Type checks inconsistent when dealing with a -> b function types

[radeusgd]

from Standard.Base import all

t1 (f : Function) =
    f "x"

t2 (f : (Text -> Any)) =
    f "x"

t3 (f : (Integer | Function)) =
    f "x"

t4 (f : (Integer | (Text -> Any))) =
    f "x"

main =
    f x = "|" + x + "|"
    IO.println (t1 f)
    IO.println (t2 f)
    IO.println (t3 f)
    IO.println (t4 f)

Running the example above yields:

|x|
|x|
|x|
Execution finished with an error: Type error: expected `f` to be Integer, but got Function.
        at <enso> inconsistent-function-type.t4(inconsistent-function-type.enso:12-13)
        at <enso> inconsistent-function-type.main<arg-1>(inconsistent-function-type.enso:20:17-20)
        at <enso> inconsistent-function-type.main(inconsistent-function-type.enso:20:5-21)

Whereas we'd expect all 4 cases to behave the same way.

Weirdly, if the type is 'alone' - both Function and Text -> Any behave the same and accept a function input.

However, once we construct a sum type using |, for some weird reason the Text -> Any seems as if it disappears.

[JaroslavTulach]

However, once we construct a sum type using |, for some weird reason the Text -> Any seems as if it disappears.

Text->Any actually disappears all the time. Just in case of t4 the Integer is left in the check. While in case of t2 there is no type check at all. Anyway #7298 fixes it all.

[enso-bot]

Jaroslav Tulach reports a new STANDUP for yesterday (2023-07-15):

Progress: - fixes #7289 ready for review: #7298

• emails & reviews
• review regex: Add Fallback to Prim_Text_Helper.compile_regex; accept Regex in Text.parse_to_table #7297 (review)
• reviewing Radek's benchmark effort: Add benchmarks comparing performance of Table operations 'vectorized' in Java vs performed in Enso #7270 (review)
• cleaning up code base integrated Avoiding dependency on EnsoContext from Module and ModuleScope #7277
• more code cleanup refactoring Compiler <-> EnsoContext dependency: Avoid Compiler compile time dependency on EnsoContext #7299 It should be finished by 2023-07-17.

Next Day: Bugfixes, cleanup & performance

[enso-bot]

Jaroslav Tulach reports a new STANDUP for yesterday (2023-07-16):

Progress: - bugfix CompilationStage: https://github.com/enso-org/enso/compare/392a8135862c029e8ee3be72d75514980d4644d9..fd0a9e70a1bc1bbda03756395bd8a8f1250d0f38

• hiding unsafe methods: c073c1b
• Shuffling CompilerContext It should be finished by 2023-07-17.

Next Day: Bugfixes, cleanup & performance

GitHub

Comparing 392a8135862c029e8ee3be72d75514980d4644d9..fd0a9e70a1bc1bbda03756395bd8a8f1250d0f38 · enso-org/enso

Hybrid visual and textual functional programming. Contribute to enso-org/enso development by creating an account on GitHub.

[enso-bot]

Jaroslav Tulach reports a new STANDUP for yesterday (2023-07-17):

Progress: - bookclub on top of runProjectManagerDistribution: #7181 - integratd

• runners: https://discord.com/channels/401396655599124480/1129451621093474335/1130378255799697418
• merged compiler changes: Avoid Compiler compile time dependency on EnsoContext #7299
• bigger rewrite of resource manager ResourceManagerTest to verify resources are cleaned on ctx.close() #7302
• types checks integrated: Type checks inconsistent when dealing with a -> b function types #7298
• emails, comments in issues
• standup meeting, meeting Dmitry. Wojciech
• thinking about Table performance & other emails It should be finished by 2023-07-17.

Next Day: Bugfixes, cleanup & performance

Discord

Discord - A New Way to Chat with Friends & Communities

Discord is the easiest way to communicate over voice, video, and text. Chat, hang out, and stay close with your friends and communities.