New Parser #42

Activity

1. EyLuismi commented on Nov 10, 2018

   

   EyLuismi

   on Nov 10, 2018

   Last edited by EyLuismi

   Contributor

   More actions

   Also the new way to add commands can be clunky at times and yield some issues with newer people wanting to contribute as there's no comments to what things are

   Yeah, I was thinking the same, what are your thoughts? First, I think maybe moving the commands definitions to another file would be a good Idea, and then in the readme I would explain with examples how you could add a new command arg type, or a command itself.

   Also, of course, more description in the code never hurts. I can prepare a PR about this.
2. EyLuismi commented on Nov 12, 2018

   

   EyLuismi

   on Nov 12, 2018

   Contributor

   More actions

   @Toyz In order to close this Issue, do you think that we would need to insert in the readme or myabe in some contribution guide, how to insert a new command? I mean, descriptions do help, but I think we could make a straightforward tutorial.

   Also now there is something about validators and expressions I want to dig in :P
3. Toyz commented on Nov 14, 2018

   

   Toyz

   on Nov 14, 2018

   ContributorAuthor

   More actions

   We probably show insert into the contributions file regrading it
4. rhdunn commented on Nov 15, 2018

   

   rhdunn

   on Nov 15, 2018

   More actions

   I was thinking of using something like the original part 1 implementation. My rationale for this is that logo is a relatively simple language that requires tokens to be space separated.

   Lexer

   1. Split on spaces
   2. If the text is a number parse it to a number object -- this can support float values
   3. Anything else (including "[" and "]") is a string object, and is thus a keyword

   There is nothing more complex in terms of tokens in logo.

   Parser

   1. Parse the tokens into a command array.
   2. Parse "[" ... "]" into a command array -- so the repeat command will take a command array as an argument.
   3. Each command is an object with a name and arguments (number tokens or a command array ("[" ... "]")), so parsing a command stops at a command name token.
   4. Return the top-level command array as a repeat 1 [ commands ] command.

   Interpreter

   1. A commands object where the keys are the command name, and the value is a function taking the argument array.
   2. An interpret helper function for looking up the command and calling it with the given arguments for each command in a command array.

   Thus, repeat could be implemented like:

   "repeat": function (arguments) {
       let n = arguments[0];
       while (n > 0) {
           interpret(arguments[1]);
           --n;
       }
   },