Bug 39249

Summary: REGRESSION (r56295): Can't create a new wave on Google Wave
Product: WebKit Reporter: Alexey Proskuryakov <ap>
Component: Page LoadingAssignee: Daniel Bates <dbates>
Status: RESOLVED FIXED    
Severity: Normal CC: abarth, ain, alin0steglinski, ap, dbates, dglazkov, webkit.review.bot
Priority: P1 Keywords: InRadar, Regression, XSSAuditor
Version: 528+ (Nightly build)   
Hardware: Mac   
OS: OS X 10.6   
Attachments:
Description Flags
Patch with test case
none
Patch with test case
none
Patch with test case
none
Patch with test case abarth: review+

Alexey Proskuryakov
Reported 2010-05-17 16:18:55 PDT
Observed with r59204 nightly. Steps to reproduce: 1. Log in to Google Wave. 2. Click "New Wave" at the right. Results: window navigates to about:blank. Expected results: a new wave should appear at the right.
Attachments
Patch with test case (10.66 KB, patch)
2010-05-18 00:50 PDT, Daniel Bates
no flags
Patch with test case (14.21 KB, patch)
2010-05-18 23:37 PDT, Daniel Bates
no flags
Patch with test case (14.60 KB, patch)
2010-05-19 01:05 PDT, Daniel Bates
no flags
Patch with test case (14.24 KB, patch)
2010-05-19 01:10 PDT, Daniel Bates
abarth: review+
Alexey Proskuryakov
Comment 1 2010-05-17 16:19:11 PDT
Alexey Proskuryakov
Comment 2 2010-05-17 17:10:11 PDT
Works in r56294, fails in r56295.
Alexey Proskuryakov
Comment 3 2010-05-17 17:24:17 PDT
The main resource has " X-Xss-Protection:1; mode=block".
Daniel Bates
Comment 4 2010-05-17 19:04:19 PDT
(In reply to comment #0) > Observed with r59204 nightly. > > Steps to reproduce: > 1. Log in to Google Wave. > 2. Click "New Wave" at the right. > > Results: window navigates to about:blank. > Expected results: a new wave should appear at the right. I am unable to reproduce following the aforementioned steps with r59204 nightly and using my Google account. I tried resetting Safari (Safari->Reset Safari), then following the steps, to no avail. Do you happen to have a copy of the console messages when this occurred? Do you have any additional information on this issue?
Alexey Proskuryakov
Comment 5 2010-05-17 19:12:54 PDT
It seems to only happen for some people, maybe Google serves different code versions? The link is <a href="javascript:;"> with some even listeners installed. Below is JS Console output: HTML WARN: The page at https://wave.google.com/wave/ displayed insecure content from http://lh3.ggpht.com/__n9uIWxqWWk/SwXq4UxcJjI/AAAAAAAAA9k/Jk0EV1lvgMc/s104-c/IMG_1611.jpg. HTML WARN: The page at https://wave.google.com/wave/ displayed insecure content from http://lh3.ggpht.com/__n9uIWxqWWk/SwXq4UxcJjI/AAAAAAAAA9k/Jk0EV1lvgMc/s104-c/IMG_1611.jpg. JS ERROR: Refused to execute a JavaScript script. Source code of script found within request. https://wave.google.com/wave/static/5DEB8B560FCA74C2AA4974967ED07055.cache.js:4218: JS ERROR: TypeError: Result of expression '(v.Y?$doc.getElementById(d):Upb(v,d))' [null] is not an object.
Alexey Proskuryakov
Comment 6 2010-05-17 19:20:52 PDT
I know! I have a semicolon in my Google Wave password - this is why others can't reproduce.
Daniel Bates
Comment 7 2010-05-18 00:50:07 PDT
Created attachment 56332 [details] Patch with test case Pass the schema portion of a JavaScript URL to the XSSAuditor as additional context.
WebKit Review Bot
Comment 8 2010-05-18 01:04:24 PDT
Daniel Bates
Comment 9 2010-05-18 01:07:26 PDT
Comment on attachment 56332 [details] Patch with test case Need to add the equivalent changes to the v8 bindings.
Daniel Bates
Comment 10 2010-05-18 23:37:22 PDT
Created attachment 56464 [details] Patch with test case Updated patch with changes to V8 bindings and added missing file "http/tests/security/xssAuditor/resources/javascript-link-safe.html".
WebKit Review Bot
Comment 11 2010-05-19 00:47:21 PDT
Daniel Bates
Comment 12 2010-05-19 01:05:10 PDT
Created attachment 56473 [details] Patch with test case
Daniel Bates
Comment 13 2010-05-19 01:10:00 PDT
Created attachment 56474 [details] Patch with test case Fixed V8 prototype for ScriptController::executeIfJavaScriptURL
Alexey Proskuryakov
Comment 14 2010-05-20 11:34:22 PDT
Adam, is this something you'd be willing to review?
Adam Barth
Comment 15 2010-05-20 12:33:47 PDT
I'd be happy to. I'm at W2SP today, but I'll try to review it tonight.
Adam Barth
Comment 16 2010-05-21 15:03:55 PDT
Comment on attachment 56474 [details] Patch with test case Very nice.
Daniel Bates
Comment 17 2010-05-22 13:51:00 PDT
Alexey Proskuryakov
Comment 18 2010-05-23 18:40:35 PDT
*** Bug 38137 has been marked as a duplicate of this bug. ***
Ain Tohvri
Comment 19 2010-05-24 01:15:20 PDT
Behaviour described in bug 39186 is still there in latest r60027.
Note You need to log in before you can comment on or make changes to this bug.