Bug 28929 – [HTML5] Support for HTMLTextAreaElement.textLength

• Component: Forms
• Status: RESOLVED FIXED

[Kent Tamura]

Need to support for HTMLTextAreaElement.textLength.

Corresponding Chromium issue: http://code.google.com/p/chromium/issues/detail?id=2435

[Kent Tamura]

Created attachment 38964 [details]
Proposed patch

run-webkit-tests and check-webkit-style passed.

[Kent Tamura]

Created attachment 38968 [details]
Proposed patch (rev.2)

The previous patch didn't have LayoutTests/fast/forms/resources/textarea-textlength.js.

[Eric Seidel (no email)]

I don't think we use "unsigned long" just "unsigned" in c++ code, but I can't remember which is used in idls.

[Kent Tamura]

Hmm,  it(In reply to comment #3)
> I don't think we use "unsigned long" just "unsigned" in c++ code, but I can't
> remember which is used in idls.


HTMLInputElement.maxLength:
  unsigned long in HTML5 IDL
  long in WebKit IDL
  int in WebKit C++

HTMLInputElement.size:
  unsigned long in HTML5 IDL
  unsigned long in WebKit IDL
  int in WebKit C++

HTMLInputElement.selectionStart and selectionEnd, HTMLTextAreaElement.cols, rows, selectionStart, and selectionEnd:
  unsigned long in HTML5 IDL
  long in WebKit IDL
  int in WebKit C++

Hmm, it's confusing ;-(
Anyway, int seems ok in C++ code.

[Kent Tamura]

Created attachment 38978 [details]
Proposed patch (rev.3)

- Changed the type of textLength() in C++ code: unsigned long -> int

[Eric Seidel (no email)]

That doesn't make sense to me.  It seems the other callsites are wrong.  if it's "unsigned" in the idl, making it signed in the c++ code seem wrong.

[Kent Tamura]

(In reply to comment #6)
> That doesn't make sense to me.  It seems the other callsites are wrong.  if
> it's "unsigned" in the idl, making it signed in the c++ code seem wrong.

I have checked modern code: DataGridColumnList.idl, FileList.idl, and HTMLVideoElement.idl.  Their "unsigned long" IDL attributes are implemented as "unsigned".

Ok, I'll change the patch to "unsgined."

[Kent Tamura]

Created attachment 38988 [details]
Proposed patch (rev.4)

int -> unsigned

[Eric Seidel (no email)]

Comment on attachment 38988 [details]
Proposed patch (rev.4)

LGTM.  Eventually we need to straighten out the rest of the unsigned/int mismatches. :(

[Eric Seidel (no email)]

Comment on attachment 38988 [details]
Proposed patch (rev.4)

Rejecting patch 38988 from commit-queue.  This patch will require manual commit.

WebKitTools/Scripts/build-webkit failed with exit code 1

[Kent Tamura]

Created attachment 39028 [details]
Proposed patch (rev.5)

I don't know the reason of the commit-queue failure.  Anyway, update the patch for the latest WebKit source, and confirmed run-webkit-tests passed.

[Eric Seidel (no email)]

Comment on attachment 39028 [details]
Proposed patch (rev.5)

Clearing flags on attachment: 39028

Committed r48038: <http://trac.webkit.org/changeset/48038>

[Eric Seidel (no email)]

All reviewed patches have been landed.  Closing bug.

[Eric Seidel (no email)]

I also do not know the reason.  I guess the build was broken but the bots hadn't noticed yet.