On 28/02/2013 16:34, Guillaume Lazzara wrote:
On 02/28/2013 04:16 PM, Roland Levillain wrote:
On 26/02/2013 16:36, Guillaume Lazzara wrote:
[...]
diff --git a/scribo/common.mk b/scribo/common.mk
index 874f313..1c5c57d 100644
--- a/scribo/common.mk
+++ b/scribo/common.mk
@@ -24,5 +24,5 @@ AM_DEFAULT_SOURCE_EXT = .cc
# Add paths to Scribo's headers.
AM_CPPFLAGS = -I$(top_srcdir)/ -I$(top_builddir)/ \
- -I$(top_srcdir)/milena -I$(top_builddir)/milena \
- -I$(top_srcdir)/scribo -I$(top_builddir)/scribo
+ -I$(abs_top_srcdir)/scribo -I$(abs_top_builddir)/scribo \
+ -I$(abs_top_srcdir)/milena -I$(abs_top_builddir)/milena
I also think this order is more logical (in scribo/common.mk). Did you
have to swap these paths because of a compilation problem or was it just
for stylistic purpose?
I had a problem with the include file data.hh which exists both in
milena/tests and scribo/tests.
This file is included in scribo tests using :
#include "tests/data.h"
Not swapping the include paths prevents g++ from including the right file.
Thanks. I didn't realize we had a name clash there. As it is not sound
to depend on the order of ``declarations'' (in a broad sense), I suggest
we rename those files.
Also the use
of `$(abs_top_srcdir)' shall not be needed here. Is it
really necessary?
I thought so too. But since the latest modifications in Makefile, it
does not work any more without an absolute directory (I don't know why).
This file is included in several subdirectories and relative paths where
computed from the path of common.mk and not from the path of the related
Makefile.
Any explanation of the problem would be welcome :)
This is weird. I'll try to have a look at it.
Thanks for the review.
No problem!