to a certain point, and return output. How can you tell if the parser actually tried to fully parse the input, but failed at a certain point (syntax error)?

[hegemonic]

I would love to have a --trace option. I'm currently trying to figure out where a parser is failing, and it's been an awfully slow process. With a trace option, I could have found the answer very quickly.

[christofmuc]

+1 here. I totally totally love PEG.js and have used it for more than once thing already, but debugging a grammar is pita. Will try patching with the suggestion above.

[dmajda]

In the past few weeks I worked on this issue and added tracing support to PEG.js-generated parsers. Instead of implementing a simple static solution, I added a flexible API which allows users to pass their own tracer to a parser. This tracer receives events about parsing progress and can decide what to do with them (display, store, etc.). There is also a built-in default tracer (used when no tracer is supplied by the user) that logs events to the console in a simple format.

Details can be found in the commit introducing tracing support.

The API is currently somewhat experimental (which is why it isn’t documented properly yet) and I expect it will evolve over time as experience is gained.

The default tracer is also somewhat bare-bones. I hope that PEG.js user community will develop more sophisticated tracers over time and I’ll be able to integrate their best ideas into the default tracer.

Big thanks goes to @ganeshv who implemented initial tracing support for PEG.js which I used as a source of inspiration.

With that, I’m closing this issue. Requests for improving the API or the default tracer should go into separate issues.