J.M. Maurer wrote:
> On Fri, 2007-08-24 at 16:14 -0400, Dominic Lachowicz wrote:
>> Aren't "uninitialized" UTF8Strings initially "empty"? What does this
>> patch do other than obfuscate the change history?
>
> Make valgrind happy...
>
> Dunno why exactly though, but maybe this is it: UT_UTF8Strings are
> indeed initially empty (their size is inited to 0). However, the 'data'
> member of the UTF8String's member variable pImpl (UT_UTF8Buffer thingy)
> is not initialized. Maybe valgrind found a way to get to that data
> member.
The default constructor *should* initialise to empty.... if it does not
it is a bug.
Hub
Received on Sun Aug 26 03:40:43 2007
This archive was generated by hypermail 2.1.8 : Sun Aug 26 2007 - 03:40:45 CEST