1
1
openmpi/autogen.sh

1301 строка
39 KiB
Bash
Исходник Обычный вид История

#! /usr/bin/env bash
2003-11-22 19:36:58 +03:00
#
# Copyright (c) 2004-2006 The Trustees of Indiana University and Indiana
# University Research and Technology
# Corporation. All rights reserved.
# Copyright (c) 2004-2005 The University of Tennessee and The University
# of Tennessee Research Foundation. All rights
# reserved.
# Copyright (c) 2004-2005 High Performance Computing Center Stuttgart,
# University of Stuttgart. All rights reserved.
# Copyright (c) 2004-2005 The Regents of the University of California.
# All rights reserved.
# Copyright (c) 2007-2008 Cisco, Inc. All rights reserved.
# $COPYRIGHT$
#
# Additional copyrights may follow
#
# $HEADER$
2003-11-22 19:36:58 +03:00
#
# This script is run on developer copies of Open MPI -- *not*
2003-11-22 19:36:58 +03:00
# distribution tarballs.
#set -x
##############################################################################
#
# User-definable parameters (search path and minimum supported versions)
#
# Note: use ';' to separate parameters
##############################################################################
ompi_aclocal_search="aclocal"
if test ! -z "$ACLOCAL"; then
ompi_aclocal_search="$ACLOCAL"
fi
ompi_autoheader_search="autoheader"
if test ! -z "$AUTOHEADER"; then
ompi_autoheader_search="$AUTOHEADER"
fi
ompi_autoconf_search="autoconf"
if test ! -z "$AUTOCONF"; then
ompi_autoconf_search="$AUTOCONF"
fi
ompi_autom4te_search="autom4te"
if test ! -z "$AUTOM4TE"; then
ompi_autom4te_search="$AUTOM4TE"
fi
ompi_libtoolize_search="libtoolize;glibtoolize"
if test ! -z "$LIBTOOLIZE"; then
ompi_libtoolize_search="$LIBTOOLIZE"
fi
ompi_automake_search="automake"
if test ! -z "$AUTOMAKE"; then
ompi_automake_search="$AUTOMAKE"
fi
ompi_automake_version="1.9.6"
ompi_autoconf_version="2.59"
ompi_libtool_version="1.5.22"
##############################################################################
2003-11-22 19:36:58 +03:00
#
# Global variables - should not need to modify defaults
2003-11-22 19:36:58 +03:00
#
##############################################################################
2003-11-22 19:36:58 +03:00
ompi_aclocal_version="$ompi_automake_version"
ompi_autoheader_version="$ompi_autoconf_version"
ompi_libtoolize_version="$ompi_libtool_version"
ompi_autom4te_version="$ompi_autoconf_version"
# program names to execute
ompi_aclocal=""
ompi_autoheader=""
ompi_autoconf=""
ompi_libtoolize=""
ompi_automake=""
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
mca_no_configure_components_file="config/mca_no_configure_components.m4"
mca_no_config_list_file="mca_no_config_list"
mca_no_config_env_file="mca_no_config_env"
mca_m4_include_file="config/mca_m4_config_include.m4"
* Rewrite ompi_mca.m4 to use m4_defined lists of projects (ompi, orte, etc.), frameworks, and components without configure scripts instead of hard-coded shell variables (for projects and frameworks) and shell variable building (for components). * Add 3rd category of component configuration (in addition to configure scripts and no-configured components): configure.m4 components. These components can only be built as part of OMPI (like no-configure), but can provide an m4 file that is run as part of the main configure script. These macros can set whether the component should be built, along with just about any other configuration wanted. More care must be taken compared to configure components, as doing things like setting variables or calling AC_MSG_ERROR now affects the top-level configure script (so calling AC_MSG_ERROR if your component can't configure probably isn't what you want) * Added support to autogen.sh for the configure.m4-style components, as well as building up the m4_define lists ompi_mca.m4 now expects * Updated a number of macros to be more config.cache friendly (both so that config.cache can be used and so the test can be quickly run multiple times in the same configrue script): - ompi_config_asm - c_weak_symbols - c_get_alignment * Added new macros to be shared when configuring components: - ompi_objc.m4 (this actually provides AC_PROG_OBJC - don't ask...) - ompi_check_xgrid - ompi_check_tm - ompi_check_bproc * Updated a number of components to use configure.m4 instead of configure.stub - btl portals - io romio - tm ras and pls - bjs, lsf_bproc ras and bproc_seed pls - xgrid ras and pls - null iof (used by tm) This commit was SVN r6412.
2005-07-09 22:52:53 +04:00
mca_m4_config_env_file="mca_m4_config_env"
autogen_subdir_file="autogen.subdirs"
topdir_file="opal/include/opal_config_bottom.h"
if echo a | grep -E '(a|b)' >/dev/null 2>&1; then
egrep='grep -E'
else
egrep=egrep
fi
############################################################################
#
# Version check - does major,minor,release check (hopefully ignoring
# beta et al)
#
# INPUT:
# - minimum version allowable
# - version we found
#
# OUTPUT:
# - 0 version is ok
# - 1 version is not ok
#
# SIDE EFFECTS:
# none
#
##############################################################################
check_version() {
local min_version="$1"
local version="$2"
min_version=`echo $min_version | sed -e 's/[A-Za-z]//g'`
version=`echo $version | sed -e 's/[A-Za-z]//g'`
local min_major_version="`echo $min_version | cut -f1 -d.`"
local min_minor_version="`echo $min_version | cut -f2 -d.`"
local min_release_version="`echo $min_version | cut -f3 -d.`"
if test "$min_release_version" = "" ; then
min_release_version=0
fi
local major_version="`echo $version | cut -f1 -d.`"
local minor_version="`echo $version | cut -f2 -d.`"
local release_version="`echo $version | cut -f3 -d.`"
if test "$release_version" = "" ; then
release_version=0
fi
if test $min_major_version -lt $major_version ; then
return 0
elif test $min_major_version -gt $major_version ; then
return 1
fi
if test $min_minor_version -lt $minor_version ; then
return 0
elif test $min_minor_version -gt $minor_version ; then
return 1
fi
if test $min_release_version -gt $release_version ; then
return 1
fi
return 0
}
##############################################################################
#
# find app - find a version of the given application that is new
# enough for use
#
# INPUT:
# - name of application (eg aclocal)
2003-11-22 19:36:58 +03:00
#
# OUTPUT:
# none
2003-11-22 19:36:58 +03:00
#
# SIDE EFFECTS:
# - sets application_name variable to working executable name
# - aborts on error finding application
#
##############################################################################
find_app() {
local app_name="$1"
local version="0.0.0"
local min_version="99.99.99"
local found=0
local tmpIFS=$IFS
eval "min_version=\"\$ompi_${app_name}_version\""
eval "search_path=\"\$ompi_${app_name}_search\""
IFS=";"
for i in $search_path ; do
IFS="$tmpIFS"
version="`${i} --version 2>&1`"
if test "$?" != 0 ; then
IFS=";"
continue
fi
version="`echo $version | cut -f2 -d')'`"
version="`echo $version | cut -f1 -d' '`"
if check_version $min_version $version ; then
eval "ompi_${app_name}=\"${i}\""
eval "ompi_${app_name}_found_version=$version"
found=1
break
fi
done
IFS="$tmpIFS"
if test "$found" = "0" ; then
2003-11-22 19:36:58 +03:00
cat <<EOF
I could not find a recent enough copy of ${app_name}.
