build: support building all tools in a single binary

Add the --enable-single-binary option to the configure file.
When enabled, this option builds a single binary file containing
the selected tools.  Which tool gets executed depends on the value
of argv[0] which can be set implicitly through symlinks to the
single program.

This setup reduces significantly the size of a complete coreutils
install, since code from lib/libcoreutils.a is not duplicated in
every one of the more than 100 binaries.  Runtime overhead is
increased due to more dynamic libraries being loaded, and extra
initialization being performed for all utils.  Also initially
a larger binary is loaded from storage, though this is usually
alleviated due to caching and lazy mmaping of unused blocks,
and in fact the single binary should have better caching
characteristics.

Comparing the size of the individual versus single binary on x86_64:
  $ cd src
  $ size coreutils
  $ size -t $(../build-aux/gen-lists-of-programs.sh --list-progs |
              grep -Ev '(coreutils|libstdbuf)') | tail -n1
     text    data     bss     dec     hex filename
  1097416    5388   88432 1191236  122d44 src/coreutils
  4901010  124964  163768 5189742  4f306e (TOTALS)

Storage requirements are reduced similarly:
  $ cd src
  $ du -h coreutils
  $ du -ch $(../build-aux/gen-lists-of-programs.sh --list-progs |
             grep -Ev '(coreutils|libstdbuf)') | tail -n1
  1.2M    coreutils
  5.3M    total

When installing, the makefile will create either symlinks or
shebangs based on the --enable-single-binary setting, for
each configured tool. In this way, all the tools are still
callable individually, but they are all implemented by the same
"coreutils" binary installed on the same directory.

* .gitignore: Add new generated files.
* Makefile.am: New rules to generate build-aux/gen-single-binary.sh
  and install symlinks.
* NEWS: Mention the new feature.
* README: Add "coreutils" to the list of utils.
* bootstrap.conf: Regenerate src/single-binary.mk
* build-aux/gen-lists-of-programs.sh: New --list-progs option.
* build-aux/gen-single-binary.sh: Regenerate
* configure.ac: New --enable-single-binary option and other variables.
Disallow --enable-single-binary=symlinks with --program-prefix et. al.
* man/coreutils.x: Manpage hook.
* man/local.mk: Add manpage hook and fix dependencies.
* src/coreutils.c: Multicall implementation.
* src/local.mk: New rules for the single binary option.
* tests/local.mk: Add $single_binary_progs to support
require_built_() from init.cfg
* tests/misc/env.sh: Avoid the use of symlink to echo.
* tests/misc/help-version.sh: Add exception for coreutils.
* tests/install/basic-1.sh: Really avoid using ginstall strip
functionality if there is an issue with the independent strip command.
* src/kill.c: Changes to call exit() in main.
* src/readlink.c: Likewise.
* src/shuf.c: Likewise.
* src/timeout.c: Likewise.
* src/truncate.c: Likewise.

2 files changed, 117 insertions, 0 deletions

diff --git a/build-aux/gen-lists-of-programs.sh b/build-aux/gen-lists-of-programs.sh
index bf63ee30a..c95e59815 100755
--- a/build-aux/gen-lists-of-programs.sh
+++ b/build-aux/gen-lists-of-programs.sh
@@ -17,6 +17,7 @@ set -e
 # use "--enable-install-program=A,B" when invoking configure.
 disabled_by_default_progs='
     arch
+    coreutils
     hostname
 '
 
