Failures in tests for obj-c++....

[Ziemowit Laski]

Well, unfortunately the patches you committed thus far did not solve 
the problems on Fedora Core 3. :-(...

On 8 Jun 2005, at 0.15, David Ayers wrote:

> Ziemowit Laski wrote:
>
>>
>> On 7 Jun 2005, at 14.19, David Ayers wrote:
>>
>>> Also we should be weary of relying on:
>>> #include "objc/runtime.h"            /* the kitchen sink */
>>> which will in turn:
>>> #include <objc/files.h>
>>
>>
>> Ah, good catch!  Is there any reason that the latter cannot be turned 
>> into
>>
>>   #include "files.h"
>>
>> instead?
>>
>
> Actually, I believe that would be correct.
>
> Would such a patch be pre-approved?

Actually, I'll try replacing _all_ occurrences of

     #include <objc/...>

with

     #include "..."

in the objc headers (via '%s/<objc\/\(.*\)>/"\1"/g' vim command), and 
then rebuild
to see if that helps, so why don't you hold off for now.  Thanks.

> PS-OT: This reminds me, I always found the following comment in the
> autoconf info pages confusing:
>
> "Use `#include <config.h>' instead
> of `#include "config.h"', and pass the C compiler a `-I.' option (or
> `-I..'; whichever directory contains `config.h').  That way, even if
> the source directory is configured itself (perhaps to make a
> distribution), other build directories can also be configured without
> finding the `config.h' from the source directory."
>
> the -I directives are not limited to system includes.  I see no gain in
> searching for config.h as a system header.

Me neither...

> Am I missing something or is there value in bringing this up on the
> autoconf list?

I really don't know much about the innards of the auto* tools, so 
perhaps
someone more knowledgable can chime in, and/or you can ask the auto* 
gurus
directly like you suggest.

--Zem
--------------------------------------------------------------
Ziemowit Laski                 1 Infinite Loop, MS 301-2K
Mac OS X Compiler Group        Cupertino, CA USA  95014-2083
Apple Computer, Inc.           +1.408.974.6229  Fax .5477