I am gonna abort. :-(
2003-11-22 19:36:58 +03:00
Please make sure you are using at least the following versions of the
GNU tools:
GNU Autoconf $ompi_autoconf_version
GNU Automake $ompi_automake_version
NOTE: You may need Automake 1.8.5 (or higher) in order to run
"make dist" successfully
GNU Libtool $ompi_libtool_version
2003-11-22 19:36:58 +03:00
EOF
exit 1
2003-11-22 19:36:58 +03:00
fi
}
##############################################################################
#
# run_and_check - run the right GNU tool, printing warning on failure
#
# INPUT:
# - name of application (eg aclocal)
# - program arguments
2003-11-22 19:36:58 +03:00
#
# OUTPUT:
# none
2003-11-22 19:36:58 +03:00
#
# SIDE EFFECTS:
# - aborts on error running application
#
##############################################################################
2003-11-22 19:36:58 +03:00
run_and_check() {
local rac_progs="$*"
echo "$indent[Running] $rac_progs"
2003-11-22 19:36:58 +03:00
eval $rac_progs
if test "$?" != 0; then
cat <<EOF
-------------------------------------------------------------------------
It seems that the execution of "$rac_progs" has failed. See above for
the specific error message that caused it to abort.
-------------------------------------------------------------------------
2003-11-22 19:36:58 +03:00
EOF
exit 1
fi
}
##############################################################################
#
# find_and_delete - look for standard files in a number of common places
# (e.g., ./config.guess, config/config.guess, dist/config.guess), and
# delete it. If it's not found there, look for AC_CONFIG_AUX_DIR in
# the configure.in script and try there. If it's not there, oh well.
2003-11-22 19:36:58 +03:00
#
# INPUT:
# - file to delete
2003-11-22 19:36:58 +03:00
#
# OUTPUT:
# none
#
# SIDE EFFECTS:
# - files may disappear
#
##############################################################################
2003-11-22 19:36:58 +03:00
find_and_delete() {
local fad_file="$1"
local fad_cfile
local auxdir
2003-11-22 19:36:58 +03:00
# Look for the file in "standard" places
if test -f $fad_file; then
rm -f $fad_file
elif test -d config/$fad_file; then
rm -f config/$fad_file
elif test -d dist/$fad_file; then
rm -f dist/$fad_file
else
# Didn't find it -- look for an AC_CONFIG_AUX_DIR line in
# configure.[in|ac]
if test -f configure.in; then
fad_cfile=configure.in
elif test -f configure.ac; then
fad_cfile=configure.ac
else
echo "--> Errr... there's no configure.in or configure.ac file!"
fi
if test -n "$fad_cfile"; then
auxdir="`grep AC_CONFIG_AUX_DIR $fad_cfile | $egrep -v '^dnl' | cut -d\( -f 2 | cut -d\) -f 1`"
fi
2003-11-22 19:36:58 +03:00
if test -f "$auxdir/$fad_file"; then
rm -f "$auxdir/$fad_file"
fi
fi
}
##############################################################################
#
# run_gnu_tools - run the GNU tools in a given directory
#
# INPUT:
# - OMPI top directory
#
# OUTPUT:
# none
#
# SIDE EFFECTS:
# - assumes that the directory is ready to have the GNU tools run
# in it (i.e., there's some form of configure.*)
# - may preprocess the directory before running the GNU tools
# (e.g., generale Makefile.am's from configure.params, etc.)
#
##############################################################################
run_gnu_tools() {
rgt_ompi_topdir="$1"
# Sanity check to ensure that there's a configure.in or
# configure.ac file here, or if there's a configure.params
# file and we need to run make_configure.pl.
if test -f configure.params -a -f configure.stub -a \
-x "$rgt_ompi_topdir/config/mca_make_configure.pl"; then
cat <<EOF
*** Found configure.stub
*** Running mca_make_configure.pl
EOF
"$rgt_ompi_topdir/config/mca_make_configure.pl" \
--ompidir "$rgt_ompi_topdir" \
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
--componentdir "`pwd`"
if test "$?" != "0"; then
echo "*** autogen.sh failed to complete!"
exit 1
fi
# If we need to make a version header template file, do so
rgt_abs_dir="`pwd`"
rgt_component_name=`basename "$rgt_abs_dir"`
rgt_component_type=`dirname "$rgt_abs_dir"`
rgt_component_type=`basename "$rgt_component_type"`
rgt_ver_header="$rgt_abs_dir/$rgt_component_type-$rgt_component_name-version.h"
rgt_ver_header_base=`basename "$rgt_ver_header"`
make_version_header_template "$rgt_ver_header_base" "$rgt_component_type" "$rgt_component_name"
happy=1
file=configure.ac
elif test -f configure.in; then
happy=1
file=configure.in
elif test -f configure.ac; then
happy=1
file=configure.ac
else
cat <<EOF
--> Err... there's no configure.in or configure.ac file in this directory
--> Confused; aborting in despair
EOF
exit 1
fi
unset happy
# Find and delete the GNU helper script files
find_and_delete config.guess
find_and_delete config.sub
find_and_delete depcomp
find_and_delete compile
find_and_delete install-sh
find_and_delete ltconfig
find_and_delete ltmain.sh
find_and_delete missing
find_and_delete mkinstalldirs
find_and_delete libtool
# Run the GNU tools
echo "*** Running GNU tools"
if test -f $topdir_file ; then
cd config
run_and_check $ompi_autom4te --language=m4sh ompi_get_version.m4sh -o ompi_get_version.sh
cd ..
fi
# We only need the libltdl stuff for the top-level
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
# configure, not any of the MCA components.
if test -f $topdir_file ; then
rm -rf libltdl opal/libltdl opal/ltdl.h
run_and_check $ompi_libtoolize --automake --copy --ltdl
if test -d libltdl; then
echo " -- Moving libltdl to opal/"
mv libltdl opal
fi
if check_version "1.10.0" $ompi_automake_found_version ; then
ver=`grep '^am__api_version' opal/libltdl/configure | cut -f2 -d=`
# eat single quotes
eval "ver=$ver"
if ! check_version "1.9.7" $ver ; then
indent=" "
echo "** Updating Automake version in libltdl package"
pushd opal/libltdl > /dev/null 2>&1
run_and_check $ompi_aclocal
run_and_check $ompi_automake
chmod u+w configure # Need this for FreeBSD.
run_and_check $ompi_autoconf
popd > /dev/null 2>&1
unset indent
fi
fi
if test ! -r opal/libltdl/ltdl.h; then
cat <<EOF
$ompi_libtoolize --ltdl apparently failed to install libltdl.
Please check that you have the libltdl development files installed.
