Suspicious code in revision 867140

Maksim Orlovich mo85 at
Fri Oct 3 15:57:39 BST 2008

> Hi,
> here are the reports for revision 864329 that had no feedback:

Thanks for posting it again, looks like we missed some.

> - kdelibs/kjs/string_object.cpp:408
> The local variable global is never changed.

This one is a false positive --- the loop will get exited via break on
line 370 if global is true (and if it's not, it'll only run once, as

> - kdelibs/khtml/khtmlview.cpp:2303
>   kdelibs/khtml/html/html_documentimpl.cpp:195
> Add a todo comment about the 0 to not forget it.

The second one has a corresponding bug report. Dunno about the first one

> - kdelibs/khtml/khtml_part.cpp:4282
> Add a todo comment about the 0 to not forget it. Does this mean that
> certificate
> chains are always valid in khtml?

Err, WTF? Gotta chase down the SSL guy looks really bad.

> - kdelibs/khtml/xml/dom_nodeimpl.cpp:333
> r_ofs and outsideEnd are used but never assigned.

The method is pretty much busted..

> - kdelibs/khtml/rendering/render_layer.cpp:894
> Line 887 indicates that m_hBar can be false. If line 893 is true, then a
> crash
> follows here.

overflow: scroll should always result in the scrollbars being there.

> - kdelibs/khtml/css/css_valueimpl.cpp:1314
> If line 1411 is false then docLoader is NULL here.

Will leave this one to Spart...

> - kdelibs/khtml/ecma/kjs_binding.cpp:336
>   kdelibs/khtml/ecma/kjs_binding.cpp:353
> The allowed code range [DOM::RangeException::_EXCEPTION_OFFSET,
> DOM::RangeException::_EXCEPTION_MAX} is wider than table
> rangeExceptionNames.

True, but latent (those values can't be produced). Will fix.


More information about the kde-core-devel mailing list