public inbox for gcc@gcc.gnu.org
 help / color / mirror / Atom feed
* strange g++ -D behavior,
@ 2003-03-05 17:29 Michael Meeks
  2003-03-05 19:20 ` Michael Meeks
  0 siblings, 1 reply; 2+ messages in thread
From: Michael Meeks @ 2003-03-05 17:29 UTC (permalink / raw)
  To: gcc; +Cc: Martin Hollmichel - Sun Germany - ham02 - Hamburg

Hi,

	This may be a FAQ - if so apologies; I seem to be getting -DFoo
substitution in some include paths.

	I have a plain gcc-3.2.2 build, on prefix:

		/home/michael/ximian-desktop/ooo/BUILD/ooo/

	I build OpenOffice.org in a sub-prefix - which goes fine building C
programs, but when I hit the first c++ program, I get a most curious
error:

[snip]
from
/home/michael/ximian-desktop/ooo/BUILD/ooo/OOO_1_0_2/xml2cmp/source/xcd/main.cxx:62:
/home/michael/ximian-desktop/ooo/BUILD/ooo/OOO_1_0_2/solver/641/unxlngi4.pro/inc/stl/cstddef:35:46:
/home/michael/ximian-desktop/ooo/7663/ooo/include/c++/3.2.2/cstddef: No
                               **^^^^**
such file or directory
[snip]

	The issue being that the first path is correct - cstddef is there;
however - the 2nd path has a 's/BUILD/7663/' substitution done on it -
for reasons unexplained.

	Creating a symlink from 'BUILD' -> '7663' results in a clean build. The
final fishy smell is that one of the pre-processor options is
'-DBUILD=7663' - which makes me suspicious of an ignorance problem on my
part, or an overly enthusiastic substitution problem somewhere inside
c++.

	Fuller compiler output:
		http://primates.ximian.com/~michael/gcc-build-err.txt

	Grokking the pre-processing output (using -E) reveals that the paths
are correct coming out of the pre-processor; and it seems that this is
an issue isolated to c++ - could I be falling over a subtle feature ?

	Help appreciated,

	Thanks,

		Michael.

-- 
 mmeeks@gnu.org  <><, Pseudo Engineer, itinerant idiot

^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: strange g++ -D behavior,
  2003-03-05 17:29 strange g++ -D behavior, Michael Meeks
@ 2003-03-05 19:20 ` Michael Meeks
  0 siblings, 0 replies; 2+ messages in thread
From: Michael Meeks @ 2003-03-05 19:20 UTC (permalink / raw)
  To: gcc
  Cc: Martin Hollmichel - Sun Germany - ham02 - Hamburg,
	Federico Mena Quintero


On Wed, 2003-03-05 at 17:16, Michael Meeks wrote:
> 	This may be a FAQ - if so apologies; I seem to be getting -DFoo
> substitution in some include paths.

	I'm a moron - please ignore me; the problem lies in stlport; wherein it
seems to do:

#define _STLP_NATIVE_INCLUDE_PATH \
	/home/michael/ximian-desktop/ooo/BUILD/ooo/include/c++/3.2.2

	And proceeds to do things like:

# define _STLP_MAKE_HEADER(path, header) <path/header>
# define _STLP_NATIVE_CPP_C_HEADER(header)  \
	_STLP_MAKE_HEADER(_STLP_NATIVE_INCLUDE_PATH,header)

	with:

#include _STLP_NATIVE_CPP_C_HEADER(foo)

	All of which is too confusing for those if little brain,

	Apologies,

		Michael.	

-- 
 mmeeks@gnu.org  <><, Pseudo Engineer, itinerant idiot

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2003-03-05 19:11 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2003-03-05 17:29 strange g++ -D behavior, Michael Meeks
2003-03-05 19:20 ` Michael Meeks

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).