alsa-driver-0.9.0rc3 patch

alsa-driver-0.9.0rc3 patch

[Ian Grant]

Hello ALSA people

Thanks for making ALSA.

Here is a patch for the toplevel Makefile to remove the $DESTDIR from the path 
of the System.map file where we check for its existence. This makes the 
existence check test the same file location as is used in the subsequent -F 
option to depmod. It works for me!

Thanks again,
Ian

PS. You will need to apply this manually - my mailer will most likely have 
broken the patch.

--- Makefile.orig	Thu Sep 12 16:09:18 2002
+++ Makefile	Thu Sep 12 16:09:52 2002
@@ -117,7 +117,7 @@
 		done \
 	fi
 
-ifeq ($(DESTDIR)$(CONFIG_SND_KERNELDIR)/System.map,$(wildcard 
$(DESTDIR)$(CONFIG_SND_KERNELDIR)/System.map))
+ifeq ($(CONFIG_SND_KERNELDIR)/System.map,$(wildcard 
$(CONFIG_SND_KERNELDIR)/System.map))
 SYSTEM_MAP_OPT = -F $(CONFIG_SND_KERNELDIR)/System.map
 endif
 

-- 
Ian Grant, Computer Lab., William Gates Building, JJ Thomson Ave., Cambridge
Phone: +44 1223 334420




-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf

Re: alsa-driver-0.9.0rc3 patch

[Takashi Iwai]

At Thu, 12 Sep 2002 16:23:56 +0100,
Ian Grant wrote:
> 
> Here is a patch for the toplevel Makefile to remove the $DESTDIR from the path 
> of the System.map file where we check for its existence. This makes the 
> existence check test the same file location as is used in the subsequent -F 
> option to depmod. It works for me!

$DESTDIR is necessary in path, too, because depmod is invoked with
"-b $DESTDIR" option.  the system-map path becomes relative to the
$DESTDIR, then. 

i think it would be much easier to stop invoking depmod if DESTDIR is
specified.  DESTDIR is used normally for building packages, and
obviously in such a case, depmod is not necessary.

(btw, the error code of depmod is ignored, so the error messages from
depmod don't matter.)


Takashi


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf

Re: alsa-driver-0.9.0rc3 patch

[Ian Grant]

> At Thu, 12 Sep 2002 16:23:56 +0100,
> Ian Grant wrote:
> > 
> > Here is a patch for the toplevel Makefile to remove the $DESTDIR from the path 
> > of the System.map file where we check for its existence. This makes the 
> > existence check test the same file location as is used in the subsequent -F 
> > option to depmod. It works for me!
> 
> $DESTDIR is necessary in path, too, because depmod is invoked with
> "-b $DESTDIR" option.  the system-map path becomes relative to the
> $DESTDIR, then. 

I don't think this is correct. The system map is loaded only at the time 
depmod is run and comes from the kernel build directory 
$(CONFIG_SND_KERNELDIR); its path is not modified by the -b option. The kernel 
build directory is an absolute path to some scratch space somewhere - a home 
directory perhaps. $DESTDIR is the place where the modules will be installed 
so there is no System.map under $(DESTDIR)$(CONFIG_SND_KERNELDIR), but in the 
Makefile that is what is tested before it is decided to supply the

       -F $(CONFIG_SND_KERNELDIR)/Sytem.map

option to depmod.

My patch brings the test in line with the option. I cannot see why these two 
things should be different.

> i think it would be much easier to stop invoking depmod if DESTDIR is
> specified.  DESTDIR is used normally for building packages, and
> obviously in such a case, depmod is not necessary.

It updates the $(DESTDIR)/lib/modules/x.x.x/modules.info which is useful. My 
package includes a kernel, the modules and the module dependencies.

> (btw, the error code of depmod is ignored, so the error messages from
> depmod don't matter.)

But they are not nice - there is an error message for every module.


-- 
Ian Grant, Computer Lab., William Gates Building, JJ Thomson Ave., Cambridge
Phone: +44 1223 334420




-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf

Re: alsa-driver-0.9.0rc3 patch

[Takashi Iwai]

At Thu, 12 Sep 2002 17:10:44 +0100,
Ian Grant wrote:
> 
> > At Thu, 12 Sep 2002 16:23:56 +0100,
> > Ian Grant wrote:
> > > 
> > > Here is a patch for the toplevel Makefile to remove the $DESTDIR from the path 
> > > of the System.map file where we check for its existence. This makes the 
> > > existence check test the same file location as is used in the subsequent -F 
> > > option to depmod. It works for me!
> > 
> > $DESTDIR is necessary in path, too, because depmod is invoked with
> > "-b $DESTDIR" option.  the system-map path becomes relative to the
> > $DESTDIR, then. 
> 
> I don't think this is correct. The system map is loaded only at the time 
> depmod is run and comes from the kernel build directory 
> $(CONFIG_SND_KERNELDIR); its path is not modified by the -b option. The kernel 
> build directory is an absolute path to some scratch space somewhere - a home 
> directory perhaps. $DESTDIR is the place where the modules will be installed 
> so there is no System.map under $(DESTDIR)$(CONFIG_SND_KERNELDIR), but in the 
> Makefile that is what is tested before it is decided to supply the
> 
>        -F $(CONFIG_SND_KERNELDIR)/Sytem.map
> 
> option to depmod.
> 
> My patch brings the test in line with the option. I cannot see why these two 
> things should be different.

ok, now it's fixed on cvs.

thanks,


Takashi


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf