TOML test matrix

This tests various parsers agains the toml-test suite.

"Decoder tests" check if valid TOML documents are processed correctly, "encoder tests" checks writing TOML files, and "invalid input" checks that the parser rejects invalid TOML documents. Implementations are marked as "compliant" (✅) if both the "valid" and "encoder" tests pass, otherwise they're marked as non-compliant (❌). Usually it's not a big deal to accept some technically invalid TOML; almost all failing "invalid" tests are uncommon edge cases.

The benchmarks are a simple decode of a 15k and 5M file. Results are displayed as "good", "slow", and "very slow" only. Doing a fully fair comparison is not straight-forward as some implementations will parse numbers, dates, etc. when reading the file, and others delay this parsing until the data is used. It's also not a contest on who can write the fastest parser – it's just to establish that performance is within reasonable bounds. Some parsers seem to be O(n²) or thereabouts and scale very poorly to large files.

		Name	Decoder tests	Encoder tests	Invalid input	15k file	5M file
✅	C++	toml++	pass: 182 fail: 0	pass: 182 fail: 0	pass: 371 fail: 0	good	good	output
✅	C++	toml11	pass: 182 fail: 0	(TODO)	pass: 371 fail: 0	good	slow	output
✅	C	toml-c	pass: 182 fail: 0	(doesn't include encoder)	pass: 350 fail: 21	good	slow	output
❌	C	tomlc99	pass: 175 fail: 7	(doesn't include encoder)	pass: 293 fail: 78	good	slow	output
❌	C#	tomlyn	pass: 179 fail: 3	(TODO)	pass: 366 fail: 5	slow	good	output
✅	Dart	toml.dart	pass: 182 fail: 0	pass: 182 fail: 0	pass: 332 fail: 39	good	good	output
❌	Fortran	toml-f	pass: 171 fail: 11	pass: 170 fail: 12	pass: 339 fail: 32	good	very slow	output
✅	Go	go-toml	pass: 182 fail: 0	pass: 182 fail: 0	pass: 371 fail: 0	good	slow	output
✅	Go	toml	pass: 182 fail: 0	pass: 182 fail: 0	pass: 356 fail: 15	good	good	output
❌	Guile	toml	pass: 177 fail: 5	pass: 178 fail: 4	pass: 356 fail: 15	(TODO)	(TODO)	output
❌	Haskell	toml-parser	pass: 181 fail: 1	pass: 182 fail: 0	pass: 369 fail: 2	good	good	output
❌	Haskell	toml-reader	pass: 177 fail: 2	(doesn't include encoder)	pass: 366 fail: 2	good	good	output
✅	JS	j-toml	pass: 182 fail: 0	pass: 182 fail: 0	pass: 357 fail: 14	good	good	output
❌	JS	smol-toml	pass: 181 fail: 0	pass: 174 fail: 7	pass: 357 fail: 14	(error)	(error)	output
✅	JS	toml-eslint-parser	pass: 182 fail: 0	(doesn't include encoder)	pass: 362 fail: 9	good	good	output
❌	Lisp	clop	pass: 179 fail: 3	(doesn't include encoder)	pass: 371 fail: 0	(TODO)	(TODO)	output
❌	OCAML	otoml	pass: 181 fail: 1	pass: 174 fail: 8	pass: 319 fail: 52	(TODO)	(TODO)	output
❌	Python	toml	pass: 164 fail: 18	pass: 168 fail: 14	pass: 298 fail: 73	good	good	output
✅	Python	tomli	pass: 182 fail: 0	pass: 182 fail: 0	pass: 371 fail: 0	good	good	output
✅	Python	tomlkit	pass: 182 fail: 0	pass: 182 fail: 0	pass: 365 fail: 6	slow	very slow	output
✅	Python	tomllib	pass: 182 fail: 0	(doesn't include encoder)	pass: 371 fail: 0	good	good	output
❌	Racket	toml-racket	pass: 0 fail: 182	pass: 175 fail: 7	pass: 0 fail: 371	(TODO)	(TODO)	output
❌	Ruby	perfect_toml	pass: 182 fail: 0	pass: 181 fail: 1	pass: 362 fail: 9	good	good	output
❌	Ruby	toml-rb	pass: 155 fail: 27	pass: 154 fail: 28	pass: 281 fail: 90	(TODO)	(TODO)	output
✅	Ruby	tomlrb	pass: 182 fail: 0	(doesn't include encoder)	pass: 351 fail: 20	good	slow	output
❌	Rust	basic-toml	pass: 161 fail: 21	(TODO)	pass: 367 fail: 4	good	good	output
❌	Rust	taplo	pass: 178 fail: 4	(TODO)	pass: 368 fail: 3	good	very slow	output
✅	Rust	toml	pass: 182 fail: 0	pass: 182 fail: 0	pass: 371 fail: 0	good	good	output
✅	Rust	toml_edit	pass: 182 fail: 0	pass: 182 fail: 0	pass: 371 fail: 0	good	good	output