EOF
exit 1
fi
echo "** Adjusting libltdl for OMPI :-("
echo " ++ patching for argz bugfix in libtool 1.5"
cd opal/libltdl
if test "`grep 'while ((before >= *pargz) && (before[-1] != LT_EOS_CHAR))' ltdl.c`" != ""; then
patch -N -p0 <<EOF
--- ltdl.c.old 2003-11-26 16:42:17.000000000 -0500
+++ ltdl.c 2003-12-03 17:06:27.000000000 -0500
@@ -682,7 +682,7 @@
/* This probably indicates a programmer error, but to preserve
semantics, scan back to the start of an entry if BEFORE points
into the middle of it. */
- while ((before >= *pargz) && (before[-1] != LT_EOS_CHAR))
+ while ((before > *pargz) && (before[-1] != LT_EOS_CHAR))
--before;
{
EOF
#'
else
echo " -- your libtool doesn't need this! yay!"
fi
cd ../..
echo " ++ patching 64-bit OS X bug in ltmain.sh"
if test ! -z "`grep otool config/ltmain.sh`" -a \
-z "`grep otool64 config/ltmain.sh`"; then
patch -N -p0 < config/ltmain_otool.diff
rm -f config/ltmain.sh.orig
else
echo " -- your libtool doesn't need this! yay!"
fi
echo " ++ RTLD_GLOBAL in libltdl"
if test -r opal/libltdl/loaders/dlopen.c && \
test ! -z "`grep 'filename, LT_LAZY_OR_NOW' opal/libltdl/loaders/dlopen.c`"; then
patch -N -p0 < config/libltdl_dlopen_global.diff
else
echo " -- your libltdl doesn't need this! yay!"
fi
if grep 'chmod -w \.' configure ; then
echo " ++ patching configure for broken -c/-o compiler test"
sed -e 's/chmod -w \./#OMPI\/MPI FIX: chmod -w ./' \
configure > configure.new
mv configure.new configure
chmod a+x configure
fi
# See http://lists.gnu.org/archive/html/bug-libtool/2008-05/msg00045.html
if check_version "2.1.9999" $ompi_libtoolize_found_version ; then
echo " ++ patching for ifort (LT 2.2.0-4)"
cd opal/libltdl/m4
patch -N -p0 < ../../../config/lt224-icc.diff > /dev/null 2>&1
rm -f libtool.m4.orig
cd ../../..
fi
else
run_and_check $ompi_libtoolize --automake --copy
fi
if test -f $topdir_file; then
run_and_check $ompi_aclocal -I config
else
run_and_check $ompi_aclocal
fi
if test "`grep AC_CONFIG_HEADER $file`" != "" -o \
"`grep AM_CONFIG_HEADER $file`" != ""; then
run_and_check $ompi_autoheader
fi
# We only need to patch the top-level aclocal.m4 for libtool stuff
# because this only affects creating C++ libraries (with pathCC).
# This must be done before we run autoconf.
if test -f $topdir_file; then
echo "** Adjusting libtool for OMPI :-("
if ! check_version "2.0.0" $ompi_libtoolize_found_version ; then
echo " ++ patching for pathscale multi-line output (LT 1.5.x)"
patch -N -p0 < config/lt1522-pathCC.diff > /dev/null 2>&1
else
echo " ++ patching for pathscale multi-line output (LT 2.x)"
patch -N -p0 < config/lt21a-pathCC.diff > /dev/null 2>&1
fi
# See http://lists.gnu.org/archive/html/bug-libtool/2008-05/msg00045.html
if check_version "2.1.9999" $ompi_libtoolize_found_version ; then
echo " ++ patching for ifort (LT 2.2.0-4)"
cd config
patch -N -p0 < lt224-icc.diff > /dev/null 2>&1
rm -f libtool.m4.orig
cd ..
fi
rm -f aclocal.m4.orig
# Libtool 1.5.2x and 2.1x automatically link in the "Cstd" STL library
# when using the Sun compilers on Linux or Solaris, even if the
# application does not use the STL (as of Feb 2008, Open MPI does not
# use any C++ STL). The problem is that Solaris has two different STL
# libraries: Cstd and stlport. Having Libtool choose that OMPI (and its
# wrapper compilers) use Cstd is problematic for users who want to
# compile their MPI applications with the other STL library. So we
# currently hack aclocal's LT macros to *not* add the Cstd library to
# any of OMPI's CXXFLAGS; the OMPI wrapper compilers can then therefore
# be used with any STL library -- it's the user's choice.
echo " ++ patching to remove solaris Cstd"
sed -e 's/-lCstd -lCrun//' \
-e 's/-library=Cstd -library=Crun//' \
aclocal.m4 > aclocal.m4.new
cp aclocal.m4.new aclocal.m4
rm -f aclocal.m4.new
fi
run_and_check $ompi_autoconf
run_and_check $ompi_automake --foreign -a --copy --include-deps
}
##############################################################################
#
# run_no_configure_component
# Prepares the non-configure component
#
# INPUT:
# - OMPI top directory
#
# OUTPUT:
# none
#
# SIDE EFFECTS:
#
##############################################################################
run_no_configure_component() {
noconf_dir="$1"
noconf_ompi_topdir="$2"
noconf_project="$3"
noconf_framework="$4"
noconf_component="$5"
# Write out to two files (they're merged at the end)
noconf_list_file="$noconf_ompi_topdir/$mca_no_config_list_file"
noconf_env_file="$noconf_ompi_topdir/$mca_no_config_env_file"
cat >> "$noconf_list_file" <<EOF
dnl ----------------------------------------------------------------
dnl No-configure component:
dnl $noconf_dir
EOF
# Tell configure to add all the PARAM_CONFIG_FILES to
# the AC_CONFIG_FILES list.
for file in $PARAM_CONFIG_FILES; do
echo "AC_CONFIG_FILES([$noconf_dir/$file])" >> "$noconf_list_file"
done
* Rewrite ompi_mca.m4 to use m4_defined lists of projects (ompi, orte, etc.), frameworks, and components without configure scripts instead of hard-coded shell variables (for projects and frameworks) and shell variable building (for components). * Add 3rd category of component configuration (in addition to configure scripts and no-configured components): configure.m4 components. These components can only be built as part of OMPI (like no-configure), but can provide an m4 file that is run as part of the main configure script. These macros can set whether the component should be built, along with just about any other configuration wanted. More care must be taken compared to configure components, as doing things like setting variables or calling AC_MSG_ERROR now affects the top-level configure script (so calling AC_MSG_ERROR if your component can't configure probably isn't what you want) * Added support to autogen.sh for the configure.m4-style components, as well as building up the m4_define lists ompi_mca.m4 now expects * Updated a number of macros to be more config.cache friendly (both so that config.cache can be used and so the test can be quickly run multiple times in the same configrue script): - ompi_config_asm - c_weak_symbols - c_get_alignment * Added new macros to be shared when configuring components: - ompi_objc.m4 (this actually provides AC_PROG_OBJC - don't ask...) - ompi_check_xgrid - ompi_check_tm - ompi_check_bproc * Updated a number of components to use configure.m4 instead of configure.stub - btl portals - io romio - tm ras and pls - bjs, lsf_bproc ras and bproc_seed pls - xgrid ras and pls - null iof (used by tm) This commit was SVN r6412.
2005-07-09 22:52:53 +04:00
cat <<EOF
--> Adding to top-level configure no-configure subdirs:
--> $noconf_dir
--> Adding to top-level configure AC_CONFIG_FILES list:
--> $PARAM_CONFIG_FILES
EOF
echo "$PARAM_CONFIG_PRIORITY $noconf_component" >> "$noconf_env_file"
* Rewrite ompi_mca.m4 to use m4_defined lists of projects (ompi, orte, etc.), frameworks, and components without configure scripts instead of hard-coded shell variables (for projects and frameworks) and shell variable building (for components). * Add 3rd category of component configuration (in addition to configure scripts and no-configured components): configure.m4 components. These components can only be built as part of OMPI (like no-configure), but can provide an m4 file that is run as part of the main configure script. These macros can set whether the component should be built, along with just about any other configuration wanted. More care must be taken compared to configure components, as doing things like setting variables or calling AC_MSG_ERROR now affects the top-level configure script (so calling AC_MSG_ERROR if your component can't configure probably isn't what you want) * Added support to autogen.sh for the configure.m4-style components, as well as building up the m4_define lists ompi_mca.m4 now expects * Updated a number of macros to be more config.cache friendly (both so that config.cache can be used and so the test can be quickly run multiple times in the same configrue script): - ompi_config_asm - c_weak_symbols - c_get_alignment * Added new macros to be shared when configuring components: - ompi_objc.m4 (this actually provides AC_PROG_OBJC - don't ask...) - ompi_check_xgrid - ompi_check_tm - ompi_check_bproc * Updated a number of components to use configure.m4 instead of configure.stub - btl portals - io romio - tm ras and pls - bjs, lsf_bproc ras and bproc_seed pls - xgrid ras and pls - null iof (used by tm) This commit was SVN r6412.
