Would you help to compare my parser?

[LongTengDao]

Hello! Would you mind to add my parser to the compare table, after @iarna/toml@1.6.0, toml-j0.4@1.1.1, toml@2.3.3, @sgarciac/bombadil@2.0.0? My parser is https://GitHub.com/LongTengDao/j-toml/, it was published at https://www.npmjs.com/package/@ltd/j-toml.

[iarna]

I'll give it a go!

[iarna]

@LongTengDao The benchmarks are failing out pretty early with parse errors, for example:

key3 = -2E-2

Throws SyntaxError: Invalid Local Date instead of parsing that as the value -0.02.

I'll make a branch that includes your parser and you can try them out yourself!

I would encourage you to get your parser passing BurntSushi's tests (https://github.com/BurntSushi/toml-test) and/or mine (https://github.com/iarna/toml-spec-tests/tree/0.5.0) before trying to move into benchmarks.

[iarna]

It doesn't have to be passing everything for me to include it, but… there are some tests that it seems to infinite loop on, and I can't run it through its paces till you at least fix those bugs.

Specifically, this benchmark 0B-types-scalar-string-multiline-1079-chars just hangs with your parser. That's it trying to parse this file: https://github.com/iarna/iarna-toml/blob/add-ltd-j-toml/benchmark/0B-types-scalar-string-multiline-1079-chars.toml

The branch add-ltd-j-toml adds support to the benchmark suite. It also (temporarily) adds a new file called ltd-benchmark-per-file.js which will run the detailed suite against only your parser, so you can more easily see where the problems are.

[LongTengDao]

@iarna Really thanks for your help!

I fixed a data type classification bug. It treated -2E-2 literal with two - as a local date, like 1111-11-11.

I found the infinite-loop-like status comes from a very slow RegExp trap /^(?:[^\\]+|\\[^])*"""/, which should be /^(?:[^\\]|\\[^])*"""/. Just a few characters will make it run a few minutes, amazing.

Then I use the tests to check over my parser (v0.5.42), now it looks good. Wish you run a good result, thank you!

[iarna]

@LongTengDao I've updated both the spec compliance test results and the benchmarks with your parser. It still can't complete all the tests, but it's running well enough that I can include it. It's very impressively fast for large strings!

[LongTengDao]

@iarna Wow! Thank you!