version 8.11

* NEWS: Record release date.

.gitignore

@@ -75,6 +75,7 @@
 /src/version.h
 /stamp-h1
 /tests/*/*.log
+/tests/.built-programs
 /tests/t?
 /tests/test-suite.log
 ID

.prev-version

@@ -1 +1 @@
-8.9
+8.10

.x-sc_GPL_version

@@ -1 +0,0 @@
-^build-aux/check\.mk$

.x-sc_bindtextdomain

@@ -1,7 +0,0 @@
-^gl/lib/randint\.c$
-^gl/tests/test-di-set\.c$
-^gl/tests/test-fadvise\.c$
-^gl/tests/test-ino-map\.c$
-^gl/tests/test-mbsalign\.c$
-^gl/tests/test-rand-isaac\.c$
-^lib/euidaccess-stat\.c$

.x-sc_error_message_uppercase

@@ -1 +0,0 @@
-build-aux/cvsu

.x-sc_file_system

@@ -1,5 +0,0 @@
-ChangeLog
-NEWS
-src/df.c
-src/stat.c
-tests/misc/df-P

.x-sc_obsolete_symbols

@@ -1 +0,0 @@
-ChangeLog

.x-sc_po_check

@@ -1 +0,0 @@
-^gl/

.x-sc_program_name

@@ -1,3 +0,0 @@
-gl/lib/randint.c
-lib/euidaccess-stat.c
-gl/tests/.*\.c

.x-sc_prohibit_always-defined_macros

@@ -1 +0,0 @@
-src/seq.c

.x-sc_prohibit_always_true_header_tests

@@ -1,2 +0,0 @@
-^m4/stat-prog\.m4$
-ChangeLog

.x-sc_prohibit_atoi_atof

@@ -1 +0,0 @@
-^lib/euidaccess-stat\.c$

.x-sc_prohibit_empty_lines_at_EOF

@@ -1 +0,0 @@
-^tests/pr/

.x-sc_prohibit_fail_0

@@ -1,4 +0,0 @@
-\.mk$
-/Makefile\.am$
-^tests/init\.sh$
-^tests/test-lib\.sh$

.x-sc_prohibit_magic_number_exit

@@ -1,2 +0,0 @@
-configure.ac
-ChangeLog*

.x-sc_prohibit_stat_st_blocks

@@ -1,4 +0,0 @@
-src/system.h
-tests/du/2g
-old/fileutils/ChangeLog-1997
-ChangeLog-2005

.x-sc_prohibit_strcmp

@@ -1 +0,0 @@
-ChangeLog

.x-sc_prohibit_tab_based_indentation

@@ -1,7 +0,0 @@
-^GNUMakefile$
-Makefile\.am$
-\.mk$
-^tests/pr/
-ChangeLog.*
-^man/help2man$
-^gl/lib/.*\.c\.diff$

.x-sc_require_config_h

@@ -1,9 +0,0 @@
-^lib/buffer-lcm\.c$
-^src/false\.c$
-^src/lbracket\.c$
-^src/ls-dir\.c$
-^src/ls-ls\.c$
-^src/ls-vdir\.c$
-^src/tac-pipe\.c$
-^src/uname-arch\.c$
-^src/uname-uname\.c$

.x-sc_require_config_h_first

@@ -1,9 +0,0 @@
-^lib/buffer-lcm\.c$
-^src/false\.c$
-^src/lbracket\.c$
-^src/ls-dir\.c$
-^src/ls-ls\.c$
-^src/ls-vdir\.c$
-^src/tac-pipe\.c$
-^src/uname-arch\.c$
-^src/uname-uname\.c$

.x-sc_space_tab

@@ -1,12 +0,0 @@
-config(ure|\.(guess|sub))
-tests/pr
-lib/regex.c
-config-log
-tests/misc/nl
-po/de.po
-m4/lib-ld.m4
-m4/lib-prefix.m4
-m4/po.m4
-aclocal.m4
-src/c99-to-c89.diff
-^gl/lib/.*\.c\.diff$

