This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
Re: libg++-2.8.1.3 and gcc-20000220 CVS
- To: "Martin v. Loewis" <martin at loewis dot home dot cs dot tu-berlin dot de>
- Subject: Re: libg++-2.8.1.3 and gcc-20000220 CVS
- From: Zack Weinberg <zack at wolery dot cumb dot org>
- Date: Thu, 16 Mar 2000 17:33:58 -0800
- Cc: gcc-bugs at gcc dot gnu dot org
- References: <200002201709.SAA08605@bolero.cs.tu-berlin.de> <20000220104913H.mitchell@codesourcery.com> <20000220124310.A13743@wolery.cumb.org> <flr9e7h8c0.fsf@poivre.cmla.ens-cachan.fr> <20000220220944.A14645@wolery.cumb.org> <200002210825.JAA00621@loewis.home.cs.tu-berlin.de>
On Mon, Feb 21, 2000 at 09:25:53AM +0100, Martin v. Loewis wrote:
> > But, -save-temps will either not give you an .i file (making it next
> > to useless) or the .i file won't have the same interpretation, when
> > fed back in, as the original source.
>
> What exactly is the difference in interpretation? If it is a matter
> whether column numbers are present or not, and whether originial token
> locations are present - I'd accept that difference.
Theoretically, it could make a huge difference to the debugging information,
for example if we generate accurate line info for code from macros.
> If something goes wrong in the integrated preprocessor that does not
> in case of preprocessed input, analysis will need all source files
> as-is. This is no different from a bug in the preprocessor today -
> preprocessed output won't help, either.
True.
zw