579305 - Permanent orange: TEST-UNEXPECTED-FAIL | test-message-header.js | test_show_all_header_mode

Status: RESOLVED FIXED

(Thunderbird :: Testing Infrastructure, defect)

Description

[Mark Banner (:standard8)]

This failure started occurring a few weeks ago on Windows and possibly Linux builds. Mac seems to be fine.

Failure symptoms:

TEST-UNEXPECTED-FAIL | e:\buildbot\win32-comm-central-check\build\mail\test\mozmill\message-header\test-message-header.js | test_show_all_header_mode
  EXCEPTION: more node should be collapsed in all header lines mode
    at: test-message-header.js line 458
       Error("more node should be collapsed in all header lines mode")  0
       subtest_change_to_all_header_mode([object XULElement]) test-message-header.js 458
       test_show_all_header_mode() test-message-header.js 342
            frame.js 468
            frame.js 520
            frame.js 562
            frame.js 411
            frame.js 568
            server.js 164
            server.js 168

Regression range from Tinderbox:

http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=fb46b4128935&tochange=26b9698505d0

There were no comm-central changes at the time.

Comment 1

[rsx11m]

This should be the scope issue due to a missing "let" for moreNode as it was introduced by bug 536542. I'm fixing this in my patch for bug 576611 (I hope).

Comment 2

[rsx11m]

Attachment: Possible fix (won't do)

In case you want to try this before the other patch gets reviewed. It's indeed an independent issue, but I ran into it while updating the tests over there.

Comment 3

[Mark Banner (:standard8)]

Comment on attachment 457860 [details] [diff] [review]
Possible fix (won't do)

I've just tried it on my Windows VM on a debug build, but it doesn't fix the issue. I've also tried adding an extra sleep before that step, but that doesn't seem to fix it either.

Comment 4

[rsx11m]

Comment on attachment 457860 [details] [diff] [review]
Possible fix (won't do)

Yeah, I had a closer look at it and the global-vs.-local issue alone couldn't be the reason, moreNode doesn't seem to be used anywhere outside of the test context. Sorry for the not-fix, I'll leave this in the other patch then.

The View > Headers > All behavior is correct in current trunk builds, expands the address list and removes the "more" node as desired, so it's a bug in the test and not the feature. Maybe there isn't enough time after expanding the headers with change_to_all_header_mode() or it needs some other event to get
it completely redrawn, but it seems that you've tested that already as well.

Comment 5

[rsx11m]

I'm wondering why subtest_change_to_all_header_mode() fails whereas a little earlier subtest_more_widget_click() in test_more_widget() proceeds and is doing the same thing (expanding the message list and collapsing the "more" button).

Note that mc.click is handled differently in both cases; also, Headers > All
may redraw the message whereas clicking more doesn't. Looking at a message with an ".eml" attachment, the message attachment's headers are expanded here as well, thus the message appears to be redrawn and may be introducing the delay.

Comment 6

[Blake Winton (:bwinton) (:☕️)]

Trying to reproduce this on Windows XP (with the latest src-central, and the upgraded mozmill), I get:

mozjs!js_Invoke+0x00000000000002FE (c:\thunderbird\src-central\mozilla\js\src\js
interp.cpp, line 693)
mozjs!js_Interpret+0x000000000000DBA6 (c:\thunderbird\src-central\mozilla\js\src
\jsops.cpp, line 2155)
mozjs!Invoke<int (__cdecl*)(JSContext *,JSObject *,unsigned int,int *,int *)>+0x
00000000000004B0 (c:\thunderbird\src-central\mozilla\js\src\jsinterp.cpp, line 6
02)
mozjs!js_Invoke+0x00000000000002FE (c:\thunderbird\src-central\mozilla\js\src\js
interp.cpp, line 693)
mozjs!js_InternalInvoke+0x000000000000009E (c:\thunderbird\src-central\mozilla\j
s\src\jsinterp.cpp, line 739)
mozjs!JS_CallFunctionValue+0x000000000000009F (c:\thunderbird\src-central\mozill
a\js\src\jsapi.cpp, line 4850)
gklayout!nsJSContext::CallEventHandler+0x000000000000031B (c:\thunderbird\src-ce
ntral\mozilla\dom\base\nsjsenvironment.cpp, line 2204)
gklayout!nsJSEventListener::HandleEvent+0x0000000000000BBE (c:\thunderbird\src-c
entral\mozilla\dom\src\events\nsjseventlistener.cpp, line 228)
gklayout!nsEventListenerManager::HandleEventSubType+0x00000000000002E4 (c:\thund
erbird\src-central\mozilla\content\events\src\nseventlistenermanager.cpp, line 1
094)
gklayout!nsEventListenerManager::HandleEventInternal+0x000000000000031D (c:\thun
derbird\src-central\mozilla\content\events\src\nseventlistenermanager.cpp, line
1192)
gklayout!nsEventListenerManager::HandleEvent+0x00000000000000D5 (c:\thunderbird\
src-central\mozilla\content\events\src\nseventlistenermanager.h, line 147)
gklayout!nsEventTargetChainItem::HandleEvent+0x0000000000000120 (c:\thunderbird\
src-central\mozilla\content\events\src\nseventdispatcher.cpp, line 213)
gklayout!nsEventTargetChainItem::HandleEventTargetChain+0x000000000000019B (c:\t
hunderbird\src-central\mozilla\content\events\src\nseventdispatcher.cpp, line 34
3)
gklayout!nsEventDispatcher::Dispatch+0x000000000000083F (c:\thunderbird\src-cent
ral\mozilla\content\events\src\nseventdispatcher.cpp, line 628)
gklayout!nsEventDispatcher::DispatchDOMEvent+0x000000000000014B (c:\thunderbird\
src-central\mozilla\content\events\src\nseventdispatcher.cpp, line 691)
gklayout!nsEventListenerManager::DispatchEvent+0x00000000000000E3 (c:\thunderbir
d\src-central\mozilla\content\events\src\nseventlistenermanager.cpp, line 1308)
gklayout!nsDOMEventRTTearoff::DispatchEvent+0x000000000000009C (c:\thunderbird\s
rc-central\mozilla\content\base\src\nsgenericelement.cpp, line 1952)
gklayout!nsContentUtils::DispatchTrustedEvent+0x0000000000000126 (c:\thunderbird
\src-central\mozilla\content\base\src\nscontentutils.cpp, line 3327)
gklayout!nsPLDOMEvent::Run+0x000000000000012B (c:\thunderbird\src-central\mozill
a\content\eTimeout: bridge.execFunction("5cf9b140-950f-11df-a3a6-000c2997fceb",
bridge.registry["{65c732e7-4fe3-4a36-a59d-95bd31868c6e}"]["runTestFile"], ["c:\\
Thunderbird\\src-central\\mail\\test\\mozmill\\message-header\\test-message-head
er.js"])

Events disconnect Error: Application unexpectedly closed
TEST-PASS |  setupModule
make: *** [mozmill-one] Error 1

I'll work on it more tomorrow, but if anyone has any suggestions, I'ld love to hear them.

Later,
Blake.

Comment 7

[Blake Winton (:bwinton) (:☕️)]

(Oh, that's just running:
    make SOLO_TEST=message-header/test-message-header.js mozmill-one
 I'll try running:
    make mozmill
 tomorrow as well.
)

Comment 8

[Blake Winton (:bwinton) (:☕️)]

Attachment: A patch to fix one of the test failures on trunk.

So, it's failing for me now, too.

The problem line is <http://mxr.mozilla.org/comm-central/source/mail/test/mozmill/message-header/test-message-header.js#353>, which says:
mc.click(new elib.Elem(mc.menus.View.viewheadersmenu.viewallheaders));

And it's a problem because mc.menus.View.viewheadersmenu.viewallheaders.disabled is true.

And so here's a patch that calls the parent menus in the hierarchy, which enables this menu, which causes the test to pass.  Woo!  :)

Thanks,
Blake.

Comment 9

[Mark Banner (:standard8)]

Comment on attachment 459565 [details] [diff] [review]
A patch to fix one of the test failures on trunk.

Ok, this looks good. a=Standard8 to land on closed tree.

Comment 10

[Blake Winton (:bwinton) (:☕️)]

Pushed to comm-central as <http://hg.mozilla.org/comm-central/rev/2ba810a84fa6>.

Comment 11

[rsx11m]

TEST-PASS | test_show_all_header_mode on latest WINNT 5.2 comm-central check.

Comment 12

[Mark Banner (:standard8)]

Yeah, this seems fixed now.