.x-sc_sun_os_names

@@ -1,2 +0,0 @@
-config-log
-build-aux/config.guess

.x-sc_system_h_headers

@@ -1,3 +0,0 @@
-^src/libstdbuf\.c$
-^src/system\.h$
-^src/copy\.h$

.x-sc_trailing_blank

@@ -1 +0,0 @@
-^tests/pr/

.x-sc_unmarked_diagnostics

@@ -1,2 +0,0 @@
-^lib/xstrtol\.h$
-^build-aux/cvsu$

.x-sc_useless_cpp_parens

@@ -1,13 +0,0 @@
-^build-aux/config.guess
-^configure
-^lib/alloca.c
-^lib/fts.c
-^lib/getdate.c
-^lib/getloadavg.c
-^lib/gettext.h
-^lib/getusershell.c
-^lib/mbswidth.c
-^lib/strtod.c
-^lib/xstrtol.c
-^m4/
-^tests/misc/pwd-unreadable-parent

ChangeLog-2005

@@ -9086,7 +9086,7 @@
 
 	split's --verbose option did nothing [broken in 4.5.10 and 5.0]
 	* src/split.c (longopts): Use `1', not `0' as the value for
-	for &verbose.  Reported by Keith Thompson.
+	&verbose.  Reported by Keith Thompson.
 
 	Test for the above fix.
 	* tests/misc/split-a: Also use --verbose and compare stderr

Makefile.am

@@ -39,35 +39,8 @@ changelog_etc =				\
   old/textutils/NEWS			\
   po/ChangeLog-2007
 
-syntax_check_exceptions =		\
-  .x-sc_GPL_version			\
-  .x-sc_bindtextdomain			\
-  .x-sc_error_message_uppercase		\
-  .x-sc_file_system			\
-  .x-sc_obsolete_symbols		\
-  .x-sc_po_check			\
-  .x-sc_program_name			\
-  .x-sc_prohibit_always-defined_macros	\
-  .x-sc_prohibit_atoi_atof		\
-  .x-sc_prohibit_empty_lines_at_EOF	\
-  .x-sc_prohibit_fail_0			\
-  .x-sc_prohibit_magic_number_exit	\
-  .x-sc_prohibit_stat_st_blocks		\
-  .x-sc_prohibit_strcmp			\
-  .x-sc_prohibit_tab_based_indentation	\
-  .x-sc_require_config_h		\
-  .x-sc_require_config_h_first		\
-  .x-sc_space_tab			\
-  .x-sc_sun_os_names			\
-  .x-sc_system_h_headers		\
-  .x-sc_trailing_blank			\
-  .x-sc_prohibit_always_true_header_tests \
-  .x-sc_unmarked_diagnostics		\
-  .x-sc_useless_cpp_parens
-
 EXTRA_DIST =				\
   $(changelog_etc)			\
-  $(syntax_check_exceptions)		\
   .mailmap				\
   .prev-version				\
   .version				\

NEWS

@@ -1,5 +1,61 @@
 GNU coreutils NEWS                                    -*- outline -*-
 
