Michael Meeks wrote:
The rtl_string2UString direction (OString -> OUString) is not that problematic, as usually there should not be translation problems.Right; and it's the same direction, we 'internConvert' from a sal_Char * -> an alloca'd buffer, then try to look it up in the intern hash. So - failure should be unlikely, so I didn't add the return.
So, yes, I made a complete contradicting fool of myself. ;) How about the following then?
While often there are no translation problems in the OString -> OUString direction, there nevertheless *can* be such problems, manifest in the RTL_TEXTTOUNICODE_FLAGS_xxx_ERROR flags. What use are those flags if you cannot tell whether an error occurred? That's where rtl_string2UString and the OUString(OString) ctor are flawed (see e.g. configmgr/qa/unit/performance.cxx:1.1.2.3 l. 84 for a consequence of that flaw). I'd wish that new API was not designed flawed from the start.
-Stephan --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