@@ -178,6 +179,12 @@ END
       echo default__progs += $progsdir/$p
     done
     ;;
+  1,--list-progs)
+    for p in $disabled_by_default_progs $build_if_possible_progs \
+        $normal_progs; do
+      echo $p
+    done
+    ;;
   *)
     echo "$0: invalid usage" >&2; exit 2
     ;;
diff --git a/build-aux/gen-single-binary.sh b/build-aux/gen-single-binary.sh
new file mode 100755
index 000000000..463570656
--- /dev/null
+++ b/build-aux/gen-single-binary.sh
@@ -0,0 +1,110 @@
+#!/bin/sh
+
+# Generate the list of rules for the single-binary option based on all the other
+# binaries found in src/local.mk.
+#
+# We need to duplicate the specific rules to build each program into a new
+# static library target. We can't reuse the existing target since we need to
+# create a .a file instead of linking the program. We can't do this at
+# ./configure since the file names need to be available when automake runs
+# to let it generate all the required rules in Makefile.in.  The configure
+# step will select which ones will be used to build, but they need to be
+# generated beforehand.
+#
+# Instead of maintaining a duplicated list of rules, we generate the
+# single-binary required rules based on the normal configuration found on
+# src/local.mk with this script.
+
+if test "x$1" = "x"; then
+  echo "Usage: $0 path/to/src/local.mk" >&2
+  exit 1
+fi
+
+set -e
+
+LOCAL_MK=$1
+GEN_LISTS_OF_PROGRAMS="`dirname "$0"`/gen-lists-of-programs.sh"
+
+ALL_PROGRAMS=$($GEN_LISTS_OF_PROGRAMS --list-progs \
+    | grep -v -F -e coreutils -e libstdbuf.so \
+    | tr '[' '_')
+
+# Compute default SOURCES. automake will assume the source file for the
+# src_${cmd} target to be src/${cmd}.c, but we will add rules to generate
+# the lib src_libsinglebin_${cmd}_a which won't match the autogenerated source
+# file. This loop will initialize the default source file and will be reset
+# later if needed.
+for cmd in $ALL_PROGRAMS; do
+  eval "src_${cmd}_SOURCES=src/${cmd}.c"
+done
+
+# Load actual values from src/local.mk. This will read all the variables from
+# the local.mk matching the src_${cmd}_... case.
+while read l; do
+  if echo "$l" | grep -E '^src_\w+ +\+?=' > /dev/null; then
+    var=$(echo $l | cut -f 1 -d ' ')
+    value=$(echo $l | cut -f 2- -d =)
+    if [ "$value" != " \$(LDADD)" ]; then
+      oldvalue=""
+      if echo $l | grep -F '+=' >/dev/null; then
+        eval "oldvalue=\${$var}"
+      fi
+      eval "$var='$oldvalue "${value//\'/\'\"\'\"\'}"'"
+    fi
+  fi
+done < $LOCAL_MK
+
+me=`echo "$0" | sed 's,.*/,,'`
+echo "## Automatically generated by $me.  DO NOT EDIT BY HAND!"
+
+# Override the sources for dir and vdir. We use a smaller version of dir and
+# vdir that relies on the ls main.
+src_dir_SOURCES="src/coreutils-dir.c"
+src_dir_LDADD+=" src/libsinglebin_ls.a"
+echo src_libsinglebin_dir_a_DEPENDENCIES = src/libsinglebin_ls.a
+src_vdir_SOURCES="src/coreutils-vdir.c"
+src_vdir_LDADD+=" src/libsinglebin_ls.a"
+echo src_libsinglebin_vdir_a_DEPENDENCIES = src/libsinglebin_ls.a
+
+# Override the sources for arch likewise, using the main from uname.
+src_arch_SOURCES="src/coreutils-arch.c"
+src_arch_LDADD+=" src/libsinglebin_uname.a"
+echo src_libsinglebin_arch_a_DEPENDENCIES = src/libsinglebin_uname.a
+
+for cmd in $ALL_PROGRAMS; do
+  echo "# Command $cmd"
+  echo noinst_LIBRARIES += src/libsinglebin_${cmd}.a
+  base="src_libsinglebin_${cmd}_a"
+  # SOURCES
+  var=src_${cmd}_SOURCES
+  eval "value=\$$var"
+  echo "${base}_SOURCES = $value"
+
+  # LDADD
+  var=src_${cmd}_LDADD
+  eval "value=\$$var"
+  if [ "x$value" != "x" ]; then
+    echo "${base}_ldadd = $value"
+  fi
+
+  # CFLAGS
+  # Hack any other program defining a main() replacing its main by
+  # _single_binary_main_$PROGRAM_NAME.
+  echo "${base}_CFLAGS = \"-Dmain=_single_binary_main_${cmd}(int, char**) " \
+       "ATTRIBUTE_NORETURN; int _single_binary_main_${cmd}\" " \
+       "-Dusage=_usage_${cmd} \$(src_coreutils_CFLAGS)"
+  var=src_${cmd}_CFLAGS
+  eval "value=\$$var"
+  if [ "x$value" != "x" ]; then
+    echo "${base}_CFLAGS += $value"
+  fi
+
+  # CPPFLAGS
+  var=src_${cmd}_CPPFLAGS
+  eval "value=\$$var"
+  if [ "x$value" != "x" ]; then
+    echo "${base}_CPPFLAGS = $value"
+  fi
+done
+
+exit 0