+* Noteworthy changes in release 8.11 (2011-04-13) [stable]
+
+** Bug fixes
+
+  cp -a --link would not create a hardlink to a symlink, instead
+  copying the symlink and then not preserving its timestamp.
+  [bug introduced in coreutils-8.0]
+
+  cp now avoids FIEMAP issues with BTRFS before Linux 2.6.38,
+  which could result in corrupt copies of sparse files.
+  [bug introduced in coreutils-8.10]
+
+  cut could segfault when invoked with a user-specified output
+  delimiter and an unbounded range like "-f1234567890-".
+  [bug introduced in coreutils-5.3.0]
+
+  du would infloop when given --files0-from=DIR
+  [bug introduced in coreutils-7.1]
+
+  sort no longer spawns 7 worker threads to sort 16 lines
+  [bug introduced in coreutils-8.6]
+
+  touch built on Solaris 9 would segfault when run on Solaris 10
+  [bug introduced in coreutils-8.8]
+
+  wc would dereference a NULL pointer upon an early out-of-memory error
+  [bug introduced in coreutils-7.1]
+
+** New features
+
+  dd now accepts the 'nocache' flag to the iflag and oflag options,
+  which will discard any cache associated with the files, or
+  processed portion thereof.
+
+  dd now warns that 'iflag=fullblock' should be used,
+  in various cases where partial reads can cause issues.
+
+** Changes in behavior
+
+  cp now avoids syncing files when possible, when doing a FIEMAP copy.
+  The sync is only needed on Linux kernels before 2.6.39.
+  [The sync was introduced in coreutils-8.10]
+
+  cp now copies empty extents efficiently, when doing a FIEMAP copy.
+  It no longer reads the zero bytes from the input, and also can efficiently
+  create a hole in the output file when --sparse=always is specified.
+
+  df now aligns columns consistently, and no longer wraps entries
+  with longer device identifiers, over two lines.
+
+  install now rejects its long-deprecated --preserve_context option.
+  Use --preserve-context instead.
+
+  test now accepts "==" as a synonym for "="
+
+
 * Noteworthy changes in release 8.10 (2011-02-04) [stable]
 
 ** Bug fixes
@@ -367,7 +423,7 @@ GNU coreutils NEWS                                    -*- outline -*-
   if it uses helper processes for compression and its parent
   ignores CHLD signals. [bug introduced in coreutils-6.9]
 
-  tail without -f no longer access uninitialized memory
+  tail without -f no longer accesses uninitialized memory
   [bug introduced in coreutils-7.6]
 
   timeout is now immune to the signal handling of its parent.
@@ -496,7 +552,7 @@ GNU coreutils NEWS                                    -*- outline -*-
 
   ls -LR exits with status 2, not 0, when it encounters a cycle
 
-  ls -is is now consistent with ls -lis in ignoring values returned
+  "ls -is" is now consistent with ls -lis in ignoring values returned
   from a failed stat/lstat.  For example ls -Lis now prints "?", not "0",
   for the inode number and allocated size of a dereferenced dangling symlink.
 

THANKS.in

@@ -71,6 +71,7 @@ Barry Kelly                         http://barrkel.blogspot.com/
 Bauke Jan Douma                     bjdouma@xs4all.nl
 Ben Elliston                        bje@air.net.au
 Ben Harris                          bjh21@netbsd.org
+Ben Walton                          bwalton@artsci.utoronto.ca
 Bengt Martensson                    bengt@mathematik.uni-Bremen.de
 Benjamin Cutler                     cutlerbc@simla.colostate.edu
 Bernard Giroud                      bernard.giroud@creditlyonnais.ch
@@ -206,7 +207,6 @@ Geoff Whale                         geoffw@cse.unsw.EDU.AU
 Gerald Pfeifer                      gerald@pfeifer.com
 Gerhard Poul                        gpoul@gnu.org
 Germano Leichsenring                germano@jedi.cs.kobe-u.ac.jp
-Gilles Espinasse                    g.esp@free.fr
 Glen Lenker                         glen.lenker@gmail.com
 Göran Uddeborg                      goeran@uddeborg.se
 Guochun Shi                         gshi@ncsa.uiuc.edu
@@ -524,6 +524,7 @@ Soeren Sonnenburg                   sonnenburg@informatik.hu-berlin.de
 Solar Designer                      solar@owl.openwall.com
 Stanislav Ievlev                    inger@altlinux.ru
 Stavros Passas                      stabat@ics.forth.gr
+Stefan Vargyas                      stvar@yahoo.com
 Stéphane Chazelas                   Stephane_CHAZELAS@yahoo.fr
 Stephen Depooter                    sbdep@myrealbox.com
 Stephen Eglen                       eglen@pcg.wustl.edu

bootstrap

@@ -1,6 +1,6 @@
 #! /bin/sh
 # Print a version string.
-scriptversion=2010-11-20.03; # UTC
+scriptversion=2011-03-03.12; # UTC
 
 # Bootstrap this package from checked-out sources.
 
@@ -42,30 +42,32 @@ local_gl_dir=gl
 bt='._bootmp'
 bt_regex=`echo "$bt"| sed 's/\./[.]/g'`
 bt2=${bt}2
+me=$0
 
 usage() {
   cat <<EOF
-Usage: $0 [OPTION]...
+Usage: $me [OPTION]...
 Bootstrap this package from the checked-out sources.
 