2005-07-09 22:52:53 +04:00
}
* Rewrite ompi_mca.m4 to use m4_defined lists of projects (ompi, orte, etc.), frameworks, and components without configure scripts instead of hard-coded shell variables (for projects and frameworks) and shell variable building (for components). * Add 3rd category of component configuration (in addition to configure scripts and no-configured components): configure.m4 components. These components can only be built as part of OMPI (like no-configure), but can provide an m4 file that is run as part of the main configure script. These macros can set whether the component should be built, along with just about any other configuration wanted. More care must be taken compared to configure components, as doing things like setting variables or calling AC_MSG_ERROR now affects the top-level configure script (so calling AC_MSG_ERROR if your component can't configure probably isn't what you want) * Added support to autogen.sh for the configure.m4-style components, as well as building up the m4_define lists ompi_mca.m4 now expects * Updated a number of macros to be more config.cache friendly (both so that config.cache can be used and so the test can be quickly run multiple times in the same configrue script): - ompi_config_asm - c_weak_symbols - c_get_alignment * Added new macros to be shared when configuring components: - ompi_objc.m4 (this actually provides AC_PROG_OBJC - don't ask...) - ompi_check_xgrid - ompi_check_tm - ompi_check_bproc * Updated a number of components to use configure.m4 instead of configure.stub - btl portals - io romio - tm ras and pls - bjs, lsf_bproc ras and bproc_seed pls - xgrid ras and pls - null iof (used by tm) This commit was SVN r6412.
2005-07-09 22:52:53 +04:00
##############################################################################
#
# run_m4_configure_component
# Prepares the component with an .m4 file that should be used to
# configure the component.
#
# INPUT:
#
# OUTPUT:
# none
#
# SIDE EFFECTS:
#
##############################################################################
run_m4_configure_component() {
m4conf_dir="$1"
m4conf_ompi_topdir="$2"
m4conf_project="$3"
m4conf_framework="$4"
m4conf_component="$5"
# Write out to two files (they're merged at the end)
m4conf_list_file="$m4conf_ompi_topdir/$mca_no_config_list_file"
m4conf_env_file="$m4conf_ompi_topdir/$mca_m4_config_env_file"
cat >> "$m4conf_list_file" <<EOF
dnl ----------------------------------------------------------------
* Rewrite ompi_mca.m4 to use m4_defined lists of projects (ompi, orte, etc.), frameworks, and components without configure scripts instead of hard-coded shell variables (for projects and frameworks) and shell variable building (for components). * Add 3rd category of component configuration (in addition to configure scripts and no-configured components): configure.m4 components. These components can only be built as part of OMPI (like no-configure), but can provide an m4 file that is run as part of the main configure script. These macros can set whether the component should be built, along with just about any other configuration wanted. More care must be taken compared to configure components, as doing things like setting variables or calling AC_MSG_ERROR now affects the top-level configure script (so calling AC_MSG_ERROR if your component can't configure probably isn't what you want) * Added support to autogen.sh for the configure.m4-style components, as well as building up the m4_define lists ompi_mca.m4 now expects * Updated a number of macros to be more config.cache friendly (both so that config.cache can be used and so the test can be quickly run multiple times in the same configrue script): - ompi_config_asm - c_weak_symbols - c_get_alignment * Added new macros to be shared when configuring components: - ompi_objc.m4 (this actually provides AC_PROG_OBJC - don't ask...) - ompi_check_xgrid - ompi_check_tm - ompi_check_bproc * Updated a number of components to use configure.m4 instead of configure.stub - btl portals - io romio - tm ras and pls - bjs, lsf_bproc ras and bproc_seed pls - xgrid ras and pls - null iof (used by tm) This commit was SVN r6412.
2005-07-09 22:52:53 +04:00
dnl m4-configure component:
dnl $m4conf_dir
EOF
* Rewrite ompi_mca.m4 to use m4_defined lists of projects (ompi, orte, etc.), frameworks, and components without configure scripts instead of hard-coded shell variables (for projects and frameworks) and shell variable building (for components). * Add 3rd category of component configuration (in addition to configure scripts and no-configured components): configure.m4 components. These components can only be built as part of OMPI (like no-configure), but can provide an m4 file that is run as part of the main configure script. These macros can set whether the component should be built, along with just about any other configuration wanted. More care must be taken compared to configure components, as doing things like setting variables or calling AC_MSG_ERROR now affects the top-level configure script (so calling AC_MSG_ERROR if your component can't configure probably isn't what you want) * Added support to autogen.sh for the configure.m4-style components, as well as building up the m4_define lists ompi_mca.m4 now expects * Updated a number of macros to be more config.cache friendly (both so that config.cache can be used and so the test can be quickly run multiple times in the same configrue script): - ompi_config_asm - c_weak_symbols - c_get_alignment * Added new macros to be shared when configuring components: - ompi_objc.m4 (this actually provides AC_PROG_OBJC - don't ask...) - ompi_check_xgrid - ompi_check_tm - ompi_check_bproc * Updated a number of components to use configure.m4 instead of configure.stub - btl portals - io romio - tm ras and pls - bjs, lsf_bproc ras and bproc_seed pls - xgrid ras and pls - null iof (used by tm) This commit was SVN r6412.
2005-07-09 22:52:53 +04:00
# Tell configure to add all the PARAM_CONFIG_FILES to
# the AC_CONFIG_FILES list.
for file in $PARAM_CONFIG_FILES; do
echo "AC_CONFIG_FILES([$m4conf_dir/$file])" >> "$m4conf_list_file"
done
# add the m4_include of the m4 file into the mca .m4 file
# directly. It shouldn't be in a macro, so this is fairly safe to
# do. By this point, there should already be a header and all
# that. m4_includes are relative to the currently included file,
# so need the .. to get us from config/ to the topsrcdir again.
echo "m4_include(${m4conf_project}/mca/${m4conf_framework}/${m4conf_component}/configure.m4)" >> "$m4conf_ompi_topdir/$mca_m4_include_file"
* Rewrite ompi_mca.m4 to use m4_defined lists of projects (ompi, orte, etc.), frameworks, and components without configure scripts instead of hard-coded shell variables (for projects and frameworks) and shell variable building (for components). * Add 3rd category of component configuration (in addition to configure scripts and no-configured components): configure.m4 components. These components can only be built as part of OMPI (like no-configure), but can provide an m4 file that is run as part of the main configure script. These macros can set whether the component should be built, along with just about any other configuration wanted. More care must be taken compared to configure components, as doing things like setting variables or calling AC_MSG_ERROR now affects the top-level configure script (so calling AC_MSG_ERROR if your component can't configure probably isn't what you want) * Added support to autogen.sh for the configure.m4-style components, as well as building up the m4_define lists ompi_mca.m4 now expects * Updated a number of macros to be more config.cache friendly (both so that config.cache can be used and so the test can be quickly run multiple times in the same configrue script): - ompi_config_asm - c_weak_symbols - c_get_alignment * Added new macros to be shared when configuring components: - ompi_objc.m4 (this actually provides AC_PROG_OBJC - don't ask...) - ompi_check_xgrid - ompi_check_tm - ompi_check_bproc * Updated a number of components to use configure.m4 instead of configure.stub - btl portals - io romio - tm ras and pls - bjs, lsf_bproc ras and bproc_seed pls - xgrid ras and pls - null iof (used by tm) This commit was SVN r6412.
