Re: demengler in libstdc++ is changed

[Carlo Wood <carlo at alinoe dot com>]

On Thu, Nov 13, 2003 at 10:05:54PM -0500, DJ Delorie wrote:
> The biggest user of libiberty's demangler, gdb, has made it clear that
> a C++-based demangler would not be acceptable.  I wish the "new"
> demangler had been written in C to start with, so we could have
> avoided all these problems in the first place.

I never wrote that demangler for libstdc++, or
libiberty or gdb.  I wrote it because I needed a
demangler for my own project that was nor GPL-ed,
because the GPL did not allow me to use the one
in libiberty (I need to be able to control it
allocation).

So, I wrote one in a way that provides an interface
that allows to pass an allocator - which uses templates
like every other C++ thing that needs control over
allocation: the project that I wrote it for is the
free, open source, C++ runtime debugging support
library, libcwd.

Then someone from libstdc++/gcc came to me and asked
if they could use the demangler - if I was willing to
donate it - release it under the GPL.  I said: No,
it will have to be the LGPL - so I can use it myself
too - and I would like your word that the interface
that allows to pass an arbitrary Allocator will be
accessible to the users - as a gcc extension.

Ever since, for months and months there have been 
endless discussions (this is not the first time)
about the difference between this demangler and the
one in libiberty/gdb, and whining that my demangler
should have been in C or at least that C++ is not
usable for them and it sucks to have two demanglers.
I must say that I am more then sick of that now.
If the libstdc++ people wouldn't been so mature (normal?)
to me till now I'd have started to regret I ever
donated this code in the first place.

-- 
Carlo Wood <carlo@alinoe.com>