Re: Q on functions as first class values

["Michael Newberry" <mnewberry@...>]

Christian,

Yes! There is an implicit conversion to a char* since the function is 
declared as returning LPCTSTR. That Microsoftism translates as "Long pointer 
to a constant C character string, using wide-char if enabled". So I did 
*exactly* what you said---I forgot about the implicit conversion of the 
CString, so the temporary char string got lost. As I said, Ugh!

Michael

----- Original Message ----- 
From: "Christian Vogler" <Christian.Vogler@gallaudet.edu>
To: "Lua list" <lua@bazar2.conectiva.com.br>
Sent: Wednesday, February 02, 2005 1:20 PM
Subject: Re: Q on functions as first class values


> On Wednesday 02 February 2005 03:10 pm, Michael Newberry wrote:
> > I think I just recognized my own problem  (cause by the other problem:
> > "coding + coffee + sleep = constant"). I violated a rule I preach all the
> > time, about not pushing volatile things on the stack!!! That CString s is
> > not persistent. Ugh! Sorry to waste people's time!
>
> Don't be too hard on yourself ... I think this example rather demonstrates 
> the
> dangers of a class design that does unsafe implicit type conversions. Is
> CString a class with an implicit conversion to char *?  If yes, that is 
> where
> the real problem lies.
>
> (I assume that char * is what LPCTSTR is supposed to be? - I am no longer 
> up
> to speed with respect to Microsoft's weird naming convention)
>
> Regards
> - Christian