 Options:
- --gnulib-srcdir=DIRNAME  Specify the local directory where gnulib
+ --gnulib-srcdir=DIRNAME  specify the local directory where gnulib
                           sources reside.  Use this if you already
                           have gnulib sources on your machine, and
                           do not want to waste your bandwidth downloading
-                          them again.  Defaults to \$GNULIB_SRCDIR.
- --bootstrap-sync         If this bootstrap script is not identical to
+                          them again.  Defaults to \$GNULIB_SRCDIR
+ --bootstrap-sync         if this bootstrap script is not identical to
                           the version in the local gnulib sources,
                           update this script, and then restart it with
-                          /bin/sh or the shell \$CONFIG_SHELL if
-                          defined.
- --no-bootstrap-sync      Do not check whether bootstrap is out of sync.
- --copy                   Copy files instead of creating symbolic links.
- --force                  Attempt to bootstrap even if the sources seem
-                          not to have been checked out.
- --skip-po                Do not download po files.
+                          /bin/sh or the shell \$CONFIG_SHELL
+ --no-bootstrap-sync      do not check whether bootstrap is out of sync
+ --copy                   copy files instead of creating symbolic links
+ --force                  attempt to bootstrap even if the sources seem
+                          not to have been checked out
+ --no-git                 do not use git to update gnulib.  Requires that
+                          --gnulib-srcdir point to a correct gnulib snapshot
+ --skip-po                do not download po files
 
-If the file $0.conf exists in the same directory as this script, its
+If the file $me.conf exists in the same directory as this script, its
 contents are read as shell variables to configure the bootstrap.
 
 For build prerequisites, environment variables like \$AUTOCONF and \$AMTAR
@@ -178,6 +180,9 @@ vc_ignore=auto
 # default.
 bootstrap_sync=false
 
+# Use git to update gnulib sources
+use_git=true
+
 # find_tool ENVVAR NAMES...
 # -------------------------
 # Search for a required program.  Use the value of ENVVAR, if set,
@@ -202,11 +207,11 @@ find_tool ()
     find_tool_error_prefix="\$$find_tool_envvar: "
   fi
   if test x"$find_tool_res" = x; then
-    echo >&2 "$0: one of these is required: $find_tool_names"
+    echo >&2 "$me: one of these is required: $find_tool_names"
     exit 1
   fi
   ($find_tool_res --version </dev/null) >/dev/null 2>&1 || {
-    echo >&2 "$0: ${find_tool_error_prefix}cannot run $find_tool_res --version"
+    echo >&2 "$me: ${find_tool_error_prefix}cannot run $find_tool_res --version"
     exit 1
   }
   eval "$find_tool_envvar=\$find_tool_res"
@@ -253,12 +258,21 @@ do
     bootstrap_sync=true;;
   --no-bootstrap-sync)
     bootstrap_sync=false;;
+  --no-git)
+    use_git=false;;
   *)
     echo >&2 "$0: $option: unknown option"
     exit 1;;
   esac
 done
 
