Bug 17250

Summary:

instanceof Array test fails when using iframes

Product:

WebKit

Reporter:

Sanjiv Jivan <sanjiv.jivan+wk>

Component:

JavaScriptCore

Assignee:

Nobody <webkit-unassigned>

Status:

RESOLVED FIXED

Severity:

Major

CC:

ap, aroben, buildbot, ddkilzer, matthew.lieder, oliver, sam, sanjiv.jivan+wk, webkit-ews, webkit.review.bot, zwarich

Priority:

Keywords:

HasReduction

Version:

523.x (Safari 3)

Hardware:

OS:

Windows XP

Bug Depends on:

57333

Bug Blocks:

Attachments:

Description	Flags
child.html	none
parent.html (test case)	none
Patch	none
Patch	none
Patch	ggaren: review+

Sanjiv Jivan

Reported 2008-02-08 22:46:41 PST

I have an iframe where I'm creating an Array with in the parent's context via new parent.Array(). I now call a function in the parent window passing this array. When the parent function receives this argument, it fails the instanceof test. I understand that passing "new Array()" from the iframe to parent will cause the "instanceof Array " test to fail in the parent because the contexts of the two are different, but in this case I am creating a array from the parent context itself yet if fails the "instanceof Array" test. Below is a simple testcase illustrating the problem. This test runs fine on IE, Firefox and Opera but fails on Safari only. parent.html <html> <head> </head> <body> <script type="text/javascript"> function isArray(arr) { alert(arr instanceof Array); } </script> </body> <iframe src="child.html"></iframe> </html> child.html <html> <head> </head> <body> <script type="text/javascript"> function test() { var arr = new parent.Array(); parent.isArray(arr); } </script> <button onclick="test()">Array Test</button> </body> </html>

Attachments
child.html (234 bytes, text/html) 2009-01-13 16:35 PST, David Kilzer (:ddkilzer)	no flags	Details
parent.html (test case) (207 bytes, text/html) 2009-01-13 16:36 PST, David Kilzer (:ddkilzer)	no flags	Details
Patch (46.29 KB, patch) 2011-03-25 18:33 PDT, Oliver Hunt	no flags	Details Formatted Diff Diff
Patch (51.72 KB, patch) 2011-03-28 13:15 PDT, Oliver Hunt	no flags	Details Formatted Diff Diff
Patch (52.08 KB, patch) 2011-03-28 14:06 PDT, Oliver Hunt	ggaren: review+	Details Formatted Diff Diff
Show Obsolete (3) View All Add attachment proposed patch, testcase, etc.

David Kilzer (:ddkilzer)

Comment 1 2009-01-13 16:35:39 PST

Created attachment 26690 [details] child.html

David Kilzer (:ddkilzer)

Comment 2 2009-01-13 16:36:23 PST

Created attachment 26691 [details] parent.html (test case)

David Kilzer (:ddkilzer)

Comment 3 2009-01-13 16:38:12 PST

Confirmed with WebKit nightly build r39872.

David Kilzer (:ddkilzer)

Comment 4 2009-01-13 16:39:24 PST

(In reply to comment #3) > Confirmed with WebKit nightly build r39872. Confirmed the test case alerts "true" in Firefox 3, but alerts "false" in WebKit/Safari.

Matthew Lieder

Comment 5 2011-03-25 09:51:38 PDT

FYI, this is a problem in Firefox 4 now too (test case alerts "false")

Oliver Hunt

Comment 6 2011-03-25 18:33:25 PDT

Created attachment 87001 [details] Patch

WebKit Review Bot

Comment 7 2011-03-25 18:35:09 PDT

Attachment 87001 [details] did not pass style-queue: Failed to run "['Tools/Scripts/check-webkit-style', '--diff-files', u'Source/JavaScriptCore/API/JSObjectRef.cpp'..." exit_code: 1 Source/JavaScriptCore/runtime/JSCell.h:38: Code inside a namespace should not be indented. [whitespace/indent] [4] Total errors found: 1 in 34 files If any of these errors are false positives, please file a bug against check-webkit-style.

Early Warning System Bot

Comment 8 2011-03-25 18:48:49 PDT

Attachment 87001 [details] did not build on qt: Build output: http://queues.webkit.org/results/8256011

Build Bot

Comment 9 2011-03-25 18:56:59 PDT

Attachment 87001 [details] did not build on win: Build output: http://queues.webkit.org/results/8256017

Oliver Hunt

Comment 10 2011-03-28 13:15:59 PDT

Created attachment 87186 [details] Patch

WebKit Review Bot

Comment 11 2011-03-28 13:19:11 PDT

Attachment 87186 [details] did not pass style-queue: Failed to run "['Tools/Scripts/check-webkit-style', '--diff-files', u'LayoutTests/ChangeLog', u'LayoutTests/fast..." exit_code: 1 Source/JavaScriptCore/runtime/JSCell.h:38: Code inside a namespace should not be indented. [whitespace/indent] [4] Total errors found: 1 in 38 files If any of these errors are false positives, please file a bug against check-webkit-style.

Build Bot

Comment 12 2011-03-28 13:56:18 PDT

Attachment 87186 [details] did not build on win: Build output: http://queues.webkit.org/results/8273473

Oliver Hunt

Comment 13 2011-03-28 14:06:51 PDT

Created attachment 87200 [details] Patch

WebKit Review Bot

Comment 14 2011-03-28 14:09:23 PDT

Attachment 87200 [details] did not pass style-queue: Failed to run "['Tools/Scripts/check-webkit-style', '--diff-files', u'LayoutTests/ChangeLog', u'LayoutTests/fast..." exit_code: 1 Source/JavaScriptCore/runtime/JSCell.h:38: Code inside a namespace should not be indented. [whitespace/indent] [4] Total errors found: 1 in 39 files If any of these errors are false positives, please file a bug against check-webkit-style.

Geoffrey Garen

Comment 15 2011-03-28 15:30:25 PDT

Comment on attachment 87200 [details] Patch r=me

Oliver Hunt

Comment 16 2011-03-28 16:40:25 PDT

Committed r82173

Adam Roben (:aroben)

Comment 17 2011-03-29 06:51:56 PDT

This caused assertion and test failures in run-javascriptcore-tests on Windows. Here's the assertion: ASSERTION FAILED: isCell() > JavaScriptCore.dll!JSC::JSValue::asCell() Line 559 + 0x30 bytes C++ JavaScriptCore.dll!JSC::asObject(JSC::JSValue value={...}) Line 393 + 0x8 bytes C++ JavaScriptCore.dll!JSC::asInternalFunction(JSC::JSValue value={...}) Line 65 + 0x12 bytes C++ JavaScriptCore.dll!JSC::constructDate(JSC::ExecState * exec=0x012201a0, const JSC::ArgList & args={...}) Line 124 + 0x3e bytes C++ JavaScriptCore.dll!JSObjectMakeDate(const OpaqueJSContext * ctx=0x012201a0, unsigned int argumentCount=1, const OpaqueJSValue * const * arguments=0x0012fdd8, const OpaqueJSValue * * exception=0x00000000) Line 171 + 0x22 bytes C++ testapi.exe!main(int argc=1, char * * argv=0x013c99d8) Line 1304 + 0x19 bytes C++ testapi.exe!__tmainCRTStartup() Line 597 + 0x17 bytes C kernel32.dll!_BaseProcessStart@4() + 0x23 bytes I'm rolling out this patch in bug 57333.

Adam Roben (:aroben)

Comment 18 2011-03-29 10:47:17 PDT

I guess this can be resolved again since Oliver fixed the Windows issues.

Note You need to log in before you can comment on or make changes to this bug.