order of execution of the install-exec-hook rule

[Vincent Torri]

Hey,

i want to use install-exec-hook to move a file. I install it with 
install-strip.


I would like to know if the install-exec-hook rule is called after 
install-strip. If yes, I have a problem :)


thank you

Vincent

Re: order of execution of the install-exec-hook rule

[Ralf Wildenhues]

Hi Vincent,

* Vincent Torri wrote on Tue, Apr 22, 2008 at 03:17:02PM CEST:
>
> i want to use install-exec-hook to move a file. I install it with  
> install-strip.
>
> I would like to know if the install-exec-hook rule is called after  
> install-strip.

Yes.  The install-exec-hook recipe is executed as part of
install-exec-am.  install-strip really calls 'make install'
with some variable settings changed.

> If yes, I have a problem :)

Why?

Cheers,
Ralf

Re: order of execution of the install-exec-hook rule

[Vincent Torri]

I would like to know if the install-exec-hook rule is called after
install-strip.


Yes.  The install-exec-hook recipe is executed as part of
install-exec-am.  install-strip really calls 'make install'
with some variable settings changed.


If yes, I have a problem :)


Why?


I have the following error message:

make  INSTALL_PROGRAM="/bin/sh /home/torri/tmp/cvsroot2/e17/libs/evas/install-sh -c 
-s" \
  install_sh_PROGRAM="/bin/sh 
/home/torri/tmp/cvsroot2/e17/libs/evas/install-sh -c -s" INSTALL_STRIP_FLAG=-s \
  `test -z 'arm-wince-cegcc-strip' || \
echo "INSTALL_PROGRAM_ENV=STRIPPROG='arm-wince-cegcc-strip'"` 
install
make[1]: entrant dans le répertoire 
Â/home/torri/tmp/cvsroot2/e17/libs/evas/src/modules/engines/software_16_wince »
make[2]: entrant dans le répertoire 
Â/home/torri/tmp/cvsroot2/e17/libs/evas/src/modules/engines/software_16_wince »
make  install-exec-hook
make[3]: entrant dans le répertoire  
/home/torri/tmp/cvsroot2/e17/libs/evas/src/modules/engines/software_16_wince Â
mv 
/home/torri/local/wince/lib/evas/modules/engines/software_16_wince/cegcc-arm/module.dll
 /home/torri/local/wince/bin/engine_buffer.dll
mv: ne peut évaluer`/home/torri/local/wince/lib/evas/modules/engines/software_16_wince/cegcc-arm/module.dll': 
Aucun fichier ou répertoire de ce type

make[3]: *** [install-exec-hook] Erreur 1

translation:

"entrant dans le répertoire: entering directory
"ne peut évaluer": can not evaluate
"Aucun fichier ou répertoire de ce type": no file or directory

The 'mv' call is done during the exec-hook. It seems that the file 
module.dll is not copied first in 
/home/torri/local/wince/lib/evas/modules/engines/software_16_wince/cegcc-arm/ 
by the install-strip rule


I thought that the install-strip rule strips the library and copied it 
where it should (or the other order: copy first, strip after). It sems 
that it is not copied.


Do you see why ?

Vincent

Re: order of execution of the install-exec-hook rule

[Ralf Wildenhues]

Apologies for mangled accents.  Thanks for the translations (though
my French isn't that bad yet).  If you want to save yourself work
translating to others, do 'LC_ALL make ...'.

