On 03/14/2010 02:08 PM, Nikolay Sivov wrote:
http://bugs.winehq.org/show_bug.cgi?id=22020
Some applications don't bother to pass properly allocated BSTR as a ::loadXML() parameter, so we crash on invalid length taken from DWORD before passed pointer.
Some tests verifying this would be nice. Also you might as well rename bstr_to_utf8() to represent what it's doing.
Vitaliy.
On 3/14/2010 23:55, Vitaliy Margolen wrote:
On 03/14/2010 02:08 PM, Nikolay Sivov wrote:
http://bugs.winehq.org/show_bug.cgi?id=22020
Some applications don't bother to pass properly allocated BSTR as a ::loadXML() parameter, so we crash on invalid length taken from DWORD before passed pointer.
Some tests verifying this would be nice. Also you might as well rename bstr_to_utf8() to represent what it's doing.
Ok, tests will be in try2. Renaming doesn't make much sense cause formal parameter is still BSTR in loadXML, the fact I don't use BSTR related calls doesn't change a type. A helper is used only in this loadXML method.
Vitaliy.