[Webkit-unassigned] [Bug 5241] Space and tab characters "sent" by an input method give totally different results than typing them directly

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=5241





--- Comment #27 from Ryosuke Niwa <rniwa at webkit.org>  2011-01-17 22:11:38 PST ---
(In reply to comment #26)
> > > Source/WebCore/editing/CompositeEditCommand.cpp:398
> > > +static inline bool isWhitespaces(const String& text)
> > 
> > How about containsOnlyWhitespace as in JavaScriptCore/wtf/text/StringImpl.h?
> > 
> Looks nice, So Replaced this with String::containsOnlyWhitespace()

Oh I didn't mean to replace it with String::containsOnlyWhitespace.  That function calls isASCIISpace and that function doesn't return true for a non-breaking space.  We should add a test for that.

> > If you make the above change, then we can move this optimization into CompositeEditCommand::rebalanceWhitespaceAt.  But I'm still not sure that'll make sense if we extended this function to walk surrounding text nodes though.
> > 
> I'd like to keep this code removed because it will pull back off-by-one tricks...

But doesn't it undo optimization though?

> > > Source/WebCore/page/EventHandler.cpp:2656
> > > -bool EventHandler::handleTextInputEvent(const String& text, Event* underlyingEvent, bool isLineBreak, bool isBackTab)
> > > +bool EventHandler::handleTextInputEvent(const String& text, Event* underlyingEvent, bool isLineBreak, bool isBackTab, bool isComposition)
> > 
> > I really don't like the fact this function takes 3! boolean arguments.  Would you mind cleaning that up in a separate patch?
> Agreed and filed Bug 52608. I'll do it sometime soon.

Great.  We should probably fix the bug 52608 before fixing this bug though.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.