Bug 48405 – [Windows] Textfields <input>s have different widths

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

[Kent Tamura]

Created attachment 71999 [details]
Test case

Windows-only (including Chromium-win).

http://crbug.com/60348 shows three different issues.

- type=search appearance is almost same as type=text on Windows, but its width is shorter.
  Probably this is intentional.  An older revision of RenderThemeWin.cpp had the following comment:
    // Override padding size to match AppKit text positioning.
# http://trac.webkit.org/changeset/42020/trunk/WebCore/rendering/RenderThemeWin.cpp

- type=color/email/tel/url is shorter because they don't have "padding:1px 0" in themeWin.css.

- Input types with spin-button is shorter.  The root cause is unknown yet.

[Tony Chang]

I think they're supposed to match the size of inputs in IE for web compatibility.  Ojan knows the details.

[Kent Tamura]

(In reply to comment #1)
> I think they're supposed to match the size of inputs in IE for web compatibility.  Ojan knows the details.

It's the reason why type=text on WebKit/Windows is different from WebKit/Mac, isn't it?
Anyway, non-search text field types should have same widths with type=text.

[Kent Tamura]

Created attachment 72163 [details]
Patch

[Kent Tamura]

(In reply to comment #3)
> Created an attachment (id=72163) [details]
> Patch

This patch will break some rendering tests.  I'll take care of them when I commit the patch.

[Kent Tamura]

(In reply to comment #3)
> Created an attachment (id=72163) [details]
> Patch

It won't change type=search width.

[Brian Weinstein]

Comment on attachment 72163 [details]
Patch

Which tests is this covered by? Will tests need to be rebaselined? Because if they are passing now, won't this change break them?

[Brian Weinstein]

(In reply to comment #6)
> (From update of attachment 72163 [details])
> Which tests is this covered by? Will tests need to be rebaselined? Because if they are passing now, won't this change break them?

Oh, saw your comment above, sorry for the noise.

[Kent Tamura]

Probably, only fast/forms/input-appearance-spinbutton-disabled-readonly.html will be affected.

[Adam Roben (:aroben)]

(In reply to comment #8)
> Probably, only fast/forms/input-appearance-spinbutton-disabled-readonly.html will be affected.

If that is the case, it sounds like we need more new tests that check the width of the other input types.

[Eric Seidel (no email)]

Yes, this is very much an Ojan question.

[Kent Tamura]

Comment on attachment 72163 [details]
Patch

Withdrawn.
I'll add new test.

[Kent Tamura]

Created attachment 76467 [details]
Patch 2

[Kent Tamura]

Created attachment 79901 [details]
Patch 3

rebased

[Dimitri Glazkov (Google)]

Comment on attachment 79901 [details]
Patch 3

ok.

[Kent Tamura]

Comment on attachment 79901 [details]
Patch 3

Clearing flags on attachment: 79901

Committed r76567: <http://trac.webkit.org/changeset/76567>

[Kent Tamura]

All reviewed patches have been landed.  Closing bug.

[WebKit Review Bot]

http://trac.webkit.org/changeset/76567 might have broken Leopard Intel Release (Tests)