Re: __func__ and C++

[Matt Austern <austern at apple dot com>]

On Sep 3, 2003, at 12:12 PM, Paul Koning wrote:

> Leaving it alone for the most part makes sense. but I see no way to
> justify "~foo" being reported as "foo".  That's clearly a bug.

I wouldn't like to see major changes in __func__: at the moment it 
doesn't include a type signature or a mangled name or scope 
qualification, and users seem to be mostly happy with that.  But users 
aren't happy with the fact that foo::~foo()'s name is reported as 
"foo".  They think that's at best counterintuitive, and I have to 
agree.

I've poked through the code a bit, and I think the problem is that 
fname_as_string() just returns whatever was stored in DECL_NAME.  I 
think that's a bug.  It's not the right way of getting a useful 
human-readable string.  Sometimes you'll get lucky and find that you 
get one, sometimes you won't.  The documented way of getting a short 
human-readable representation of a decl's name is to use 
lang_hooks.decl_printable_name with verbosity 0.  I'll work up the 
patch and make sure it does something reasonable, and then submit it.

			--Matt

P.S.  Sorry about your getting this mail again, Nathan.  I tried to 
send it before, and just noticed that I sent it to you only instead of 
to the whole list.