1261602 - Image insertion broken. Regression from M-C editor bug 1257363.

Status: RESOLVED FIXED

(Thunderbird :: Message Compose Window, defect)

Description

[Jorg K (CEST = GMT+2)]

Can be reproduced locally:
mozmake SOLO_TEST=composition/test-image-insertion-dialog.js mozmill-one

Local error message:
Chrome file doesn't exist: c:\mozilla-source\comm-central\obj-i686-pc-mingw32\dir\icons\img-align-top.gif

Then it hangs for a while, then we get:

TEST-UNEXPECTED-FAIL | Disconnect Error: Application unexpectedly closed
Timeout: bridge.set("4821049e-f8e9-11e5-90d0-08606e868e3d", Components.utils.import('resource://mozmill/modules/mozmill.js'))

img-align-top.gif exists in the source tree but not in the object directory.

Comment 1

[Richard Marti (:Paenglab)]

Attachment: Bug1261602.patch

The two images where removed from jar.mn only from Windows in 2012. Adding them again.

Comment 2

[Jorg K (CEST = GMT+2)]

Richard, sorry, thanks for the patch, the images were missing and it's great to have them again. Would you mind moving the addition to another bug.

The test still fails:
TEST-UNEXPECTED-FAIL | Disconnect Error: Application unexpectedly closed
Timeout: bridge.set("00cc0bc0-f8f0-11e5-a585-08606e868e3d", Components.utils.import('resource://mozmill/modules/mozmill.js'))

So we noticed the missing images, but adding them doesn't fix the test.

Comment 3

[Richard Marti (:Paenglab)]

Comment on attachment 8737518 [details] [diff] [review]
Bug1261602.patch

Okay, I file a new bug.

Comment 4

[aleth [:aleth]]

The timeout happens on the acceptDialog() call on the insert image window. It might be good to check if there is also a problem when these UI steps are done by hand in a current TB build, or if it is a problem with the mozmill test run.

m-c regression window: bccb11375f2a to 538d248fa252

Comment 5

[Jorg K (CEST = GMT+2)]

If run be hand, it crashes straight away. STR:
Open a new composition window.
Click in the body text, then "Insert > Image".
Type name: doesntmatter.png, click "Don't use alternate text".
Press OK.
Crash!!

Comment 6

[Jorg K (CEST = GMT+2)]

This crashes so quickly and so badly, that the Windows debugger (VS 2013) doesn't catch it and I don't get a chance to see what's on the debug console either.

Comment 7

[:aceman]

Crashes on linux too, but there is no usual stack dump. That is strange.

Comment 8

[Jorg K (CEST = GMT+2)]

OK, WinDbg gets me there:
Crashes in nsHTMLEditRules::CheckForEmptyBlock.
Regressed by a bug I landed. Great. Bug 1257363.

Comment 9

[:aceman]

I mean the test crashes.
Manually doing steps from comment 5 does NOT cause a crash for me. A missing image frame is inserted into the message correctly.

But I get this in console:
Warning: ReferenceError: reference to undefined property window.outerWidth
Source File: chrome://editor/content/EdDialogCommon.js
Line: 545

Also, are the warnings about "inner window" harmless?
++DOMWINDOW == 39 (0xd6276c00) [pid = 5032] [serial = 39] [outer = (nil)]
[5032] WARNING: No inner window available!: file mozilla/dom/base/nsGlobalWindow.cpp, line 9776

Comment 10

[aleth [:aleth]]

(In reply to Jorg K (GMT+2) from comment #8)
> OK, WinDbg gets me there:
> Crashes in nsHTMLEditRules::CheckForEmptyBlock.
> Regressed by a bug I landed. Great. Bug 1257363.

Thanks for the quick diagnosis. Good job we have tests ;)

Comment 11

[aleth [:aleth]]

(In reply to :aceman from comment #9)
> Also, are the warnings about "inner window" harmless?
> ++DOMWINDOW == 39 (0xd6276c00) [pid = 5032] [serial = 39] [outer = (nil)]
> [5032] WARNING: No inner window available!: file
> mozilla/dom/base/nsGlobalWindow.cpp, line 9776

You'd have to look at exactly what's happening there...

Comment 12

[Jorg K (CEST = GMT+2)]

(In reply to :aceman from comment #9)
> Also, are the warnings about "inner window" harmless?
> ++DOMWINDOW == 39 (0xd6276c00) [pid = 5032] [serial = 39] [outer = (nil)]
> [5032] WARNING: No inner window available!: file
> mozilla/dom/base/nsGlobalWindow.cpp, line 9776
Hmm, good question, I see them all the time and they don't cause trouble. We should investigate that a bit.

Comment 13

[Jorg K (CEST = GMT+2)]

Fixed by reopening and fixing bug 1257363 again.

(In reply to :aceman from comment #9)
> Manually doing steps from comment 5 does NOT cause a crash for me. A missing
> image frame is inserted into the message correctly.
Most likely you didn't try in paragraph mode. The bug that caused it was about paragraphs.