1126409 - jit-test/tests/ion/simd-bug1121299.js is going to permafail when Gecko 38 merges to Aurora

Status: VERIFIED FIXED

(Core :: JavaScript Engine: JIT, defect)

Description

[Ryan VanderMeulen [:RyanVM]]

[Tracking Requested - why for this release]: Permafail when Gecko 38 is uplifted to Aurora.

https://treeherder.mozilla.org/logviewer.html#?job_id=4499401&repo=try

TEST-UNEXPECTED-FAIL | js/src/jit-test/tests/ion/simd-bug1121299.js | /builds/slave/try_lx-d_sm-arm-sim-0000000000/src/js/src/jit-test/tests/ion/simd-bug1121299.js:7:6 ReferenceError: SIMD is not defined (code 3, args "--baseline-eager --no-fpu")

Comment 1

[Dan Gohman [:sunfish]]

This needs a check like the on added in bug 1115217.

Comment 2

[Benjamin Bouvier [:bbouvier] (inactive)]

Attachment: Add SIMD check in test + move SIMD tests into jit-tests/tests/SIMD

nbp is on PTO today, and it's an easy change.  I've also moved the tests into
the jit-tests/tests/SIMD dir, which was already existing somehow.

Comment 3

[Hannes Verschore [:h4writer]]

Comment on attachment 8555754 [details] [diff] [review]
Add SIMD check in test + move SIMD tests into jit-tests/tests/SIMD

Review of attachment 8555754 [details] [diff] [review]:
-----------------------------------------------------------------

::: js/src/jit-test/tests/SIMD/bug1121299.js
@@ +1,2 @@
> +if (!this.hasOwnProperty("SIMD"))
> +  quit();

Why not:
typeof SIMD === undefined

which is how we normally test this?

Comment 4

[Hannes Verschore [:h4writer]]

s/undefined/"undefined"

Comment 5

[Benjamin Bouvier [:bbouvier] (inactive)]

As discussed on IRC, this semantically doesn't make a difference, so I let it like this.

https://hg.mozilla.org/integration/mozilla-inbound/rev/c4df6e6c0321

Comment 6

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/c4df6e6c0321

Comment 7

[Ryan VanderMeulen [:RyanVM]]

Looks great, thanks :)