913885 - Differential Testing: Different output message involving schedulegc

Status: RESOLVED FIXED

(Core :: JavaScript Engine, defect)

Description

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

try {
    Object.defineProperty(this, "a1", {
        get: function() {
            return [({}), eval("eval(\"\",\"\")", x), ({})]
        }
    })
    a1;
    var x, u;
} catch (e) {}
try {
    schedulegc(5);
    a1
} catch (e) {}

prints the following on js opt shell on m-c changeset 3697f962bb7b without any CLI arguments, when the testcase is passed in as a CLI argument:

1613-interesting.js:4:12 warning: Support for eval(code, scopeObject) has been removed. Use |with (scopeObject) eval(code);| instead.
1613-interesting.js:4:0 warning: Support for eval(code, scopeObject) has been removed. Use |with (scopeObject) eval(code);| instead.

However it prints the following with --ion-eager:

1613-interesting.js:4:12 warning: Support for eval(code, scopeObject) has been removed. Use |with (scopeObject) eval(code);| instead.
1613-interesting.js:4:0 warning: Support for eval(code, scopeObject) has been removed. Use |with (scopeObject) eval(code);| instead.
1613-interesting.js:4:0 warning: Support for eval(code, scopeObject) has been removed. Use |with (scopeObject) eval(code);| instead.


My configure flags are:

CC="clang -Qunused-arguments" AR=ar CXX="clang++ -Qunused-arguments" sh ./configure --target=x86_64-apple-darwin11.4.0 --enable-optimize --disable-debug --enable-profiling --enable-gczeal --enable-debug-symbols --enable-methodjit --enable-type-inference --disable-tests --enable-more-deterministic --with-ccache --enable-threadsafe <more NSPR compilation flags>

Comment 1

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

autoBisect shows this is probably related to the following changeset:

The first bad revision is:
changeset:   http://hg.mozilla.org/mozilla-central/rev/77e2eaaf2fbb
user:        Jason Orendorff
date:        Fri Sep 06 12:05:30 2013 -0500
summary:     Bug 893186, part 5 - Delete JS_GetPropertyAttributes and many similar APIs; remove half a dozen methods from JS::Class. r=Waldo.

Jason, is bug 893186 a possible regressor?

Comment 2

[Jason Orendorff [:jorendorff]]

It's possible, but the bug doesn't reproduce for me in tip. Building 77e2eaaf2fbb now...

Comment 3

[Jason Orendorff [:jorendorff]]

I can't reproduce it in 77e2eaaf2fbb either.

Whether or not we generate the third warning depends on whether we get an eval cache hit. So I think it's nondeterministic, and it would be good to skip this warning when fuzzing.

However I think it's better to delete this warning entirely now. Waldo, do you agree?

Comment 4

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

(In reply to Jason Orendorff [:jorendorff] from comment #3)
> I can't reproduce it in 77e2eaaf2fbb either.

I tested on Mac 10.8, and my configure options are as in comment 0. How about rev 3697f962bb7b as specified there?

It may have gone away though..

Comment 5

[Jeff Walden [:Waldo]]

I think at this point we can just kill the warning.  We haven't shipped two-argument eval in years, now.

Comment 6

[Jason Orendorff [:jorendorff]]

Attachment: v1

OK then.

Comment 7

[Tom S [:evilpie]]

Comment on attachment 803788 [details] [diff] [review]
v1

Review of attachment 803788 [details] [diff] [review]:
-----------------------------------------------------------------

::: js/src/builtin/Eval.cpp
@@ -407,5 @@
> -WarnOnTooManyArgs(JSContext *cx, const CallArgs &args)
> -{
> -    if (args.length() > 1) {
> -        Rooted<JSScript*> script(cx, cx->currentScript());
> -        if (script && !script->warnedAboutTwoArgumentEval) {

You forgot about warnedAboutTwoArgumentEval in JSScript

Comment 8

[Jeff Walden [:Waldo]]

Comment on attachment 803788 [details] [diff] [review]
v1

Review of attachment 803788 [details] [diff] [review]:
-----------------------------------------------------------------

Yeah, remove the deadweight JSScript field, too.

Comment 9

[Jason Orendorff [:jorendorff]]

Thanks, Tom!

https://hg.mozilla.org/integration/mozilla-inbound/rev/29458f6c07e5

Comment 10

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/29458f6c07e5