Improve toml-test integration #11
Conversation
Make the scripts executable and add hashbang; this way it can be run a
bit easier:
% toml-test -int-as-float ./toml-test-parse.js | tail -n3
toml-test v2023-10-18 [./toml-test-parse.js]: using embedded tests
valid tests: 163 passed, 3 failed
invalid tests: 342 passed, 13 failed
% toml-test ./toml-test-encode.js -encoder | tail -n2
toml-test v2023-10-18 [./toml-test-encode.js]: using embedded tests
encoder tests: 157 passed, 9 failed
Also add a run-toml-test.bash script to the root, with the correct flags
to run the tests and skipping known failures; this way it's easy to test
for regressions.
I just added the `-int-as-float` flag to toml-test, so that you don't
get loads of "failed" tests on this. In my reading of the specification
("64-bit signed integers (from −2^63 to 2^63−1) should be accepted")
it's not actually *required* to support the full int64 range. I'll send
a PR to clarify the spec on this later.
There was a problem hiding this comment.
Thank you for the contribution!
Yeah, the integer limit is not marked as required by the spec and uses the term "should"; the only enforced requirement is to properly reject integers if they cannot be represented which I do.
I also found a few tests for the encoder fail due to expectations of sub-millisecond precision in the output, and I had to modify my un-tag method to handle +Inf with an uppercase I - IMHO it should be lowercase everywhere for consistency with the other tests and the way it is written in TOML
|
The I need to look at the other failures; I just added smol-toml to https://arp242.github.io/toml-test-matrix; one of the reason I made that is so I can have some insight what failures happen with different implementations, and fix/improve toml-test as needed. I'll get round to it ... eventually. |
Make the scripts executable and add hashbang; this way it can be run a bit easier:
Also add a run-toml-test.bash script to the root, with the correct flags to run the tests and skipping known failures; this way it's easy to test for regressions.
I just added the
-int-as-floatflag to toml-test, so that you don't get loads of "failed" tests on this. In my reading of the specification ("64-bit signed integers (from −2^63 to 2^63−1) should be accepted") it's not actually required to support the full int64 range. I'll send a PR to clarify the spec on this later.