search

Dervall / Piglet

An easier to use parsing and lexing tool that is configurable in fluent code without a pre-build step.
binarysculpting.com
MIT License
91 stars 11 forks source link

Provide lexer state in reduce function callback #53

Open tgjones opened 11 years ago

tgjones commented 11 years ago

I am building a compiler in which semantic analysis will be done a separate pass, after the AST has been built. But in order to report line numbers for semantic errors, I need to associate AST nodes with their corresponding token position.

The token position is currently only provided (to SetErrorFunction) if a parse error occurs. But perhaps there could be another overload of SetReduceFunction, where the callback is given the lexer state as well as the list of objects to reduce?

Dervall commented 11 years ago

Thanks for reporting a possible improvement!

Sure, this can be done. It will require a bit of thought but it's doable.

I'll see if I can get some time with this in the coming week, unless you want to take a stab at it yourself.

tgjones commented 11 years ago

Thanks! I did have a go at implementing it, but I didn't get very far. It was a little while ago now, but from what I remember I could maybe get the line number, but not the column number of the current token. Ideally both would be available to the callback.

I will try to find time to take another stab at it.

Dervall commented 11 years ago

There's been a lot of life between me and programming in my spare time unfortunately. :) It's a great feature, and I should implement it, but there is not likely to be time to do so currently for (starting) a family reasons :-)