Re: Feedback on -iquote

[Carlo Wood <carlo at alinoe dot com>]

On Wed, Jun 02, 2004 at 04:42:43PM -0700, Mike Stump wrote:
> On Monday, May 31, 2004, at 05:33 AM, Carlo Wood wrote:
> >In other words, I don't like this change and I protest ;).
> 
> So, let me recap.  You object to the removal of -I-, as you want the 
> feature of not searching the directory containing the current file.  
> You don't mind that we add -iquote, just you can't use it.  Did I get 
> that right?

Yes

> If -I- were not deprecated, would you be happy?

That is what I intended to propose.  But read on.

> I think Zack will have to weigh in, as that idea was his, not my desire.
> 
> Have you tried the new compiler?  Does it work for you (ignoring extra 
> messages by inform)?

I am using gcc cvs version for the purpose of being ready for it when
it is released (my main project, the C++ debugging support library is
a likely candidate to be used by developers who also want to use
bleeding edge compilers - so I have to be *extra* early) (I also use
it so that I can serve as C++ tester and report bugs way before a new
release is made).

Therefore, as a result of the deprecation of -I-, this issue was seriously
discussed on libcwd-devel and a new design of how to deal with debug
declarations in various header files (of libcwd, of third party libraries
and of the end-applications) has been made that will work without the
use of -I-.

During this process I realized that it is actually my opinion that -I-
should not be used as a tool, but rather and at most as a workaround for
a given situation where a given installation otherwise cannot compile.
So, I am happy to have gotten rid of the need of -I- and will do so
even if it is not deprecated.

The question for you remains (assuming you agree with me on the above)
whether or not -iquote can be used to workaround all situation that -I-
can be a workaround for.  If not then it shouldn't be deprecated, but
when it can - then well, why not.

-- 
Carlo Wood <carlo@alinoe.com>