Use tasty-golden for golden tests and ditch vendored hspec-expectations package. #125
Conversation
Not sure where this is coming from, but we don't want the builders to get stuck.
|
|
||
| source-repository-package | ||
| type: git | ||
| location: https://github.com/rewinfrey/hspec-expectations-pretty-diff |
There was a problem hiding this comment.
Down to only two vendored packages left (the hard ones, unfortunately).
| testsForLanguage language = do | ||
| let dir = "test/fixtures" </> language </> "corpus" | ||
| let items = unsafePerformIO (examples dir) | ||
| localOption (mkTimeout 3000000) $ testGroup language $ fmap testForExample items |
There was a problem hiding this comment.
These tests sometimes get held up; I’m not sure why (I think it’s a race condition w/r/t file I/O), but this timeout remedies it so the builders won’t get stuck for 10 minutes before it happens. (Worth filing a bug about, I think.)
| -- tests and find this output hard to read, install the `jd` CLI tool | ||
| -- (https://github.com/josephburnett/jd), which will print a detailed | ||
| -- summary of the differences between these JSON files. | ||
| renderDiff :: String -> String -> [String] |
There was a problem hiding this comment.
The major compromise associated with giving up hspec-expectations-pretty-diff is that the quality of the diffs went down. However, tasty-golden lets us provide our own diffing invocation, which means we can retain the excellent output.
| let nativeSpecs = tests session | ||
| asTastySpecs <- Tasty.testSpecs $ legacySpecs session | ||
| let allSpecs = nativeSpecs <> asTastySpecs | ||
| pure . Tasty.localOption Tasty.Success $ testGroup "semantic" allSpecs |
There was a problem hiding this comment.
Tasty has no concept of pending tests, so we mark our pending tests as successes.
| {-(KeyValue | ||
| {-(TextElement)-} | ||
| {-(Float)-})-} | ||
| {-(KeyValue |
There was a problem hiding this comment.
Now that we don’t use verbatim, we care about whitespace here.
| @@ -1,36 +1,53 @@ | |||
| {-# LANGUAGE ImplicitParams, LambdaCase, NamedFieldPuns #-} | |||
There was a problem hiding this comment.
Countdown to Rob requesting NoImplicitParams in 3… 2… 1… 😆
There was a problem hiding this comment.
(I do think implicit params are a better solution than plumbing parameters everywhere or Reader + umpteen calls to pure. But we should arguably do the let ?session in the top-level Spec.hs file.)
There was a problem hiding this comment.
u wot m8
The purist in me would prefer a Reader effect, but I see the appeal. Tell you what: you maintain it and we’re good 😉
There was a problem hiding this comment.
(I kid, I kid. I’m cool with this 👍)
| other-modules: Paths_semantic | ||
| build-depends: base >= 4.12 && < 5 | ||
| , ansi-terminal ^>= 0.8.2 | ||
| , ansi-terminal >= 0.8.2 && <1 |
There was a problem hiding this comment.
Not technically needed, but ansi-terminal just hit 0.9.0, and the old hspec-expectations-pretty-diff was complaining about it.
| import: haskell, dependencies, executable-flags | ||
| type: exitcode-stdio-1.0 | ||
| hs-source-dirs: test | ||
| ghc-options: -Wunused-imports |
There was a problem hiding this comment.
Can we apply this in a common stanza?
There was a problem hiding this comment.
Yup. We should really just turn on -Wall everywhere, but I’ll file a separate issue about that.
There was a problem hiding this comment.
I dig it; the thing fused-effects does is working pretty nicely.
| @@ -1,36 +1,53 @@ | |||
| {-# LANGUAGE ImplicitParams, LambdaCase, NamedFieldPuns #-} | |||
There was a problem hiding this comment.
u wot m8
The purist in me would prefer a Reader effect, but I see the appeal. Tell you what: you maintain it and we’re good 😉
| {-(KeyValue | ||
| {-(TextElement)-} | ||
| {-(Float)-})-} | ||
| {-(KeyValue |
This patch:
hspec-expectations-pretty-diff, which partially unblocks Publish to Hackage/Stackage. #16.unsafePerformIOin places where it is somewhat unavoidable (I defendunsafePerformIO’s use in tests)tasty-goldenlets us provide our own diff command, and attempts to usejdfor JSON comparisons.hspec-expectationsFixes #120.