697279 - "Assertion failure: !tc->inStrictMode(),"

Status: RESOLVED FIXED

(Core :: JavaScript Engine, defect)

Description

[Gary Kwong [:gkw] [:nth10sd] (NOT official MoCo now)]

Attachment: stack

Reflect.parse("\"use strict\";*")

asserts js debug shell on m-c changeset f2fa4ae74ee1 without any CLI flags at Assertion failure: !tc->inStrictMode()

Definitely a regression, autoBisect is on it.

Comment 1

[Christian Holler (:decoder)]

I also saw this on LangFuzz because one of the jit-tests is failing with that assertion. I minimized that test and got:

eval("'use strict'; @7");

Comment 2

[Gary Kwong [:gkw] [:nth10sd] (NOT official MoCo now)]

autoBisect shows this is probably related to the following changeset:

The first bad revision is:
changeset:   79134:0cff4fe76772
user:        Brendan Eich
date:        Sun Oct 23 22:42:29 2011 -0700
summary:     Ban E4X in ES5 strict mode (bug 695577, r=luke).

This blows up the fuzzers, including decoder's.

Comment 3

[Gary Kwong [:gkw] [:nth10sd] (NOT official MoCo now)]

> This blows up the fuzzers, including decoder's.

While this bug is on our ignore list, it will prove really really helpful to have this fixed as it shows up on almost every iteration of fuzzing on debug shells, at least on jsfunfuzz.

Can we have this fixed soon, please? (or at least a patch would be nice?)

Comment 4

[Tom S]

Never really looked into that stuff, but i will take a shot at it.

Comment 5

[Tom S]

Attachment: v1

Maybe this could be done somewhere else, but this fixes it.

Comment 6

[Gary Kwong [:gkw] [:nth10sd] (NOT official MoCo now)]

(In reply to Tom Schuster (evilpie) from comment #5)
> Created attachment 572612 [details] [diff] [review] [diff] [details] [review]
> v1
> 
> Maybe this could be done somewhere else, but this fixes it.

This patch seems to hold up fine after 10 minutes of fuzzing, and it prevents the blowing up of fuzzers.

Comment 7

[Brendan Eich [:brendan]]

Comment on attachment 572612 [details] [diff] [review]
v1

Thanks.

/be

Comment 8

[Tom S]

http://hg.mozilla.org/integration/mozilla-inbound/rev/6b839530a88a

Comment 9

[Ed Morley [:emorley]]

https://hg.mozilla.org/mozilla-central/rev/6b839530a88a

Comment 10

[Gary Kwong [:gkw] [:nth10sd] (NOT official MoCo now)]

Nominating for esr10 probably after 10.0.3, this will be nice to have for fuzzing on that branch.

Without the patch a lot of noise is generated - even though we could ignore this assert, due to the simple nature of the testcase, this may mask other asserts.

There have not been regressions on other branches since this patch landed some months ago.

Comment 11

[Lukas Blakk [:lsblakk] use ?needinfo]

Comment on attachment 572612 [details] [diff] [review]
v1

[Triage Comment]
Please go ahead and land, see https://wiki.mozilla.org/Release_Management/ESR_Landing_Process for details.

Comment 12

[Gary Kwong [:gkw] [:nth10sd] (NOT official MoCo now)]

Landed on esr10:

http://hg.mozilla.org/releases/mozilla-esr10/rev/27b1269dac4d

Comment 13

[u279076]

Verified fixed in Firefox 10.0.5esrpre 2012-05-31 js-shell.