Bug 74686 – [Forms] The "maxlength" attribute on "textarea" tag miscounts hard newlines

• Status: RESOLVED
• Resolution: FIXED
• Component: Forms

From Chromium http://crbug.com/104971

== Steps ==
1. Give a textarea tag a maxlength=10
2. Hit Enter key many times
3. Submit form containing textarea

== Result ==
10 CRLF pairs are submitted.

== Expected ==
5 CRLF pairs are submitted.

== Note ==
According to WHATWG, line break characters in DOM are LF, and are CRLF in submission value.

Created attachment 119578 [details]
Patch 1

[Kent Tamura]

Comment on attachment 119578 [details]
Patch 1

View in context: https://bugs.webkit.org/attachment.cgi?id=119578&action=review

> Source/WebCore/html/HTMLTextAreaElement.cpp:58
> +static unsigned computeLengthForSubmission(const String& str)
> +{
> +    unsigned count =  numGraphemeClusters(str);
> +    unsigned len = str.length();

We prefer complete words for variable names.  'str' 'len' are not acceptable.

> LayoutTests/ChangeLog:10
> +        * fast/forms/textarea/textarea-maxlength-newline-expected.txt: Added.
> +        * fast/forms/textarea/textarea-maxlength-newline.html: Added.

The new test is not needed.
Your change for textarea-maxlength.js covers the behavior change.

Created attachment 119584 [details]
Patch 2

[Kent Tamura]

Comment on attachment 119584 [details]
Patch 2

ok

[WebKit Review Bot]

Comment on attachment 119584 [details]
Patch 2

Clearing flags on attachment: 119584

Committed r103051: <http://trac.webkit.org/changeset/103051>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.

[Aaron Nance]

This issue needs to be reopened. All hard carriage returns are being counted as 2 against the maxlength in Safari. I've tested this against Safari 9 on Mac 10.10 and in mobile Safari on iOS 9.2.

[Zach Leatherman]

Per Boris Zbarsky’s comment on the Mozilla issue tracker, the specification was recently changed to make hard newlines represent a single character: https://bugzilla.mozilla.org/show_bug.cgi?id=702296#c41

So, I think this needs to be reopened?

Per research in https://github.com/scottjehl/Device-Bugs/issues/73 only Chrome and Safari are going against the new spec.

The original case report should be amended to read:

== Expected ==
10 CRLF pairs are submitted.