This is the mail archive of the
mailing list for the GCC project.
Re: Comments with -fpreprocessed
- To: Fergus Henderson <fjh at cs dot mu dot oz dot au>
- Subject: Re: Comments with -fpreprocessed
- From: Michael Matz <matzmich at cs dot tu-berlin dot de>
- Date: Tue, 29 May 2001 12:42:47 +0200 (MET DST)
- cc: Nathan Sidwell <nathan at codesourcery dot com>, Neil Booth <neil at daikokuya dot demon dot co dot uk>, <gcc at gcc dot gnu dot org>
On Fri, 25 May 2001, Fergus Henderson wrote:
> > Maybe we should allow -C to indicate that they are. I.e.
> > gcc -E -C foo.c > foo.i ;# generate with comments
> > gcc -c -C foo.i ;# accept with comments
> > should do what you want. Does that make sense?
> That sounds to me like a reasonable solution, so long as `-C'
> is mentioned in the error message that occurs if you do `gcc -c foo.i'
> and forget the `-C'.
This then is an argument for reverting Nathan's patch, as usability-wise
if the compiler is able to say "you need to do this", it can also itself
do "this", especially if "this" is "running gcc again with -C". It would
make me angry by a large amount if I ever would see this message. In
Nathans view it was a regression because it was'nt behaving the way it was
before, but I consider the old behaviour broken (I was disturbed multiple
times by cc1 not accepting comments), so it's a negative regression, or a
bug fix. (Well, it all depends on definition of "preprocessed"; does it
include comments or not; if it doesn't we can make "gcc -c a.i" mean:
"accept preprocessed source with optional comments"). So, I vote for