1735775 - Permanent failure js/src/jit-test/tests/arrays/change-array-by-copy.js | Error: Invalid long option: --enable-change-array-by-copy (code 1, args "--enable-change-array-by-copy") [0.0 s] when Gecko 95 merges to Beta on 2021-11-01

Status: RESOLVED FIXED

(Core :: JavaScript Engine, defect)

Description

[Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout)]

• central-as-beta simulation
• How to run these simulations
• Failure log

TEST-UNEXPECTED-FAIL | js/src/jit-test/tests/arrays/change-array-by-copy.js | Error: Invalid long option: --enable-change-array-by-copy (code 1, args "--enable-change-array-by-copy") [0.0 s]

That's from bug 1729563. Matthew, can you drive the fix?

Comment 1

[Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout)]

Please also take notice of the other failing tests which are also fixed if the changes in bug 1729563 get reverted.

Comment 2

[Matthew Gaudet (he/him) [:mgaudet]]

Ok. So the original 'Invalid long option: --enable-change-array-by-copy' bug is fixable with relatively little work, but the other failures are more concerning. I'm seeing lots of failures

Assertion failure: efs->argCount == args.length() - 1, at /builds/worker/checkouts/gecko/js/src/vm/SelfHosting.cpp:334 

Looking at this failure in a debugger, I'm seeing that assert being issued when we're trying to issue a JSMSG_OBJECT_REQUIRED_ARG error from self-hosted code.

These do indeed go away with backout. Given two things: 1) I can't figure out why these failures only show up on a central-as-beta sim, 2) How this is showing up non-array-change-by-copy code, I'm gong to request backout. This is a disabled-by-default feature, so should have no impact on beta at all.

I'll update this bug when the backout is processed.

Comment 3

[Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout)]

Backout got merged to central in bug 1729563 and central-as-beta simulation doesn't report any of those errors anymore.