Closed
Bug 494199
Opened 17 years ago
Closed 17 years ago
New Yahoo Mail Categories Broken
Categories
(Core :: JavaScript Engine, defect)
Core
JavaScript Engine
Tracking
()
RESOLVED
DUPLICATE
of bug 494544
People
(Reporter: Portfolioso, Assigned: jorendorff)
References
()
Details
(Keywords: regression)
Attachments
(2 files)
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2a1pre) Gecko/20090521 Minefield/3.6a1pre (.NET CLR 3.5.30729)
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2a1pre) Gecko/20090521 Minefield/3.6a1pre (.NET CLR 3.5.30729)
The new Yahoo mail does not work in the latest nightly. Selecting a different category in the left menu results in it being highlighted, rather than loading messages from the selected category
Reproducible: Always
Steps to Reproduce:
1. Visit http://us.mg1.mail.yahoo.com/dc/launch
2. Click on various categories (ex. sent, drafts, spam, etc)
Actual Results:
The categories highlight and the messages do not load
Expected Results:
The categories should load the appropriate messages
This regressed on 5/21's build. I don't have an hourly, but I'm assuming it's something to do with all the javascript checkins.
Broken with JIT enabled and disabled.
Flags: blocking-firefox3.6?
Keywords: regression
Updated•17 years ago
|
Version: unspecified → Trunk
Updated•17 years ago
|
Broken for me too.
Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2a1pre) Gecko/20090521 Minefield/3.6a1pre
Updated•17 years ago
|
Status: UNCONFIRMED → NEW
Ever confirmed: true
Could you please download hourly builds and pinpoint the regression range between hourly builds? Then post the changesets that are listed in about:buildconfig so we can get a narrow regression range.
Comment 4•17 years ago
|
||
Comment 5•17 years ago
|
||
This occurs on Mac as well using Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US; rv:1.9.2a1pre) Gecko/20090521 Minefield/3.6a1pre, changing to all.
OS: Windows XP → All
Hardware: x86 → All
Comment 6•17 years ago
|
||
pushlog range based on comment 4:
http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=13b38e2a54ee&tochange=cb4d2ce3b5db
pitty that there don't seem to be archieved hourly TM builds for 05.19/20th :-/
Assignee: nobody → general
Component: General → JavaScript Engine
Flags: blocking-firefox3.6?
Product: Firefox → Core
QA Contact: general → general
Updated•17 years ago
|
Flags: blocking1.9.2?
Updated•17 years ago
|
Flags: blocking1.9.1?
Comment 7•17 years ago
|
||
Does this affect branch? Please nominate as blocker if so.
Comment 8•17 years ago
|
||
Yes, affects latest-nightly on 1.9.1, blocking. Screenie:
http://www.grabup.com/uploads/74ddaeab9ee79268ca07fd5a77af3a79.png
(note that the error console shows:
Error: Permission denied for <http://ad.yieldmanager.com> to call method Location.toString on <http://us.mg1.mail.yahoo.com>.
though I think that is unrelated)
Flags: blocking1.9.1? → blocking1.9.1+
Comment 9•17 years ago
|
||
(I, too, think that is unrelated.)
Yeah, wish we had some more TM hourlies around. To the try server, or making builds?
Comment 10•17 years ago
|
||
(In reply to comment #9)
> (I, too, think that is unrelated.)
>
> Yeah, wish we had some more TM hourlies around. To the try server, or making
> builds?
http://hourly-archive.localgho.st/hourly-archive2/ includes TM hourly builds
Comment 11•17 years ago
|
||
(In reply to comment #10)
> (In reply to comment #9)
> > (I, too, think that is unrelated.)
> >
> > Yeah, wish we had some more TM hourlies around. To the try server, or making
> > builds?
>
> http://hourly-archive.localgho.st/hourly-archive2/ includes TM hourly builds
yes, but - as i already mentioned in comment 6 - there are none of may 19th/20th what yould help here finding the causing checkin :-/. actually there are non past may 16th.
