search

NetLogo / Tortoise

Compiler and runtime engine for NetLogo models that runs in JavaScript 🐢
https://netlogoweb.org
Other
56 stars 27 forks source link

Get language tests working with existing features #60

Closed TheBizzle closed 10 years ago

TheBizzle commented 10 years ago

TortoiseFixture marks many of the language tests that ought to already be passing as "freebies" that are, for currently-unknown reason, not yet passing.

The tests listed as "to be investigated" ought to be investigated and classified. Any test that is ultimately not marked "obscure" or given some other low-priority classification ought to get taken care of within the near future.

SethTisue commented 10 years ago

some background

3:10 PM <SethTisue> this business of getting individual models working is going to result in detours where we investigate problems and they turn out to be due to things the language tests would have caught
3:11 PM <SethTisue> it's great to have an example model like Climate Change in order to drive the front-end effort
3:12 PM <SethTisue> but on the compiler/engine front, at this point I think we've done enough models and we should be concentrating on getting language tests passing. if the language tests pass, the models will follow.
3:12 PM <SethTisue> not saying any wrong choices have been made, just looking ahead to what the rest of 2014 holds.
3:24 PM <TheBizzle> SethTisue, I'm not sure I understand the situation.  Is it not true that we have all of the language tests passing for all of the features that we've already implemented?
3:26 PM <SethTisue> sort of
3:27 PM <SethTisue> hardly
3:27 PM <SethTisue> see TortoiseFixture.scala. many tests are marked as known to fail
3:28 PM <SethTisue> they don't fail the build
3:28 PM <SethTisue> but they fail.
3:28 PM <SethTisue> a couple are marked as "egregious", about 8 are marked as "significant", and about 25 are marked as "to be investigated", of unknown severity
3:29 PM <TheBizzle> Ohhh, interesting
3:29 PM <SethTisue> furthermore...
3:29 PM <SethTisue> "passing for all of the features that we've already implemented"
3:29 PM <SethTisue> nearly any language test exercises multiple features
3:29 PM <SethTisue> even a single unsupported feature means the entire test is skipped
3:29 PM <TheBizzle> Ah, right
TheBizzle commented 10 years ago

The situation here is not perfect, but is now pretty acceptable.

Relevant commits: 815f949467b3d4863f76fd0e37fdee6628f79f4e, 7da36d2d9440d85e1b2a5c778de160f5d4664c45, 167daa9c09d606f639430ee077aad8456d07b39e, e707a19d41374df14ebb2b2020d3e0cc59b3ef37, 19a5a936cbf97d418ee7a5af9bf2b72a7c6b308e, c99e17b53a00abbf9364c5843833f89f2b920cfb, 05238302f66872e4390d87d591c2f532fb285fe6, 46d9c9edc32288a595164907aa69063e3b1f91b3, 28635623884c251b7c32c7f2e2b72195fef4fd91, 25a95074fbe7079b965a7383dde2c75c4699a8a2, b9ac78a77449c4e3efc7f024a83b433d9959764d, b35cd3af21a9d2f68f6b999f4d2b3248167e55da, 20f83de295bd765df5d6411cd46b012cf7b304f4, 4e37054b406f6d09ea4f1d0d47dbc45d5567332e, 36061719cb95e630d5b992dddb77e3f5b826f083, 4ff145ade0a52adbc4e8a1f2eb0d4363be6eddde, 76bce4eb739b484ee0f6c27b8583520a548f6f57, 0108c5a3ed95a7389aa66167c5aa74e3d818bffd, 4725b4308a5c515ac4bc17ac0c4369ea6e0fe2cf, 700e917223f2811166e5b70ce0522ff4f2836582, 173bc4b6077e111fd47a0d8d488cb7095c73c647, 42a0f2d17a96b6059248d9c3f8ca100f33def366, dcd50fc62bc9d79d96cc534ef6c55d7e86889005, 70052cfb1b8212d2ca7ab7944696e8b24b65aff6, 162fec785277f539372fc87a4e775dd99009c8ee