2005-07-09 22:52:53 +04:00
cat <<EOF
* Rewrite ompi_mca.m4 to use m4_defined lists of projects (ompi, orte, etc.), frameworks, and components without configure scripts instead of hard-coded shell variables (for projects and frameworks) and shell variable building (for components). * Add 3rd category of component configuration (in addition to configure scripts and no-configured components): configure.m4 components. These components can only be built as part of OMPI (like no-configure), but can provide an m4 file that is run as part of the main configure script. These macros can set whether the component should be built, along with just about any other configuration wanted. More care must be taken compared to configure components, as doing things like setting variables or calling AC_MSG_ERROR now affects the top-level configure script (so calling AC_MSG_ERROR if your component can't configure probably isn't what you want) * Added support to autogen.sh for the configure.m4-style components, as well as building up the m4_define lists ompi_mca.m4 now expects * Updated a number of macros to be more config.cache friendly (both so that config.cache can be used and so the test can be quickly run multiple times in the same configrue script): - ompi_config_asm - c_weak_symbols - c_get_alignment * Added new macros to be shared when configuring components: - ompi_objc.m4 (this actually provides AC_PROG_OBJC - don't ask...) - ompi_check_xgrid - ompi_check_tm - ompi_check_bproc * Updated a number of components to use configure.m4 instead of configure.stub - btl portals - io romio - tm ras and pls - bjs, lsf_bproc ras and bproc_seed pls - xgrid ras and pls - null iof (used by tm) This commit was SVN r6412.
2005-07-09 22:52:53 +04:00
--> Adding to top-level configure m4-configure subdirs:
--> $m4conf_dir
--> Adding to top-level configure AC_CONFIG_FILES list:
--> $PARAM_CONFIG_FILES
EOF
echo "$PARAM_CONFIG_PRIORITY $m4conf_component" >> "$m4conf_env_file"
}
##############################################################################
#
# make_template_version_header -- make a templated version header
# file, but only if we have a PARAM_VERSION_FILE that exists
#
# INPUT:
# - filename base
# - component type name
# - component name
#
# OUTPUT:
# none
#
# SIDE EFFECTS:
#
##############################################################################
make_version_header_template() {
mvht_filename="$1"
mvht_component_type="$2"
mvht_component_name="$3"
# See if we have a VERSION file
PARAM_CONFIG_FILES_save="$PARAM_CONFIG_FILES"
. ./configure.params
if test -z "$PARAM_VERSION_FILE"; then
if test -f "VERSION"; then
PARAM_VERSION_FILE="VERSION"
fi
else
if test ! -f "$PARAM_VERSION_FILE"; then
PARAM_VERSION_FILE=
fi
fi
if test -n "$PARAM_VERSION_FILE" -a -f "$PARAM_VERSION_FILE" -a \
"$pd_component_type" != "common"; then
rm -f "$mvht_filename.template.in"
cat > "$mvht_filename.template.in" <<EOF
/*
* This file is automatically created by autogen.sh; it should not
* be edited by hand!!
*
* List of version number for this component
*/
#ifndef MCA_${mvht_component_type}_${mvht_component_name}_VERSION_H
#define MCA_${mvht_component_type}_${mvht_component_name}_VERSION_H
#define MCA_${mvht_component_type}_${mvht_component_name}_MAJOR_VERSION @MCA_${mvht_component_type}_${mvht_component_name}_MAJOR_VERSION@
#define MCA_${mvht_component_type}_${mvht_component_name}_MINOR_VERSION @MCA_${mvht_component_type}_${mvht_component_name}_MINOR_VERSION@
#define MCA_${mvht_component_type}_${mvht_component_name}_RELEASE_VERSION @MCA_${mvht_component_type}_${mvht_component_name}_RELEASE_VERSION@
#define MCA_${mvht_component_type}_${mvht_component_name}_GREEK_VERSION "@MCA_${mvht_component_type}_${mvht_component_name}_GREEK_VERSION@"
#define MCA_${mvht_component_type}_${mvht_component_name}_SVN_VERSION "@MCA_${mvht_component_type}_${mvht_component_name}_SVN_VERSION@"
#define MCA_${mvht_component_type}_${mvht_component_name}_VERSION "@MCA_${mvht_component_type}_${mvht_component_name}_VERSION@"
#endif /* MCA_${mvht_component_type}_${mvht_component_name}_VERSION_H */
EOF
fi
PARAM_CONFIG_FILES="$PARAM_CONFIG_FILES_save"
unset PARAM_VERSION_FILE PARAM_CONFIG_FILES_save
}
component_list_sort() {
cls_filename="$1"
# why, oh, why can't non-gnu sort support the -s (stable) option?
# Solaris sort supports -r -n and -u, so we'll assume that works everywhere
# get the list of priorities
component_list=
cls_priority_list=`sort -r -n -u "$cls_filename" | cut -f1 -d' ' | xargs`
for cls_priority in $cls_priority_list ; do
component_list="$component_list "`grep "^$cls_priority " "$cls_filename" | cut -f2 -d' ' | xargs`
done
}
##############################################################################
#
# process_dir - look at the files present in a given directory, and do
# one of the following:
# - skip/ignore it
# - run custom autogen.sh in it
# - run the GNU tools in it
# - get a list of Makefile.am's to add to the top-level configure
#
# INPUT:
# - directory to run in
# - OMPI top directory
#
# OUTPUT:
# none
#
# SIDE EFFECTS:
# - skips directories with .ompi_no_gnu .ompi_ignore
# - uses provided autogen.sh if available
#
##############################################################################
process_dir() {
pd_dir="$1"
pd_ompi_topdir="$2"
pd_project="$3"
pd_framework="$4"
pd_component="$5"
pd_cur_dir="`pwd`"
# Convert to absolutes
if test -d "$pd_dir"; then
cd "$pd_dir"
pd_abs_dir="`pwd`"
cd "$pd_cur_dir"
fi
if test -d "$pd_ompi_topdir"; then
cd "$pd_ompi_topdir"
pd_ompi_topdir="`pwd`"
cd "$pd_cur_dir"
fi
# clean our environment a bit, since we might evaluate a configure.params
unset PARAM_CONFIG_FILES
PARAM_CONFIG_PRIORITY="0"
if test -d "$pd_dir"; then
cd "$pd_dir"
# See if the package doesn't want us to set it up
if test -f .ompi_no_gnu; then
cat <<EOF
*** Found .ompi_no_gnu file -- skipping GNU setup in:
*** `pwd`
EOF
elif test -f .ompi_ignore -a ! -f .ompi_unignore; then
# Note that if we have an empty (but existant)
# .ompi_unignore, then we ignore the .ompi_ignore file
# (and therefore build the component)
cat <<EOF
*** Found .ompi_ignore file -- skipping entire tree:
*** `pwd`
EOF
# Use && instead of -a here for the test conditions because if
# you use -a, then "test" will execute *all* condition clauses
# (even if the first one is false), meaning that grep will
# fail if there is no .ompi_unignore file. If you use &&,
# then the latter tests will not be executed if a prior one
# fails (i.e., grep won't run if .ompi_unignore does not
# exist).
elif test -f .ompi_ignore && \
test -s .ompi_unignore && \
test -z "`$egrep $USER\$\|$USER@$HOST .ompi_unignore`" ; then
# If we have a non-empty .ompi_unignore and our username
# is in there somewhere, we ignore the .ompi_ignore (and
# therefore build the component). Otherwise, this
# condition is true and we don't configure.
cat <<EOF
*** Found .ompi_ignore file and .ompi_unignore didn't invalidate --
*** skipping entire tree:
*** `pwd`
EOF
elif test "$pd_abs_dir" != "$pd_ompi_topdir" -a -x autogen.sh; then
cat <<EOF
*** Found custom autogen.sh file in:
*** `pwd`
EOF
./autogen.sh
if test ! $? -eq 0 ; then
echo "Error running autogen.sh -l in `pwd`. Aborting."