Comment 12•17 years ago
|
||
Ccing shaver with a question mark on my face: where are our hourlies?
Updated•17 years ago
|
Assignee: general → jorendorff
Comment 13•17 years ago
|
||
The 1.9.1 branch (Shiretoko) nightlies broke the same exact day,
Gecko/20090521. Cross referencing the changes listed at
http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=13b38e2a54ee&tochange=cb4d2ce3b5db
with the ones listed at:
http://hg.mozilla.org/releases/mozilla-1.9.1/pushloghtml?startdate=may+20%2C+2009&enddate=may+21%2C+2009
should eliminate some of the changesets. It looks like the following bugs were
in both branches simultaneously:
Bug 493657, Bug 493760, Bug 492904, Bug 493545
This doesn't include any TraceMonkey merges done to the trunk, but it doesn't
look like 1.9.1 got any of those merges any way.
Comment 14•17 years ago
|
||
Just an update. I checked my update logs to see what version I was running the other day that worked.
Shiretoko 3.5pre (20090521031853) worked, but Shiretoko 3.5pre (20090521152500) did not.
| Assignee | ||
Comment 15•17 years ago
|
||
I can reproduce this only using general.useragent.override to pass as Ff 3.0. I had no luck getting the application to run locally after saving the page or figuring out how that tree control is supposed to work. I know the blueness is applied by hacking element.style.backgroundColor, which I think happens in combo_004.js. Presumably what's supposed to happen is that the new button is highlighted and the old one is un-highlighted, so I'll poke at that code a bit more.
Maybe I should instead try to make a web page that uses the YUI Library and makes a TreeView... but so far I see nothing that tells me they're using a particular version of YUI, so that could be a big waste of time.
Comment 16•17 years ago
|
||
The new Yahoo Mail is built using AJAX so it's highly doubtful that you'll be able to run it locally. You can run the new Yahoo Mail without needing to change any preferences by going to:
http://us.mg.mail.yahoo.com/dc/launch?sysreq=ignore
As for the issue, clicking the "folder" is supposed to trigger the code that actually switches the displayed folder to the selected one and the removes the highlighting from the old folder. That code apparently never kicks off.
I do notice that if you select the "Home" tab first and then click a folder, the folder will appear to load in that the tab will have the folder's name and message count, but the actual list won't update.
Another weird thing, is that if you try to add a folder, it will add an "untitled" as it usually does, but selected the entitled so you can type in it, it will make the "inbox" so you can type in it. That makes me wonder if some index on Yahoo's side is getting messed up.
Comment 17•17 years ago
|
||
Another thing to try is I tried sending a new email and got an alert box that said J is undefined.
Comment 18•17 years ago
|
||
I used Firebug 1.4.0a29 to track errors on the Yahoo Mail page. In the Minefield Gecko/20090522 build I get a ton of try/catched errors in http://mail.yimg.com/d/combo?/pim/r/dclient/cg1277_43/js/us/strings.js&/pim/r/dclient/cg1277_43/js/main/main.js&/bc/bc_2.0.4.js
It always throws on "line" 851 in the GU function, which is called by the FX function. I say "line", because the code in bc_2.0.4.js is all on one big line. The number 851 is what Firebug is reporting.
Minefield Gecko/20090520 generated no errors at all when changing folders (or doing anything else for that matter).
I just remembered that there is a way to view the Yahoo debug console in Yahoo Mail. If you hold down SHIFT and CTRL while clicking help, it will bring up a menu where you can click on Debug log.
