analysis of libiberty/cp-demangle.c for libsupc++/__cxa_demangle

[Zack Weinberg]

Gabriel Dos Reis <gdr@integrable-solutions.net> writes:

> | > I don't think anybody had proposed actually removing the Wood/C++
> | > demangler.  Did I miss an email?
> | 
> | The only reason I've seen for keeping it is the user-selectable
> | allocator capability.  Has Ian now added that to the libiberty
> | demangler?
>
> So in the end, you're proposing the removal of Wood/C++-based
> demangler?  I need a clear answer, not a circumvoluted one.
> Mark has said that wasn't his intent -- even if he would not have
> choosed that.

I was considering that possibility *if* the C++ demangler did not
provide any additional functionality.  From further discussion, it
is clear that it does, so I am fine with keeping it.

zw