[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kfm-devel
Subject:    Re: Win32 Port of Safari
From:       Alan Gutierrez <ajglist () izzy ! net>
Date:       2003-01-10 12:54:40
[Download RAW message or body]

David Faure wrote:

 > +#if defined (WIN32_NATIVE)
 > +typedef int64_t DOMTimeStamp;
 > +#else
 >  typedef unsigned long long DOMTimeStamp;
 > +#endif
 >
 > Isn't there a way to have a global typedef for "long long"?
 > Hmm, since it's not a single word that might be hard.
 > Maybe we need longlong and ulonglong...

KHTML imports stdint.h which defines uint64_t. You can use that.

VC++ has no unsigned 64 bit integer, so this ugly #ifdef will have to stay.



 > +#if defined(WIN32_NATIVE)
 > +protected:
 > +#endif
 > For changes like this we need some explanation (where is the calling code
 > that needs this).

c:\codearea\khtml_win32\webcore\kwq\kwqkhtmlpart.h(124) :
     error C2876: 'KHTMLPart::xmlDocImpl' : not all overloads are accessible
c:\codearea\khtml_win32\webcore\khtml\khtml_part.h(133) :
     see declaration of 'KHTMLPart'

class KWQKHTMLPart : public KHTMLPart
{
public:

     ...

     using KHTMLPart::xmlDocImpl;

     ...
};

Apple's bad? Can't tell. Don't have time now to devine their intent right now.

I know this change should not be checked into KHTML. I submitted the diff for
discussion, not as a patch. Please don't anyone panic.



 > - -    bool hasPseudoStyle() const { return pseudoStyle; }
 > +    bool hasPseudoStyle() const { return pseudoStyle != NULL; } // AJG
 >
 > Seems the bool handling is very different :(((

For:

bool hasPseudoStyle() const { return pseudoStyle; }

Says VC++:

c:\codearea\khtml_win32\webcore\khtml\rendering\render_style.h(632) :
     warning C4800: 'class khtml::RenderStyle *const ' :
     forcing value to bool 'true' or 'false' (performance warning)

Seems reasonable to ask that you test for null rather than cast.

Alan Gutierrez

[prev in list] [next in list] [prev in thread] [next in thread]