I clicked on a folder test in a working version of Minefield and this is what I got:
53. 5/22 19:38:01.354
SwitchToFolder: test
54. 5/22 19:38:02.023
handleItemSelect: deselectAll; stype: null
55. 5/22 19:38:02.088
Switch Folder Profile
total 853
showMainTabIfHidden 121
focusPane 1
selectFolderRow 0
kickFoldersPane 1
kickMainPane 81
loadDataForRange1 1
56. 5/22 19:38:02.139
refreshMessagePane: false, i2_3
57. 5/22 19:38:02.140
blankOutMessageDisplay from RefreshMessagePane
58. 5/22 19:38:02.140
blankOutMessageDisplay
59. 5/22 19:38:02.387
Ad Info: Ad INFO REMOVED
60. 5/22 19:38:02.387
Ad Urls: Ad URL REMOVED
61. 5/22 19:38:02.825
MON ad took 437ms
62. 5/22 19:38:05.516
CmdName:MT:launch:toggle_help
----
I clicked on the same folder in a broken version of Minefield and this is what I got:
----
53. 5/22 19:35:52.549
Tab switch profile: welcome -> main
total 131
ts.pre 0
ts.render 131
ts.render.content 122
ts.render.content.aboutToDeactivate 0
ts.render.content.addClass 1
ts.render.content.activate 45
ts.render.content.deactivate 3
ts.render.tabs.2 9
ts.fixupPanes 0
54. 5/22 19:35:52.550
SwitchToFolder: test
55. 5/22 19:35:52.672
Error Message: Caught JS error after launch Script error. url:http://mail.yimg.com/d/combo?/pim/r/dclient/cg1277_43/js/us/strings.js&/pim/r/dclient/cg1277_43/js/main/main.js&/bc/bc_2.0.4.js line:0
56. 5/22 19:35:52.672
Stack Trace: > handleErr
57. 5/22 19:35:52.677
Caught an error --> ScriptError-otherBrows-win-aLaunch-otherBrows-win-non-launch-error
58. 5/22 19:35:56.856
CmdName:MT:launch:toggle_help
So the latest nightlies are causing a javascript errors to be thrown in
http://mail.yimg.com/d/combo?/pim/r/dclient/cg1277_43/js/us/strings.js&/pim/r/dclient/cg1277_43/js/main/main.js&/bc/bc_2.0.4.js
That file is highly encapsulated so I doubt looking at it will be much use. I think the easiest thing to do would be to revert all the changes made in the changeset that triggered this issue and then add them one at a time until it breaks again.
Comment 19•17 years ago
|
||
I switched from Firebug to using the Javascript Debugger extension since it does a better job at logging caught exceptions. The difference between a version of Minefield that works and one that does not, is the one that does not throws the following exceptions:
Error ``A is undefined'' [x-] in file ``http://mail.yimg.com/d/combo?/pim/r/dclient/cg1277_43/js/us/strings.js&/pim/r/dclient/cg1277_43/js/main/main.js&/bc/bc_2.0.4.js'', line 177, character 0.
Exception ``TypeError: A is undefined'' thrown from function anonymous(B=HTMLSpanElement:{2}) in <http://mail.yimg.com/d/combo?/pim/r/dclient/cg1277_43/js/us/strings.js&/pim/r/dclient/cg1277_43/js/main/main.js&/bc/bc_2.0.4.js> line 177.
[e] message = [string] "A is undefined"
Exception ``TypeError: A is undefined'' thrown from function Sn(B=Element:{0}) in <http://mail.yimg.com/d/combo?/pim/r/dclient/cg1277_43/js/us/strings.js&/pim/r/dclient/cg1277_43/js/main/main.js&/bc/bc_2.0.4.js> line 621.
[e] message = [string] "A is undefined"
Exception ``TypeError: A is undefined'' thrown from function anonymous() in <http://mail.yimg.com/d/combo?/pim/r/dclient/cg1277_43/js/us/strings.js&/pim/r/dclient/cg1277_43/js/main/main.js&/bc/bc_2.0.4.js> line 617.
[e] message = [string] "A is undefined"
Exception ``TypeError: A is undefined'' thrown from function anonymous() in <http://mail.yimg.com/d/combo?/pim/r/dclient/cg1277_43/js/us/strings.js&/pim/r/dclient/cg1277_43/js/main/main.js&/bc/bc_2.0.4.js> line 616.
