Re: [xml] -fexceptions

[Sean Middleditch <elanthis awesomeplay com>]

On Thu, 2003-04-03 at 15:26, Stefan Seefeld wrote:
> Graham Bennett wrote:
>
> > > shouldn't 'configure CC=g++' do the job just fine ?
> >
> >
> > Sounds reasonable, but I then get:
> >                                                                                 
> > checking for g++ option to accept ANSI C... no
> > ...
> > configure: error: Compiler not ANSI compliant
>
>
> ah, the joy of autoconf...:-/
>
> I remember running into problems with configure tests when done with
> a C++ compiler instead of C. The problems were related to tests not
> being correct C++ code. The autoconf m4 macros are full of them. :-(

Well, in defense of autoconf (why am I doing that?) you are supposed to
intentionally state you are using C++ and not C, since C++ is *not*
forwards compatible from C (as you've clearly observed).

One might try using libxml++ (http://libxmlplusplus.sf.net) - it seems a
good wrapper, tho I haven't used it much yet.  Just starting playing
around with using it to replace my own C++ wrapper in AweMUD.  Should
also note I don't touch exceptions (they being one of the few things
that are real to the "C++ is bloated and slow" fud), so I can't testify
on how well libxml++ handles those in parsers.

> Well, what I tried to point out was merely that you should build libxml2
> with a C++ compiler to let stack unwinding 'pass through'. It seems
> it's not as simple as it could to make the build system understand that.

It'd probably take less than half an hour to update the build system to
compile using CXX instead of CC.  Automake might need some tricky
coercing tho.  There's also no guarantee libxml itself is C++ safe
(forwards incompatibility thing and all).

> Regards,
>               Stefan
>
> _______________________________________________
> xml mailing list, project page  http://xmlsoft.org/
> xml gnome org
> http://mail.gnome.org/mailman/listinfo/xml