* Vincent Torri wrote on Tue, Apr 22, 2008 at 07:36:41PM CEST:
>
> make  INSTALL_PROGRAM="/bin/sh 
> /home/torri/tmp/cvsroot2/e17/libs/evas/install-sh -c -s" \
>   install_sh_PROGRAM="/bin/sh 
> /home/torri/tmp/cvsroot2/e17/libs/evas/install-sh -c -s" 
> INSTALL_STRIP_FLAG=-s \
>   `test -z 'arm-wince-cegcc-strip' || \
> echo "INSTALL_PROGRAM_ENV=STRIPPROG='arm-wince-cegcc-strip'"` 
> install
> make[1]: entrant dans le répertoire 
> Â/home/torri/tmp/cvsroot2/e17/libs/evas/src/modules/engines/software_16_wince 
> »
> make[2]: entrant dans le répertoire 
> Â/home/torri/tmp/cvsroot2/e17/libs/evas/src/modules/engines/software_16_wince 
> »
> make  install-exec-hook
> make[3]: entrant dans le répertoire  
> /home/torri/tmp/cvsroot2/e17/libs/evas/src/modules/engines/software_16_wince Â
> mv 
> /home/torri/local/wince/lib/evas/modules/engines/software_16_wince/cegcc-arm/module.dll
>  /home/torri/local/wince/bin/engine_buffer.dll
> mv: ne peut 
> évaluer`/home/torri/local/wince/lib/evas/modules/engines/software_16_wince/cegcc-arm/module.dll':
>  
> Aucun fichier ou répertoire de ce type
> make[3]: *** [install-exec-hook] Erreur 1

> I thought that the install-strip rule strips the library and copied it  
> where it should (or the other order: copy first, strip after). It sems  
> that it is not copied.
>
> Do you see why ?

Maybe if you show us software_16_wince/Makefile.am?

Cheers,
Ralf

Re: order of execution of the install-exec-hook rule

[Vincent Torri]

On Tue, 22 Apr 2008, Ralf Wildenhues wrote:


Apologies for mangled accents.  Thanks for the translations (though
my French isn't that bad yet).  If you want to save yourself work
translating to others, do 'LC_ALL make ...'.

* Vincent Torri wrote on Tue, Apr 22, 2008 at 07:36:41PM CEST:


make  INSTALL_PROGRAM="/bin/sh /home/torri/tmp/cvsroot2/e17/libs/evas/install-sh -c 
-s" \
  install_sh_PROGRAM="/bin/sh 
/home/torri/tmp/cvsroot2/e17/libs/evas/install-sh -c -s" INSTALL_STRIP_FLAG=-s \
  `test -z 'arm-wince-cegcc-strip' || \
echo "INSTALL_PROGRAM_ENV=STRIPPROG='arm-wince-cegcc-strip'"` 
install
make[1]: entrant dans le répertoire 
Â/home/torri/tmp/cvsroot2/e17/libs/evas/src/modules/engines/software_16_wince »
make[2]: entrant dans le répertoire 
Â/home/torri/tmp/cvsroot2/e17/libs/evas/src/modules/engines/software_16_wince »
make  install-exec-hook
make[3]: entrant dans le répertoire  
/home/torri/tmp/cvsroot2/e17/libs/evas/src/modules/engines/software_16_wince Â
mv 
/home/torri/local/wince/lib/evas/modules/engines/software_16_wince/cegcc-arm/module.dll
 /home/torri/local/wince/bin/engine_buffer.dll
mv: ne peut
évaluer`/home/torri/local/wince/lib/evas/modules/engines/software_16_wince/cegcc-arm/module.dll':
Aucun fichier ou répertoire de ce type
make[3]: *** [install-exec-hook] Erreur 1



I thought that the install-strip rule strips the library and copied it
where it should (or the other order: copy first, strip after). It sems
that it is not copied.

Do you see why ?


Maybe if you show us software_16_wince/Makefile.am?


Here it is:


MAINTAINERCLEANFILES = Makefile.in

AM_CPPFLAGS = \
-I. \
-I$(top_srcdir)/src/lib \
-I$(top_srcdir)/src/lib/include \
-I$(top_srcdir)/src/modules/engines \
-I$(top_srcdir)/src/modules/engines/software_16 \
@FREETYPE_CFLAGS@

if BUILD_ENGINE_SOFTWARE_16_WINCE

pkgdir = $(libdir)/evas/modules/engines/software_16_wince/$(MODULE_ARCH)

pkg_LTLIBRARIES = module.la

module_la_SOURCES = \
evas_engine.h \
evas_engine.c \
evas_wince_fb_buffer.c \
evas_wince_gapi_buffer.c

module_la_CFLAGS = @WIN32_CFLAGS@
module_la_LIBADD = $(top_builddir)/src/lib/libevas.la
module_la_LDFLAGS = @lt_no_undefined@ -module -avoid-version
module_la_LIBTOOLFLAGS = --tag=disable-static
module_la_DEPENDENCIES = $(top_builddir)/config.h

include_HEADERS = Evas_Engine_Software_16_WinCE.h

install-exec-hook:
	mv 
$(libdir)/evas/modules/engines/software_16_wince/$(MODULE_ARCH)/module.dll 
$(bindir)/engine_buffer.dll


endif

CPPFLAGS for LIBOBJS

[Russ Allbery]

