740784 - Undo (Ctrl+z) in textarea adding a newline (\n) to the text

Status: RESOLVED FIXED

(Core :: DOM: Editor, defect)

Description

[Drew Waddell]

User Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:11.0) Gecko/20100101 Firefox/11.0
Build ID: 20120312181643

Steps to reproduce:

I have text in a textarea that is delete and then restored using Ctrl+z.  When the text is restored an mysterious \n is added to the end of the text.

Here is an example of the the bug:
http://jsfiddle.net/dwaddell/S9mAv/18/

I have tested this in IE and Chorme and it works properly.


Actual results:

If the text was "abc" it will end up being "abc\n".


Expected results:

The text should be restored as "abc".

Comment 1

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

Can you test this with a current Firefox Nightly build from nightly.mozilla.org and see if it still happens?

For me in Nightly, if I follow the steps in that jsfiddle, I see the following things logged to the console, and there is no \n in the textarea:
[23:52:45.290] 9 Delete Me @ http://fiddle.jshell.net/dwaddell/S9mAv/18/show/:33
--
[23:53:01.114] 0  @ http://fiddle.jshell.net/dwaddell/S9mAv/18/show/:41
--
[23:53:13.893] 11 Delete Me
 @ http://fiddle.jshell.net/dwaddell/S9mAv/18/show/:41

Comment 2

[Nickolay_Ponomarev]

I can reproduce on a nightly, here's a jquery-less testcase: http://jsfiddle.net/S9mAv/20/

Note also bug 471319 and bug 523441.

Comment 3

[Nickolay_Ponomarev]

Attachment: testcase

Comment 4

[Alice0775 White]

Regression window(m-c)
Works:
http://hg.mozilla.org/mozilla-central/rev/145c98d55ae1
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:8.0a1) Gecko/20110814 Firefox/8.0a1 ID:20110814030749
Fails:
http://hg.mozilla.org/mozilla-central/rev/2de3cff973b2
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:8.0a1) Gecko/20110814 Firefox/8.0a1 ID:20110814044816
Pushlog:
http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=145c98d55ae1&tochange=2de3cff973b2

Regression window(m-i)
Works:
http://hg.mozilla.org/integration/mozilla-inbound/rev/56eaf58dc5fb
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:8.0a1) Gecko/20110812 Firefox/8.0a1 ID:20110812113455
Fails:
http://hg.mozilla.org/integration/mozilla-inbound/rev/437f3eb67c18
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:8.0a1) Gecko/20110812 Firefox/8.0a1 ID:20110812125630
Pushlog:
http://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=56eaf58dc5fb&tochange=437f3eb67c18

Suspected:
437f3eb67c18	Graeme McCutcheon — Bug 483651 - Trailing <br> node not removed when it should be; r=ehsan

Comment 5

[(no longer active)]

The cause of this bug is that nsTextEditorState::mCachedValue contains the value with the trailing BR.  The value that the underlying editor has is actually correct.

Comment 6

[Graeme McCutcheon [:graememcc]]

Attachment: v1

Essentially the opposite of bug 483651. Post-undo, the trailing BR does not get the mozBR attribute re-instated, so it gets interpreted as a newline.

Green on try, and fixes another test that had been forced to rely on the broken behaviour.

Comment 7

[(no longer active)]

Comment on attachment 614715 [details] [diff] [review]
v1

Review of attachment 614715 [details] [diff] [review]:
-----------------------------------------------------------------

r=me with the comments below addressed.  Thanks for your patch!

::: editor/libeditor/text/nsTextEditRules.cpp
@@ +1109,5 @@
> +    // Check to see if the trailing BR is a former bogus node - this will have stuck
> +    // around if we previously morphed a trailing node into a bogus node
> +    nsCOMPtr<nsIContent> lastBR = do_QueryInterface(lastChild);
> +    if (!mEditor->IsMozEditorBogusNode(lastBR))
> +      return NS_OK;

Nit: please use braces.

@@ +1112,5 @@
> +    if (!mEditor->IsMozEditorBogusNode(lastBR))
> +      return NS_OK;
> +
> +    // Morph it back to a mozBR
> +    dom::Element* elem = lastBR->AsElement();

Nit: Please add MOZ_ASSERT(lastBR->IsElement()) before this line as a sanity check.

::: editor/libeditor/text/tests/test_bug740784.html
@@ +14,5 @@
> +  <script type="application/javascript" src="/tests/SimpleTest/SimpleTest.js"></script>
> +  <link rel="stylesheet" type="text/css" href="/tests/SimpleTest/test.css">
> +</head>
> +
> +<body onload="doTest();">

Please use waitForFocus as you're sending keyboard events.

Comment 8

[(no longer active)]

Comment on attachment 614715 [details] [diff] [review]
v1

Zero risk for mobile as the undo functionality does not exist there.

Comment 9

[Lukas Blakk [:lsblakk] use ?needinfo]

Comment on attachment 614715 [details] [diff] [review]
v1

[triage comment]
low/no risk to mobile.

Comment 10

[(no longer active)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/a7d54fc7d18b

Comment 11

[(no longer active)]

https://hg.mozilla.org/mozilla-central/rev/a7d54fc7d18b

Comment 12

[Graeme McCutcheon [:graememcc]]

I changed the test to waitForFocus, per comment 7:
https://hg.mozilla.org/integration/mozilla-inbound/rev/304cd5d5124b

The others should be taken care of Ms2ger's patch in bug 747346

Comment 13

[Ed Morley [:emorley]]

https://hg.mozilla.org/mozilla-central/rev/304cd5d5124b