Re: [cxx-abi-dev] C++0x: mangling of char16_t and char32_t

[Sean Perry <perry@xxxxxxxxxx>]

What are the overload rules for char32_t and wchar_t?  If foo(char32_t) and
foo(wchar_t) are unique then you can't use the same mangling.
--
Sean Perry
Compiler Development
IBM Canada Lab
(905)-413-6031 (tie 313-6031), fax (905)-413-4839



                                                                           
             Dennis Handly                                                 
             <dhandly@xxxxxxxx                                             
             om>                                                        To 
                                       cxx-abi-dev@xxxxxxxxxxxxxxxx,       
             03/25/2008 10:29          doug.gregor@xxxxxxxxx               
             PM                                                         cc 
                                                                           
                                                                   Subject 
                                       Re:  [cxx-abi-dev] C++0x: mangling  
                                       of char16_t and char32_t            
                                                                           
                                                                           
                                                                           
                                                                           
                                                                           
                                                                           




>From: "Doug Gregor" <doug.gregor@xxxxxxxxx>
>We don't have a mangling for the C++0x char16_t or char32_t types. It
>has been suggested that we use 'k' for char16_t and 'q' for char32_t.
>Does that seem reasonable?

I was thinking for HP-UX, char32_t would be the same as wchar_t.
If we need a mangling that treats it differently, we probably need to
add some rationale words saying why we have w, c, a, h and now k/q.

>From: David Vandevoorde <daveed@xxxxxxx>
>Fine by me, although I'd prefer something a little more "mnemonical";
>maybe W2 and W4.

I would think we would want to use something like W# just so we preserve
the one letter combinations.