Re: PATCH: Support the new demangler written in C++

["Zack Weinberg" <zack at codesourcery dot com>]

Nathanael Nerode <neroden@twcny.rr.com> writes:

>>> Seventh, if we're doing this, is there a reason we still have 
>>> cplus-dem.o (the *old old* demangler), or am I missing something?
> You missed the point; that's the "old old" demangler, as opposed to 
> cp-demangle.c which is the "old" demangler (but is newer than 
> cplus-dem.c).  Surely we don't need *three* demanglers?

I will point out that the "old old" demangler is for mangling schemes
*other* than that of the new C++ ABI.  Neither the "old" demangler nor
the "new" demangler knows how to demangle anything but the new C++ ABI
scheme.

I don't know how much people still need the ability to demangle these
other schemes, since the whole world seems to be converging on the
genericized IA64 ABI (and there was much rejoicing) but it may not be
totally unnecessary yet.

Now, I am definitely of the opinion that, if this brand new spiffy
written-in-C++ demangler goes in, the old demangler (cp-demangle.c)
must be removed.  Whatever solution is eventually found, I'm not going
to be happy if it involves the user being forced to choose between
cp-demangle.c and its known bugs, and new-spiffy-demangler and its
build dependencies.  Hence, of your listed solutions, I do not like #7
(introducing a --with-broken-demangler option)...

zw