[e] message = [string] "A is undefined"
Exception ``TypeError: A is undefined'' thrown from function anonymous(B=integer:0) in <http://mail.yimg.com/d/combo?/pim/r/dclient/cg1277_43/js/us/strings.js&/pim/r/dclient/cg1277_43/js/main/main.js&/bc/bc_2.0.4.js> line 617.
[e] message = [string] "A is undefined"
Exception ``TypeError: A is undefined'' thrown from function anonymous() in <http://mail.yimg.com/d/combo?/pim/r/dclient/cg1277_43/js/us/strings.js&/pim/r/dclient/cg1277_43/js/main/main.js&/bc/bc_2.0.4.js> line 170.
[e] message = [string] "A is undefined"
Exception ``TypeError: A is undefined'' thrown from function anonymous() in <http://mail.yimg.com/d/combo?/pim/r/dclient/cg1277_43/js/us/strings.js&/pim/r/dclient/cg1277_43/js/main/main.js&/bc/bc_2.0.4.js> line 151.
[e] message = [string] "A is undefined"
Error ``TypeError: A is undefined'' [x-] in file ``http://mail.yimg.com/d/combo?/pim/r/dclient/cg1277_43/js/us/strings.js&/pim/r/dclient/cg1277_43/js/main/main.js&/bc/bc_2.0.4.js'', line 177, character 0.
Unless the line numbers are off, the only references to "A" I see on line 117 are:
for(var A=0;A<B;A++){s[A]=A;}
var A=F;var B=A!=null?YAHOO.widget.PagedTableView.RecordsetSelection.SELECT:YAHOO.widget.PagedTableView.RecordsetSelection.DESELECT_ALL;
Neither should cause an A is undefined exception.
Comment 20•17 years ago
|
||
Is that repeated var A in a function? If so, please extract the function and any enclosing functions (other than a top level "module pattern" function -- that is worth noting, but no need to pull out everything in it).
/be
Comment 21•17 years ago
|
||
It's literally 836 KB of obfuscated javascript code. I can't make heads or tails out of it. I'll attach the whole thing in case someone can make some sense out of it.
At this point, if Firefox 3.5 is really pushing for a June release candidate, your best bet is to either bring in someone from Yahoo or revert out all the changesets made between http://ftp.mozilla.org/pub/mozilla.org/firefox/nightly/2009-05-21-04-mozilla-1.9.1/firefox-3.5pre.en-US.win32.zip and http://ftp.mozilla.org/pub/mozilla.org/firefox/nightly/2009-05-21-15-mozilla-1.9.1/firefox-3.5pre.en-US.win32.zip
Comment 22•17 years ago
|
||
looks like a side effect of bug 493760.
Comment 23•17 years ago
|
||
This may be fixed by the patch in bug 494544. If so, please dup.
/be
Comment 24•17 years ago
|
||
This works in today's tracemonkey nightly with bug 494544 fixed. Sending mail works too.
Status: NEW → RESOLVED
Closed: 17 years ago
Resolution: --- → DUPLICATE
Comment 25•17 years ago
|
||
The code that was failing on line 177 was...
176: ... function MD(){var A;YAHOO.util.Access={ ...
177: ... B:function(F,E,B){if(!A.A()) ... }};A=YAHOO.util.Access
The "A" that is undefined is the first "A" of "!A.A()".
So there was an uninitialized var A that was assigned the object YAHOO.util.Access in the same closure.
data:text/html,<script>var Y;(function(){var A;Y={A:function(){return typeof A;}};A=Y;alert(Y.A())})()</script>
Comment 26•17 years ago
|
||
In the latest seamonkey (nightly) I'm still seeing the Yahoo mail problem.
Comment 27•17 years ago
|
||
I think Seamonkey uses the 1.9.1 branch and that fix is not there yet.
Comment 28•17 years ago
|
||
OK Thanks. I misread the later comments.
Updated•16 years ago
|
Flags: blocking1.9.2?
You need to log in
before you can comment on or make changes to this bug.
Description
•