accurate position of syntax error when parsing

[dragoncoder047]

When I call lips.parse() and it gives me a syntax error there seems to be no way to see at what position in the source the parse error occurred, cause the comments and whitespace get thrown out.

I also tried the "parse with metadata" option you shared earlier in #414 (comment), but it gave the same result (no __line__ or __col__ appeared).

is there a better way to find the exact index of a syntax error?

[jcubic]

It may not be as simple as adding the metadata.

[jcubic]

Actually, I just checked, and all parse errors came from the parser that have info about the line and col. I will see if I can add this information to the error when meta is used.

[jcubic]

Added the same metadata to the parse errors.

[dragoncoder047]

Hmmm. I updated my html files to point to https://cdn.jsdelivr.net/gh/LIPS-scheme/lips@b372c7e4f6a1963788192612e34170a56976dca4/dist/lips.js which is this plus the #417 fix, and am using this function to parse

/**
 * @param {string} source
 * @returns {Promise<lips.Pair[]>}
 */
export async function parseWithMetadata(source) {
    const parser = new lips.Parser({ env: lips.env, meta: true });
    parser.prepare(source);
    return await lips.parse(parser);
}

but am not getting __line__ and __col__ on the errors thrown. Must be a browser caching problem, will update later...

[jcubic]

It works, I've put your code into CodPen: https://codepen.io/jcubic/pen/MYgexza?editors=0012

[dragoncoder047]

Now that I've force-reloaded many times, the cache has cleared and I am getting line/columns on some of the syntax errors now.

However I am still not getting position info for the "Parser: expected parenthesis but eof found" error. Indeed, when you open the codepen you shared above and change the ")" to "(", the position information all shows up, but is all undefined.

[jcubic]

The problem was that lexer was returning eof object instead of token with metadata, when it reaches the end of the input. But the last token is still accessible from parser.__lexer__.__token__.

[jcubic]

I've updated the demo with the latest version https://codepen.io/jcubic/pen/MYgexza?editors=0012

[jcubic]

Added line number to all error messages, when using higher level lips.Interpreter interface.