+if $use_git || test -d "$GNULIB_SRCDIR"; then
+  :
+else
+  echo "$0: Error: --no-git requires --gnulib-srcdir" >&2
+  exit 1
+fi
+
 if test -n "$checkout_only_file" && test ! -r "$checkout_only_file"; then
   echo "$0: Bootstrapping from a non-checked-out distribution is risky." >&2
   exit 1
@@ -384,18 +398,22 @@ check_versions() {
     if test "$app" = libtool; then
       app=libtoolize
     fi
+    # Exempt git if --no-git is in effect.
+    if test "$app" = git; then
+      $use_git || continue
+    fi
     # Honor $APP variables ($TAR, $AUTOCONF, etc.)
     appvar=`echo $app | tr '[a-z]-' '[A-Z]_'`
     test "$appvar" = TAR && appvar=AMTAR
     eval "app=\${$appvar-$app}"
     inst_ver=$(get_version $app)
     if [ ! "$inst_ver" ]; then
-      echo "Error: '$app' not found" >&2
+      echo "$me: Error: '$app' not found" >&2
       ret=1
     elif [ ! "$req_ver" = "-" ]; then
       latest_ver=$(sort_ver $req_ver $inst_ver | cut -d' ' -f2)
       if [ ! "$latest_ver" = "$inst_ver" ]; then
-        echo "Error: '$app' version == $inst_ver is too old" >&2
+        echo "$me: Error: '$app' version == $inst_ver is too old" >&2
         echo "       '$app' version >= $req_ver is required" >&2
         ret=1
       fi
@@ -428,9 +446,9 @@ fi
 if ! printf "$buildreq" | check_versions; then
   echo >&2
   if test -f README-prereq; then
-    echo "See README-prereq for how to get the prerequisite programs" >&2
+    echo "$0: See README-prereq for how to get the prerequisite programs" >&2
   else
-    echo "Please install the prerequisite programs" >&2
+    echo "$0: Please install the prerequisite programs" >&2
   fi
   exit 1
 fi
@@ -442,11 +460,11 @@ if test -d .git && (git --version) >/dev/null 2>/dev/null ; then
   if git config merge.merge-changelog.driver >/dev/null ; then
     :
   elif (git-merge-changelog --version) >/dev/null 2>/dev/null ; then
-    echo "initializing git-merge-changelog driver"
+    echo "$0: initializing git-merge-changelog driver"
     git config merge.merge-changelog.name 'GNU-style ChangeLog merge driver'
     git config merge.merge-changelog.driver 'git-merge-changelog %O %A %B'
   else
-    echo "consider installing git-merge-changelog from gnulib"
+    echo "$0: consider installing git-merge-changelog from gnulib"
   fi
 fi
 
@@ -462,7 +480,7 @@ git_modules_config () {
 }
 
 gnulib_path=`git_modules_config submodule.gnulib.path`
-: ${gnulib_path=gnulib}
+test -z "$gnulib_path" && gnulib_path=gnulib
 
 # Get gnulib files.
 
@@ -533,7 +551,7 @@ gnulib_tool=$GNULIB_SRCDIR/gnulib-tool
 download_po_files() {
   subdir=$1
   domain=$2
-  echo "$0: getting translations into $subdir for $domain..."
+  echo "$me: getting translations into $subdir for $domain..."
   cmd=`printf "$po_download_command_format" "$domain" "$subdir"`
   eval "$cmd"
 }
@@ -567,7 +585,7 @@ update_po_files() {
         ! test -f "$po_dir/$po.po" ||
         ! $SHA1SUM -c --status "$cksum_file" \
             < "$new_po" > /dev/null; then
-      echo "updated $po_dir/$po.po..."
+      echo "$me: updated $po_dir/$po.po..."
       cp "$new_po" "$po_dir/$po.po" \
           && $SHA1SUM < "$new_po" > "$cksum_file"
     fi
@@ -612,13 +630,13 @@ symlink_to_dir()
     if $copy; then
       {
         test ! -h "$dst" || {
-          echo "$0: rm -f $dst" &&
+          echo "$me: rm -f $dst" &&
           rm -f "$dst"
         }
       } &&
       test -f "$dst" &&
       cmp -s "$src" "$dst" || {
-        echo "$0: cp -fp $src $dst" &&
+        echo "$me: cp -fp $src $dst" &&
         cp -fp "$src" "$dst"
       }
     else
@@ -632,7 +650,7 @@ symlink_to_dir()
         *)
           case /$dst/ in
           *//* | */../* | */./* | /*/*/*/*/*/)
-             echo >&2 "$0: invalid symlink calculation: $src -> $dst"
+             echo >&2 "$me: invalid symlink calculation: $src -> $dst"
              exit 1;;
           /*/*/*/*/)	dot_dots=../../../;;
           /*/*/*/)	dot_dots=../../;;
@@ -640,7 +658,7 @@ symlink_to_dir()
           esac;;
         esac
 
