JS bool set invalid output with BSD

[beef331]

Akin to #17076 there is an issue with set[bool] on bsd in js.
CI error

Example

import macros
macro enmRange(enm: typed): untyped = result = newNimNode(nnkCurly).add(enm.getType[1][1..^1])

proc fullSet*(T: typedesc): auto {.inline.} =
  when T is Ordinal:
    {T.low..T.high}
  else: # Hole filled enum
    enmRange(T)
assert fullSet(bool) == {true, false}

Current Output

Program fails at runtime for whatever reason.

`fullSet(bool) == {true, false}`  [AssertionDefect]

[timotheecour]

• for whoever has a freebsd/openbsd VM, please minimize; i'm assuming this would also fail:

proc fullSet*(T: typedesc): auto {.inline.} =
  {T.low..T.high}
doAssert fullSet(bool) == {true, false}

and possibly simpler even

• please also look at:

import std/jsconsole
proc fullSet*(T: typedesc): auto {.inline.} =
  {T.low..T.high}
template fn(a) =
  echo a
  console.log a
fn fullSet(bool)
fn {true, false}

(/cc @euantorano )

[euantorano]

Will give this a go tomorrow. The most common causes for errors like this are different versions of node. It seems the CI run used node 15.8.0.

…

On Thu, 18 Feb 2021, at 22:51, Timothee Cour wrote: * for whoever has a freebsd/openbsd VM, please minimize; i'm assuming this would also fail: proc fullSet*(T: typedesc): auto {.inline.} = {T.low..T.high} doAssert fullSet(bool) == {true, false} and possibly simpler even * please also look at: import std/jsconsole proc fullSet*(T: typedesc): auto {.inline.} = {T.low..T.high} template fn(a) = echo a console.log a fn fullSet(bool) fn {true, false} (/cc @euantorano <https://github.com/euantorano> ) — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#17093 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAFW24MTXCCEKGQBZ73KZNDS7WKWXANCNFSM4X3HUOIA>.

[timotheecour]

I finally found root cause: see #17097 (@euantorano please read that issue since it may affect you in other PR's)

[euantorano]

I finally found root cause: see #17097 (@euantorano please read that issue since it may affect you in other PR's)

Yeah, I just saw your email to lists.sr.ht and was reading through it. I wonder if we can fix it in the CI script if the actual platform can’t/won’t change.