Hi,
I'm looking into adding support for the apple airport extreme to scli, and while building scli 0.3.0 (from the testing subdirectory) i get:
configure: WARNING: netinet/if_ether.h: present but cannot be compiled configure: WARNING: netinet/if_ether.h: check for missing prerequisite headers? configure: WARNING: netinet/if_ether.h: see the Autoconf documentation configure: WARNING: netinet/if_ether.h: section "Present But Cannot Be Compiled" configure: WARNING: netinet/if_ether.h: proceeding with the preprocessor's result configure: WARNING: netinet/if_ether.h: in the future, the compiler will take precedence configure: WARNING: ## ------------------------------- ## configure: WARNING: ## Report this to the scli lists. ## configure: WARNING: ## ------------------------------- ##
from configure.
on my system (NetBSD/i386 3.0), i can include netinet/if_ether.h cleanly by including:
#include <net/if.h> #include <netinet/in.h>
beforehand.
P.S. did the anon cvs server for scli ever come back online? or are the cvs snapshots available anywhere?
On Mon, Feb 06, 2006 at 03:06:21PM +0000, Jasper Wallace wrote:
I'm looking into adding support for the apple airport extreme to scli, and while building scli 0.3.0 (from the testing subdirectory) i get:
configure: WARNING: netinet/if_ether.h: present but cannot be compiled configure: WARNING: netinet/if_ether.h: check for missing prerequisite headers? configure: WARNING: netinet/if_ether.h: see the Autoconf documentation configure: WARNING: netinet/if_ether.h: section "Present But Cannot Be Compiled" configure: WARNING: netinet/if_ether.h: proceeding with the preprocessor's result configure: WARNING: netinet/if_ether.h: in the future, the compiler will take precedence configure: WARNING: ## ------------------------------- ## configure: WARNING: ## Report this to the scli lists. ## configure: WARNING: ## ------------------------------- ##
from configure.
on my system (NetBSD/i386 3.0), i can include netinet/if_ether.h cleanly by including:
#include <net/if.h> #include <netinet/in.h>
beforehand.
So what is the apropriate fix? My guess is that the warning somehow comes from the AC_CHECK_HEADERS macro. Can you fix this problem by changing configure.in and send a patch?
P.S. did the anon cvs server for scli ever come back online? or are the cvs snapshots available anywhere?
We moved to subversion some time ago. You should have read access by going to:
https://svn.eecs.iu-bremen.de/svn/schoenw/src/scli https://svn.eecs.iu-bremen.de/svn/schoenw/src/gsnmp
/js
On Mon, 6 Feb 2006, Juergen Schoenwaelder wrote:
So what is the apropriate fix? My guess is that the warning somehow comes from the AC_CHECK_HEADERS macro. Can you fix this problem by changing configure.in and send a patch?
attached.
Looking through the autoconf docs there where some notes onnetinet/if_ether.h and darwin/macos x, so i did those too.
My flatmate has a macos laptop and we've got gentoo and ubuntu machines kicking around here too, so when he gets back i'll test it on those.
On Mon, 6 Feb 2006, Jasper Wallace wrote:
On Mon, 6 Feb 2006, Juergen Schoenwaelder wrote:
So what is the apropriate fix? My guess is that the warning somehow comes from the AC_CHECK_HEADERS macro. Can you fix this problem by changing configure.in and send a patch?
attached.
Looking through the autoconf docs there where some notes on netinet/if_ether.h and darwin/macos x, so i did those too.
ok,
the patch dosn't break gentoo.
on the Mac OS X front the patch works fine, but it looks as if the advice in the autoconf docs are now out of date.
Building on Mac OS X did highlight another issue - -lgmodule is hardcoded in scli/Makefile.am in scli_LDADD, the machine i built on didn't have a gmodule lib, but did have a gmodule-2.0 lib, so it didn't build
I guess the -lgmodule should be $(GLIB_LIBS) instead, or do something like:
PKG_CHECK_MODULES(gmodule-2.0, gmodule-2.0, [], [])
and then use whatever LIB thing you get out of that.
unfortunatly after that things broke badly, Mac OS X has a libreadline in /usr/lib thats a sumlink a libedit, which is a BSD licenced line editor like readline.
and libedit defines symbols like _UP which are also defined by ncurses, so things go downhill from there.
i gave up on mac os x at that point, i'll look at how the vairous package systems that run on macos x handle it.
on the airport support front, i've got basic reporting working.
On Tue, Feb 07, 2006 at 09:53:51PM +0000, Jasper Wallace wrote:
the patch dosn't break gentoo.
I have checked the patch into svn and I also fixed the -lgmodule problem.
unfortunatly after that things broke badly, Mac OS X has a libreadline in /usr/lib thats a sumlink a libedit, which is a BSD licenced line editor like readline.
[...]
i gave up on mac os x at that point, i'll look at how the vairous package systems that run on macos x handle it.
I use darwinports and I have the GNU tools including readline via darwinports. This seems to take precedence and kind of works for me. Perhaps someone has to figure out how to write Portfiles for gsnmp and scli to make things easier to use on Mac OS X...
/js
participants (2)
-
Jasper Wallace
-
Juergen Schoenwaelder