exit 1
fi
* Rewrite ompi_mca.m4 to use m4_defined lists of projects (ompi, orte, etc.), frameworks, and components without configure scripts instead of hard-coded shell variables (for projects and frameworks) and shell variable building (for components). * Add 3rd category of component configuration (in addition to configure scripts and no-configured components): configure.m4 components. These components can only be built as part of OMPI (like no-configure), but can provide an m4 file that is run as part of the main configure script. These macros can set whether the component should be built, along with just about any other configuration wanted. More care must be taken compared to configure components, as doing things like setting variables or calling AC_MSG_ERROR now affects the top-level configure script (so calling AC_MSG_ERROR if your component can't configure probably isn't what you want) * Added support to autogen.sh for the configure.m4-style components, as well as building up the m4_define lists ompi_mca.m4 now expects * Updated a number of macros to be more config.cache friendly (both so that config.cache can be used and so the test can be quickly run multiple times in the same configrue script): - ompi_config_asm - c_weak_symbols - c_get_alignment * Added new macros to be shared when configuring components: - ompi_objc.m4 (this actually provides AC_PROG_OBJC - don't ask...) - ompi_check_xgrid - ompi_check_tm - ompi_check_bproc * Updated a number of components to use configure.m4 instead of configure.stub - btl portals - io romio - tm ras and pls - bjs, lsf_bproc ras and bproc_seed pls - xgrid ras and pls - null iof (used by tm) This commit was SVN r6412.
2005-07-09 22:52:53 +04:00
elif test -f configure.params -a -f configure.m4 ; then
cat <<EOF
*** Found configure.params and configure.m4
*** `pwd`
EOF
. ./configure.params
if test -z "$PARAM_CONFIG_FILES"; then
cat <<EOF
*** No PARAM_CONFIG_FILES!
*** Nothing to do -- skipping this directory
EOF
else
# temporary workaround - remove possibly there configure code
rm -f "configure" "configure.ac*" "acinclude*" "aclocal.m4"
run_m4_configure_component "$pd_dir" "$pd_ompi_topdir" \
"$pd_project" "$pd_framework" "$pd_component"
fi
elif test -f configure.ac -o -f configure.in; then
# If we have configure.ac or configure.in, run the GNU
# tools here
cat <<EOF
*** Found configure.(in|ac)
*** `pwd`
EOF
run_gnu_tools "$pd_ompi_topdir"
elif test -f configure.params -a -f configure.stub; then
cat <<EOF
*** Found configure.params and configure.stub
*** `pwd`
EOF
run_gnu_tools "$pd_ompi_topdir"
elif test -f configure.params; then
cat <<EOF
*** Found configure.params
*** `pwd`
EOF
. ./configure.params
if test -z "$PARAM_CONFIG_FILES"; then
cat <<EOF
*** No PARAM_CONFIG_FILES!
*** Nothing to do -- skipping this directory
EOF
else
run_no_configure_component "$pd_dir" "$pd_ompi_topdir" \
"$pd_project" "$pd_framework" "$pd_component"
fi
else
cat <<EOF
*** Nothing found; directory skipped
*** `pwd`
EOF
fi
# See if there's a file containing additional directories to
# traverse. Shell scripts aren't too good at recursion (no
# local state!), so just have a child autogen.sh do the work.
if test -f $autogen_subdir_file; then
pd_subdir_start_dir="`pwd`"
echo ""
echo "==> Found $autogen_subdir_file -- sub-traversing..."
echo ""
for dir in `cat $autogen_subdir_file`; do
if test -d "$dir"; then
echo "*** Running autogen.sh in $dir"
echo "*** (started in $pd_subdir_start_dir)"
cd "$dir"
$pd_ompi_topdir/autogen.sh -l
if test ! $? -eq 0 ; then
echo "Error running autogen.sh -l in $dir. Aborting."
exit 1
fi
cd "$pd_subdir_start_dir"
echo ""
fi
done
echo "<== Back in $pd_subdir_start_dir"
echo "<== autogen.sh continuing..."
fi
# Go back to the topdir
cd "$pd_cur_dir"
fi
unset PARAM_CONFIG_FILES PARAM_VERSION_FILE
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
unset pd_dir pd_ompi_topdir pd_cur_dir pd_component_type
}
process_framework() {
framework_path="$1"
rg_cwd="$2"
project="$3"
framework="$4"
if test "$framework" != "base" -a \
-d "$framework_path" ; then
if test "$framework" = "common" -o \
-r "${framework_path}/${framework}.h" ; then
framework_list="$framework_list $framework"
# Add the framework's configure file into configure,
# if there is one
if test -r "${framework_path}/configure.m4" ; then
echo "m4_include(${framework_path}/configure.m4)" >> "$mca_m4_include_file"
fi
echo "AC_CONFIG_FILES(${framework_path}/Makefile)" >> "$mca_no_config_list_file"
rm -f "$mca_no_config_env_file" "$mca_m4_config_env_file"
touch "$mca_no_config_env_file" "$mca_m4_config_env_file"
for component_path in "$framework_path"/*; do
if test -d "$component_path"; then
if test -f "$component_path/configure.in" -o \
-f "$component_path/configure.params" -o \
-f "$component_path/configure.ac"; then
component=`basename "$component_path"`
process_dir "$component_path" "$rg_cwd" \
"$project" "$framework" "$component"
