Bug 72195 – [chromium] Expose mock scrollbars to window.internals

• Status: RESOLVED
• Resolution: FIXED
• Component: New Bugs

[Adrienne Walker]

[chromium] Expose mock scrollbars to window.internals

[Adrienne Walker]

Created attachment 114797 [details]
Patch

[WebKit Review Bot]

Please wait for approval from [e-mail] before submitting because this patch contains changes to the Chromium public API.

[Adrienne Walker]

I realize that some ports set mock scrollbars on all the time, but I felt like exposing a per-test setting seemed more flexible.

I'm hoping this patch will alleviate some of our scrollbar pixel difference woes.

[Gustavo Noronha (kov)]

Comment on attachment 114797 [details]
Patch

Attachment 114797 [details] did not pass gtk-ews (gtk):
Output: http://queues.webkit.org/results/10448591

[Adrienne Walker]

Created attachment 114810 [details]
Now with more exports

[Gustavo Noronha (kov)]

Comment on attachment 114810 [details]
Now with more exports

Attachment 114810 [details] did not pass gtk-ews (gtk):
Output: http://queues.webkit.org/results/10350700

[Adrienne Walker]

Created attachment 114841 [details]
Patch

[Adrienne Walker]

Created attachment 114860 [details]
Fourth try's a charm

[Adrienne Walker]

(In reply to comment #8)
> Created an attachment (id=114860) [details]
> Fourth try's a charm

This link error really makes no sense.  Why is WebKit.dll referencing Internals.h? Shouldn't that be contained within WebCoreTestSupport?

11>   Creating library C:\cygwin\home\buildbot\Webkit\WebKitBuild\Debug\lib\WebKit.lib and object C:\cygwin\home\buildbot\Webkit\WebKitBuild\Debug\lib\WebKit.exp
11>WebKit.exp : error LNK2001: unresolved external symbol "public: void __thiscall WebCore::Internals::setMockScrollbarsEnabled(class WebCore::Document *,bool,int &)" (?setMockScrollbarsEnabled@Internals@WebCore@@QAEXPAVDocument@2@_NAAH@Z)

[Darin Fisher (:fishd, Google)]

Comment on attachment 114860 [details]
Fourth try's a charm

View in context: https://bugs.webkit.org/attachment.cgi?id=114860&action=review

> Source/WebKit/chromium/public/WebSettings.h:107
> +    virtual void setMockScrollbarsEnabled(bool) = 0;

Chromium WebKit API changes LGTM

[Adam Roben (:aroben)]

(In reply to comment #9)
> (In reply to comment #8)
> > Created an attachment (id=114860) [details] [details]
> > Fourth try's a charm
> 
> This link error really makes no sense.  Why is WebKit.dll referencing Internals.h? Shouldn't that be contained within WebCoreTestSupport?

WebCoreTestSupport is a static library that is linked into DRT and WTR and calls various WebCore functions. WebCore (on Windows) is a static library that is linked into WebKit.dll.

For DRT/WTR to be able to call WebCore functions (via WebCoreTestSupport), those functions need to be exported from WebKit.dll.

You can fix your linker error by adding the appropriate mangled symbols to WebKit2.def. You can see there are already multiple Internals-related functions listed there toward the bottom.

[Adrienne Walker]

Created attachment 115179 [details]
Oops. Don't export the wrong symbol from WebKit2.def, causing a linker error in WebKit.dll

[Adrienne Walker]

(In reply to comment #11)
> (In reply to comment #9)
> > (In reply to comment #8)
> > > Created an attachment (id=114860) [details] [details] [details]
> > > Fourth try's a charm
> > 
> > This link error really makes no sense.  Why is WebKit.dll referencing Internals.h? Shouldn't that be contained within WebCoreTestSupport?
> 
> WebCoreTestSupport is a static library that is linked into DRT and WTR and calls various WebCore functions. WebCore (on Windows) is a static library that is linked into WebKit.dll.
> 
> For DRT/WTR to be able to call WebCore functions (via WebCoreTestSupport), those functions need to be exported from WebKit.dll.
> 
> You can fix your linker error by adding the appropriate mangled symbols to WebKit2.def. You can see there are already multiple Internals-related functions listed there toward the bottom.

Thanks for all the info and explanation.  :)

I think I see what happened.  I copied and pasted the wrong symbol (the caller, not the callee) into WebKit2.def, so WebKit.dll expected that function to exist when it was building.  Oops.

[Adrienne Walker]

Created attachment 115182 [details]
Rebaselined

[James Robinson]

Comment on attachment 115182 [details]
Rebaselined

R=me

[Adrienne Walker]

Committed r100545: <http://trac.webkit.org/changeset/100545>