Today I ran into a need to set CPPFLAGS when building files for LIBOBJS
and discovered that there doesn't appear to be any way to do it without
setting CPPFLAGS globally.

I have:

portable_libportable_a_LIBADD = $(LIBOBJS)

If I add:

portable_libportable_a_CPPFLAGS = $(KRB5_LDFLAGS) $(KAFS_LDFLAGS)

it has no effect, which after thinking about it for a moment makes some
sense.  However, setting LIBOBJS_CPPFLAGS doesn't do anything either.

Could some facility to handle this be added?

-- 
Russ Allbery ([EMAIL PROTECTED]) 

Re: CPPFLAGS for LIBOBJS

[Russ Allbery]

Russ Allbery <[EMAIL PROTECTED]> writes:

> If I add:
>
> portable_libportable_a_CPPFLAGS = $(KRB5_LDFLAGS) $(KAFS_LDFLAGS)

Er, obviously that should be KRB5_CPPFLAGS and KAFS_CPPFLAGS.  Sorry about
the confusion.

> it has no effect, which after thinking about it for a moment makes some
> sense.  However, setting LIBOBJS_CPPFLAGS doesn't do anything either.
>
> Could some facility to handle this be added?

-- 
Russ Allbery ([EMAIL PROTECTED]) 

Dynamic AC_OUPUT

[Gary Kramlich]

I'm rewriting the build system for project that is using autotools.  The
problem is right now there are over 50 makefiles output which is
creating a bit of a maintenance headache.  As such I've moved to a XML
config file that does the heavy lifting of dependency checking for each
moudule (the 50 makefiles).

This python script also determines what directories should have
makefiles created and so on.  I've gotten far enough to trick configure
into outputting all of the make files.  (This is done by appending the
directories to ac_config_files before AC_OUTPUT).

I was working on this in an already configured tree, and thought I had
it with the ac_config_files hack.  However, turns out automake when
being run from my autogen doesn't like this.  I realize it's just
reading in the contents of AC_OUTPUT, and tried to work around this by
specifying the directories to the initial call of automake.  However, it
drops out every time refusing to create depmod.

So after that long winded explanation, is there any way I can get this
to work during autogen?  I realize this is really abusing automake, but
I'd prefer to stay with automake, but this project just can't keep it's
build system the way it is.

If examples are needed, please let me know so I can provide them.

Thanks,

-- 
Gary Kramlich <[EMAIL PROTECTED]>



signature.asc
Description: OpenPGP digital signature

Re: Dynamic AC_OUPUT

[Ralf Wildenhues]

Hello Gary,

* Gary Kramlich wrote on Wed, Apr 23, 2008 at 03:08:38AM CEST:
> This python script also determines what directories should have
> makefiles created and so on.  I've gotten far enough to trick configure
> into outputting all of the make files.  (This is done by appending the
> directories to ac_config_files before AC_OUTPUT).

Bad.  ac_config_files is not defined in the Autoconf manual.  Please use
AC_CONFIG_FILES([file ...]).

> I was working on this in an already configured tree, and thought I had
> it with the ac_config_files hack.  However, turns out automake when
> being run from my autogen doesn't like this.  I realize it's just
> reading in the contents of AC_OUTPUT, and tried to work around this by
> specifying the directories to the initial call of automake.  However, it
> drops out every time refusing to create depmod.

FYI, automake heavily uses something called "autoconf traces" which is
pretty much asking m4 "show me the args whenever the macros
AC_CONFIG_FILES, AC_INIT, ... are called".

> If examples are needed, please let me know so I can provide them.

Yes, an example would be good.

Cheers,
Ralf