fi
fi
done
# make list of components that are "no configure".
# Sort the list by priority (stable, so things stay in
# alphabetical order at the same priority), then munge
# it into form we like
component_list=
component_list_sort $mca_no_config_env_file
component_list_define="m4_define(mca_${framework}_no_config_component_list, ["
component_list_define_first="1"
for component in $component_list ; do
if test "$component_list_define_first" = "1"; then
component_list_define="${component_list_define}${component}"
component_list_define_first="0"
else
component_list_define="${component_list_define}, ${component}"
fi
done
component_list_define="${component_list_define}])"
echo "$component_list_define" >> "$mca_no_configure_components_file"
# make list of components that are "m4 configure"
component_list=
component_list_sort $mca_m4_config_env_file
component_list_define="m4_define(mca_${framework}_m4_config_component_list, ["
component_list_define_first="1"
for component in $component_list ; do
if test "$component_list_define_first" = "1"; then
component_list_define="${component_list_define}${component}"
component_list_define_first="0"
else
component_list_define="${component_list_define}, ${component}"
fi
done
component_list_define="${component_list_define}])"
echo "$component_list_define" >> "$mca_no_configure_components_file"
fi
fi
}
process_project() {
project_path="$1"
rg_cwd="$2"
project="$3"
project_list="$project_list $project"
framework_list=""
for framework_path in $project_path/mca/*; do
framework=`basename "$framework_path"`
process_framework $framework_path $rg_cwd $project $framework
done
# make list of frameworks for this project
framework_list_define="m4_define(mca_${project}_framework_list, ["
framework_list_define_first="1"
for framework in $framework_list ; do
if test "$framework_list_define_first" = "1"; then
framework_list_define="${framework_list_define}${framework}"
framework_list_define_first="0"
else
framework_list_define="${framework_list_define}, ${framework}"
fi
done
framework_list_define="${framework_list_define}])"
echo "$framework_list_define" >> "$mca_no_configure_components_file"
# process contributed software packages (ompi/contrib/*)
if test "$project" = "ompi"; then
for contrib_path in $project_path/contrib/*; do
process_dir $contrib_path $rg_cwd
done
fi
}
##############################################################################
#
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
# run_global - run the config in the top OMPI dir and all MCA components
#
# INPUT:
# none
#
# OUTPUT:
# none
#
# SIDE EFFECTS:
#
##############################################################################
run_global() {
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
# [Re-]Create the mca_component_list file
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
rm -f "$mca_no_configure_components_file" "$mca_no_config_list_file" \
* Rewrite ompi_mca.m4 to use m4_defined lists of projects (ompi, orte, etc.), frameworks, and components without configure scripts instead of hard-coded shell variables (for projects and frameworks) and shell variable building (for components). * Add 3rd category of component configuration (in addition to configure scripts and no-configured components): configure.m4 components. These components can only be built as part of OMPI (like no-configure), but can provide an m4 file that is run as part of the main configure script. These macros can set whether the component should be built, along with just about any other configuration wanted. More care must be taken compared to configure components, as doing things like setting variables or calling AC_MSG_ERROR now affects the top-level configure script (so calling AC_MSG_ERROR if your component can't configure probably isn't what you want) * Added support to autogen.sh for the configure.m4-style components, as well as building up the m4_define lists ompi_mca.m4 now expects * Updated a number of macros to be more config.cache friendly (both so that config.cache can be used and so the test can be quickly run multiple times in the same configrue script): - ompi_config_asm - c_weak_symbols - c_get_alignment * Added new macros to be shared when configuring components: - ompi_objc.m4 (this actually provides AC_PROG_OBJC - don't ask...) - ompi_check_xgrid - ompi_check_tm - ompi_check_bproc * Updated a number of components to use configure.m4 instead of configure.stub - btl portals - io romio - tm ras and pls - bjs, lsf_bproc ras and bproc_seed pls - xgrid ras and pls - null iof (used by tm) This commit was SVN r6412.
2005-07-09 22:52:53 +04:00
"$mca_no_config_env_file" "$mca_m4_config_env_file" "$mca_m4_include_file"
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
touch "$mca_no_configure_components_file" "$mca_no_config_list_file" \
* Rewrite ompi_mca.m4 to use m4_defined lists of projects (ompi, orte, etc.), frameworks, and components without configure scripts instead of hard-coded shell variables (for projects and frameworks) and shell variable building (for components). * Add 3rd category of component configuration (in addition to configure scripts and no-configured components): configure.m4 components. These components can only be built as part of OMPI (like no-configure), but can provide an m4 file that is run as part of the main configure script. These macros can set whether the component should be built, along with just about any other configuration wanted. More care must be taken compared to configure components, as doing things like setting variables or calling AC_MSG_ERROR now affects the top-level configure script (so calling AC_MSG_ERROR if your component can't configure probably isn't what you want) * Added support to autogen.sh for the configure.m4-style components, as well as building up the m4_define lists ompi_mca.m4 now expects * Updated a number of macros to be more config.cache friendly (both so that config.cache can be used and so the test can be quickly run multiple times in the same configrue script): - ompi_config_asm - c_weak_symbols - c_get_alignment * Added new macros to be shared when configuring components: - ompi_objc.m4 (this actually provides AC_PROG_OBJC - don't ask...) - ompi_check_xgrid - ompi_check_tm - ompi_check_bproc * Updated a number of components to use configure.m4 instead of configure.stub - btl portals - io romio - tm ras and pls - bjs, lsf_bproc ras and bproc_seed pls - xgrid ras and pls - null iof (used by tm) This commit was SVN r6412.
2005-07-09 22:52:53 +04:00
"$mca_m4_config_env_file" "$mca_m4_include_file"
# create header for the component m4 include file
cat > "$mca_m4_include_file" <<EOF
dnl
dnl \$HEADER
dnl
dnl This file is automatically created by autogen.sh; it should not
dnl be edited by hand!!
EOF
#create header for the component config file
cat > "$mca_no_configure_components_file" <<EOF
dnl
dnl \$HEADER
dnl
dnl This file is automatically created by autogen.sh; it should not
dnl be edited by hand!!
EOF
# Now run the config in every directory in <location>/mca/*/*
# that has a configure.in or configure.ac script
#
# In order to deal with components that have .m4 files, we need to
# build up m4_defined lists along the way. Unfortunately, there
# is no good way to do this at the end (stupid sh), so we have to
# do it as we are going through the lists of frameworks and
# components. Use a file to keep the list of components (we don't
# want every component in a framework included, as the
# determination about skipping components or whether the component
# has its own configure script are made later on in the process.
rg_cwd="`pwd`"
echo $rg_cwd
project_list=""
for project_path in $config_project_list; do
project=`basename "$project_path"`
process_project $project_path $rg_cwd $project
done
# create the m4 defines for the list of projects. The list of
# frameworks for each project is already created and in the file.
project_list_define="m4_define(mca_project_list, ["
project_list_define_first="1"
for project in $project_list ; do
if test "$project_list_define_first" = "1"; then
project_list_define="${project_list_define}${project}"
project_list_define_first="0"
else
project_list_define="${project_list_define}, ${project}"
fi
done
project_list_define="${project_list_define}])"
echo "$project_list_define" >> "$mca_no_configure_components_file"
cat >> "$mca_no_configure_components_file" <<EOF
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
dnl List all the no-configure components that we found, and AC_DEFINE
dnl their versions
* Rewrite ompi_mca.m4 to use m4_defined lists of projects (ompi, orte, etc.), frameworks, and components without configure scripts instead of hard-coded shell variables (for projects and frameworks) and shell variable building (for components). * Add 3rd category of component configuration (in addition to configure scripts and no-configured components): configure.m4 components. These components can only be built as part of OMPI (like no-configure), but can provide an m4 file that is run as part of the main configure script. These macros can set whether the component should be built, along with just about any other configuration wanted. More care must be taken compared to configure components, as doing things like setting variables or calling AC_MSG_ERROR now affects the top-level configure script (so calling AC_MSG_ERROR if your component can't configure probably isn't what you want) * Added support to autogen.sh for the configure.m4-style components, as well as building up the m4_define lists ompi_mca.m4 now expects * Updated a number of macros to be more config.cache friendly (both so that config.cache can be used and so the test can be quickly run multiple times in the same configrue script): - ompi_config_asm - c_weak_symbols - c_get_alignment * Added new macros to be shared when configuring components: - ompi_objc.m4 (this actually provides AC_PROG_OBJC - don't ask...) - ompi_check_xgrid - ompi_check_tm - ompi_check_bproc * Updated a number of components to use configure.m4 instead of configure.stub - btl portals - io romio - tm ras and pls - bjs, lsf_bproc ras and bproc_seed pls - xgrid ras and pls - null iof (used by tm) This commit was SVN r6412.
