gh-127146: Update test skips for Emscripten 4.0.1

[hoodmane]

PR emscripten-core/emscripten#23306 fixed stating pipes and will be released as part of Emscripten 4.0.2. I removed skips for various other testes that were fixed in 4.0.1 or earlier.

• Issue: Emscripten: Get test suite passing #127146

[freakboy3742]

All makes sense; the only issue preventing merge is the RTD build failure... not sure what is going on there. I'll poke around and see what I can find out.

[zware]

Dumb question, but can the Emscripten version be detected so as to skip based on that version? I think there's prior art for this for various other platforms.

[freakboy3742]

Dumb question, but can the Emscripten version be detected so as to skip based on that version? I think there's prior art for this for various other platforms.

It almost certainly can be; but it will require the analog of platform.ios_ver()/platform.mac_ver() (et al) to expose the platform version identity. I imagine platform.emscripten_ver() (or wasm_ver, if there's an opportunity to share with WASI) will eventually exist; but it doesn't yet.

[hoodmane]

We have sys._emscripten_info.emscripten_version. So we could add the following to support:

defis_emscripten_lt(ver):
ifnotis_emscripten:
returnFalsever_tuple=tuple(int(x) forxinver.split("."))
returnsys._emscripten_info.emscripten_version<ver_tuple

and then put skipIf(is_emscripten_lt("4.0.1")).

[freakboy3742]

We'll eventually want to surface the version in platform for consistency, rather than the semi-private name on _sys. I'd also suggest using a tuple, rather than parsing the string at runtime,

We could also consider making this a feature of is_emscripten , so that skipIf(is_emscripten()) is "skip any version" and skipIf(is_emscripten((4,0,1)) does an additional version check.