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

[Carlo Wood <carlo at alinoe dot com>]

On Wed, Feb 25, 2004 at 04:25:17PM +0100, Gabriel Dos Reis wrote:
>   Yes, you're entirely right and I think Carlo is wrong in accusing you.

Sorry, I couldn't believe that Benjamin was the one who initiated the
removal in his patch and jumped conclusions; because Benjamin is the
one who approached me in the first place in the past and has always
taken the standpoint like "no matter what libiberty does, I want to keep
your demangler in libstdc++".  Anyway, I mailed him in private about it
for clarification.

>   As I said earlier, we should not disguise the removal (if that were
> to happen) as a bug fix to make libsupc++ self-contained.  I'm of the
> opinion that the demangler has proven to be useful and is still useful
> (yes, it could be improved, but that is a different story).  I do not
> think we should take actions to discourage people to contribute.
> Even if I happened to disagree with Carlo on some technical aspects, I
> do not believe that the removal of its contributed demangler (which we
> actually asked for and acclamed!) is fair or the right thing to do.
> It is not like he has announced to no longer maintain the demangler
> and we're shortage of maintainers. 

Despite that I am the emotional kind of guy with high peaks, low and high ;)
I consider myself I mature and responsible.  I will keep maintaining the
demangler when it is kept.  Not that I think a lot of maintenance is needed
though.  If for example there comes a request to extend it to also support
a different type of demangling, then I am not sure if I will have time to
integrate that.

Anyway - sorry for my emotional reactions so far :(.  I've always taken
that standpoint that such mails are best never sent to mailinglist, but
I couldn't help myself *sigh*.  I'll be over it tomorrow, no matter what
is decided.

If it is decided to keep the demangler, then I surely hope it won't be
removed "tomorrow" (or in 3.5, or 3.6): I don't like code duplication,
and once it is a part of libstdc++ I will start to use it from there and
remove it from libcwd (as such providing a garantee for maintenance and
testing too).  You understand that I am reluctant to do this unless it
becomes a more or less permanent extension to libstdc++.

-- 
Carlo Wood <carlo@alinoe.com>