2005-07-09 22:52:53 +04:00
AC_DEFUN([MCA_NO_CONFIG_CONFIG_FILES],[
`cat $mca_no_config_list_file`
])dnl
EOF
* Rewrite ompi_mca.m4 to use m4_defined lists of projects (ompi, orte, etc.), frameworks, and components without configure scripts instead of hard-coded shell variables (for projects and frameworks) and shell variable building (for components). * Add 3rd category of component configuration (in addition to configure scripts and no-configured components): configure.m4 components. These components can only be built as part of OMPI (like no-configure), but can provide an m4 file that is run as part of the main configure script. These macros can set whether the component should be built, along with just about any other configuration wanted. More care must be taken compared to configure components, as doing things like setting variables or calling AC_MSG_ERROR now affects the top-level configure script (so calling AC_MSG_ERROR if your component can't configure probably isn't what you want) * Added support to autogen.sh for the configure.m4-style components, as well as building up the m4_define lists ompi_mca.m4 now expects * Updated a number of macros to be more config.cache friendly (both so that config.cache can be used and so the test can be quickly run multiple times in the same configrue script): - ompi_config_asm - c_weak_symbols - c_get_alignment * Added new macros to be shared when configuring components: - ompi_objc.m4 (this actually provides AC_PROG_OBJC - don't ask...) - ompi_check_xgrid - ompi_check_tm - ompi_check_bproc * Updated a number of components to use configure.m4 instead of configure.stub - btl portals - io romio - tm ras and pls - bjs, lsf_bproc ras and bproc_seed pls - xgrid ras and pls - null iof (used by tm) This commit was SVN r6412.
2005-07-09 22:52:53 +04:00
# Remove temp files
* Rewrite ompi_mca.m4 to use m4_defined lists of projects (ompi, orte, etc.), frameworks, and components without configure scripts instead of hard-coded shell variables (for projects and frameworks) and shell variable building (for components). * Add 3rd category of component configuration (in addition to configure scripts and no-configured components): configure.m4 components. These components can only be built as part of OMPI (like no-configure), but can provide an m4 file that is run as part of the main configure script. These macros can set whether the component should be built, along with just about any other configuration wanted. More care must be taken compared to configure components, as doing things like setting variables or calling AC_MSG_ERROR now affects the top-level configure script (so calling AC_MSG_ERROR if your component can't configure probably isn't what you want) * Added support to autogen.sh for the configure.m4-style components, as well as building up the m4_define lists ompi_mca.m4 now expects * Updated a number of macros to be more config.cache friendly (both so that config.cache can be used and so the test can be quickly run multiple times in the same configrue script): - ompi_config_asm - c_weak_symbols - c_get_alignment * Added new macros to be shared when configuring components: - ompi_objc.m4 (this actually provides AC_PROG_OBJC - don't ask...) - ompi_check_xgrid - ompi_check_tm - ompi_check_bproc * Updated a number of components to use configure.m4 instead of configure.stub - btl portals - io romio - tm ras and pls - bjs, lsf_bproc ras and bproc_seed pls - xgrid ras and pls - null iof (used by tm) This commit was SVN r6412.
2005-07-09 22:52:53 +04:00
rm -f $mca_no_config_list_file $mca_no_config_env_file $mca_m4_config_env_file
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
# Finally, after we found all the no-configure MCA components, run
# the config in the top-level directory
process_dir . . "" "" ""
unset project project_path framework framework_path component component_path
}
##############################################################################
#
# check_for_svk_checkout - determine whether this is an SVK checkout
#
# INPUT:
# none
#
# OUTPUT:
# none
#
# SIDE EFFECTS:
#
##############################################################################
check_for_svk_checkout() {
is_svk_checkout=0
svk_path=`which svk 2>/dev/null`
if test -x "$svk_path" -a -d "$HOME/.svk/local"; then
top_level_dir="`dirname $0`"
svk info $top_level_dir >/dev/null 2>&1
if test "$?" = 0 ; then
is_svk_checkout=1
fi
fi
}
##############################################################################
#
# main - do the real work...
#
##############################################################################
# announce
echo "[Checking] command line parameters"
# Check the command line to see if we should run the whole shebang, or
# just in this current directory.
want_local=no
ompidir=
no_ompi=0
no_orte=0
for arg in $*; do
case $arg in
-l) want_local=yes ;;
-ompidir|--ompidir|-ompi|--ompi) ompidir="$2" ;;
-no-ompi) no_ompi=1 ;;
-no-orte) no_orte=1 ;;
*) ;;
esac
shift
done
# announce
echo "[Checking] prerequisites"
# sanity check to make sure user isn't being stupid
check_for_svk_checkout
if test ! -d .svn -a ! $is_svk_checkout ; then
cat <<EOF
This doesn't look like a developer copy of Open MPI. You probably do not
want to run autogen.sh - it is normally not needed for a release source
tree. Giving you 5 seconds to reconsider and kill me.
EOF
sleep 5
fi
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
# figure out if we're at the top level of the OMPI tree, a component's
# top-level directory, or somewhere else.
if test -f VERSION -a -f configure.ac -a -f $topdir_file ; then
# locations to look for mca modules
config_project_list="opal orte ompi"
if test "$no_ompi" = "1" ; then
config_project_list="opal orte"
fi
if test "$no_orte" = "1" ; then
config_project_list="opal"
fi
echo "Configuring projects: $config_project_list"
# Top level of OMPI tree
ompidir="`pwd`"
elif test -f configure.in -o -f configure.ac -o -f configure.params ; then
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
# Top level of a component directory
want_local=yes
if test -z "$ompidir"; then
ompidir="../../../.."
fi
2003-11-22 19:36:58 +03:00
else
cat <<EOF
You must run this script from either the top level of the OMPI
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
directory tree or the top-level of an MCA component directory tree.
2003-11-22 19:36:58 +03:00
EOF
exit 1
fi
# find all the apps we are going to run
find_app "autom4te"
find_app "aclocal"
find_app "autoheader"
find_app "autoconf"
find_app "libtoolize"
find_app "automake"
2003-11-22 19:36:58 +03:00
# do the work
if test "$want_local" = "yes"; then
process_dir . $ompidir
else
run_global
2003-11-22 19:36:58 +03:00
fi
# BWB --- temporary time to update autoconf / automake hack
bad=0
if ! check_version "2.60.0" $ompi_autoconf_found_version ; then
bad=1
fi
if ! check_version "1.10.0" $ompi_automake_found_version ; then
bad=1
fi
if test $bad -eq 1 ; then
cat <<EOF
-------------------------------------------------------------------------
WARNING WARNING WARNING WARNING WARNING WARNING WARNING WARNING WARNING
You are currently using Autoconf $ac_version and Automake $am_version.
Starting Saturday, May 12th, Autoconf 2.60 and Automake 1.10 or higher
will be required to built configure for the trunk. Note that these
autotool versions can also safely build the v1.2 branch. You will
still be able to use either Libtool 1.5.22 or 2.1a.
WARNING WARNING WARNING WARNING WARNING WARNING WARNING WARNING WARNING
-------------------------------------------------------------------------
EOF
fi
# BWB --- end evil hack
2003-11-22 19:36:58 +03:00
# All done
exit 0