Bug 165483 – REGRESSION(r209399): Causes crashes when dumping JIT disassembly

• Status: RESOLVED
• Resolution: FIXED
• Component: JavaScriptCore

[Michael Saboff]

Looks like the RELEASE_ASSERT() at the end of sixCharacterHashStringToInteger() has an inverted check.  We want to assert that the string is terminated with a null character.

[Michael Saboff]

Created attachment 296306 [details]
Patch

[Geoffrey Garen]

Comment on attachment 296306 [details]
Patch

r=me

[Darin Adler]

Comment on attachment 296306 [details]
Patch

Oops! Thanks very much for fixing this.

[Darin Adler]

Wonder why EWS missed it.

[Michael Saboff]

(In reply to comment #4)
> Wonder why EWS missed it.

I don't think the debug tests exercise this path.  That would require something to dump the hashed name of a compiled JS function.

[WebKit Commit Bot]

Comment on attachment 296306 [details]
Patch

Clearing flags on attachment: 296306

Committed r209413: <http://trac.webkit.org/changeset/209413>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.

[Ryan Haddad]

*** Bug 165490 has been marked as a duplicate of this bug. ***

[Michael Saboff]

(In reply to comment #5)
> (In reply to comment #4)
> > Wonder why EWS missed it.
> 
> I don't think the debug tests exercise this path.  That would require
> something to dump the hashed name of a compiled JS function.

I was wrong, the debug bots did catch this (https://bugs.webkit.org/show_bug.cgi?id=165490).  It just took a little time.

[Darin Adler]

Are "the debug bots" part of EWS, or part of some other automated WebKit testing?

[Darin Adler]

Also, Michael, do you know why these need to be RELEASE_ASSERT and not just ASSERT?

[Michael Saboff]

(In reply to comment #10)
> Are "the debug bots" part of EWS, or part of some other automated WebKit
> testing?

There is a mac-debug EWS bot.  I don't know if this bot would / did catch the issue.  It would depend on what tests that bot runs.

> Also, Michael, do you know why these need to be RELEASE_ASSERT and not just ASSERT?

I don't know why this isn't simply an ASSERT.

[Ryan Haddad]

(In reply to comment #12)
> (In reply to comment #10)
> > Are "the debug bots" part of EWS, or part of some other automated WebKit
> > testing?
> 
> There is a mac-debug EWS bot.  I don't know if this bot would / did catch
> the issue.  It would depend on what tests that bot runs.

The failure was caught after the patch was landed by the bots that run JSC tests. EWS does not currently run JSC tests.