This is the mail archive of the
mailing list for the GCC project.
Re: Experiment: gcc and doxygen
- From: Chris Lattner <sabre at nondot dot org>
- To: Zack Weinberg <zack at codesourcery dot com>
- Cc: <gcc at gcc dot gnu dot org>
- Date: Sat, 16 Feb 2002 22:03:52 -0600 (CST)
- Subject: Re: Experiment: gcc and doxygen
> > For example from http://www.nondot.org/gcc/ssa_8c.html, we can easily see
> > that it is not neccesary for ssa.c to #include system.h, partition.h,
> > valarray.h, or sbitmap.h directly.
> Please be careful when making assertions like this. It's a policy
> decision that every .c file shall include config.h and system.h. More
> generally, having direct includes for things that are directly
> referenced means that the code is more robust against changes in
> header files.
Certainly, I just wasn't aware of that policy. In that case, perhaps
varray.h shouldn't included system.h. Regardless, I was just hoping that
the graphs could help visualization of the source, with one possible
application of cleaning up #include's.