-        echo "$0: ln -fs $dot_dots$src $dst" &&
+        echo "$me: ln -fs $dot_dots$src $dst" &&
         ln -fs "$dot_dots$src" "$dst"
       }
     fi
@@ -673,7 +691,7 @@ cp_mark_as_generated()
       cmp -s "$cp_src" "$cp_dst" || {
         # Copy the file first to get proper permissions if it
         # doesn't already exist.  Then overwrite the copy.
-        echo "$0: cp -f $cp_src $cp_dst" &&
+        echo "$me: cp -f $cp_src $cp_dst" &&
         rm -f "$cp_dst" &&
         cp "$cp_src" "$cp_dst-t" &&
         sed "s!$bt_regex/!!g" "$cp_src" > "$cp_dst-t" &&
@@ -691,7 +709,7 @@ cp_mark_as_generated()
       if cmp -s "$cp_dst-t" "$cp_dst"; then
         rm -f "$cp_dst-t"
       else
-        echo "$0: cp $cp_src $cp_dst # with edits" &&
+        echo "$me: cp $cp_src $cp_dst # with edits" &&
         mv -f "$cp_dst-t" "$cp_dst"
       fi
     fi
@@ -710,7 +728,7 @@ version_controlled_file() {
   elif test -d .svn; then
     svn log -r HEAD "$dir/$file" > /dev/null 2>&1 && found=yes
   else
-    echo "$0: no version control for $dir/$file?" >&2
+    echo "$me: no version control for $dir/$file?" >&2
   fi
   test $found = yes
 }
@@ -734,18 +752,18 @@ slurp() {
         remove_intl='/^[^#].*\/intl/s/^/#/;'"s!$bt_regex/!!g"
         sed "$remove_intl" $1/$dir/$file |
         cmp - $dir/$gnulib_mk > /dev/null || {
-          echo "$0: Copying $1/$dir/$file to $dir/$gnulib_mk ..." &&
+          echo "$me: Copying $1/$dir/$file to $dir/$gnulib_mk ..." &&
           rm -f $dir/$gnulib_mk &&
           sed "$remove_intl" $1/$dir/$file >$dir/$gnulib_mk &&
           gnulib_mk_hook $dir/$gnulib_mk
         }
       elif { test "${2+set}" = set && test -r $2/$dir/$file; } ||
            version_controlled_file $dir $file; then
-        echo "$0: $dir/$file overrides $1/$dir/$file"
+        echo "$me: $dir/$file overrides $1/$dir/$file"
       else
         copied=$copied$sep$file; sep=$nl
         if test $file = gettext.m4; then
-          echo "$0: patching m4/gettext.m4 to remove need for intl/* ..."
+          echo "$me: patching m4/gettext.m4 to remove need for intl/* ..."
           rm -f $dir/$file
           sed '
             /^AC_DEFUN(\[AM_INTL_SUBDIR],/,/^]/c\
@@ -856,7 +874,7 @@ grep -E '^[	 ]*AC_CONFIG_HEADERS?\>' configure.ac >/dev/null ||
 
 for command in \
   libtool \
-  "${ACLOCAL-aclocal} --force -I m4 $ACLOCAL_FLAGS" \
+  "${ACLOCAL-aclocal} --force -I '$m4_base' $ACLOCAL_FLAGS" \
   "${AUTOCONF-autoconf} --force" \
   "${AUTOHEADER-autoheader} --force" \
   "${AUTOMAKE-automake} --add-missing --copy --force-missing"
@@ -867,7 +885,7 @@ do
     command="${LIBTOOLIZE-libtoolize} -c -f"
   fi
   echo "$0: $command ..."
-  $command || exit
+  eval "$command" || exit
 done
 
 

bootstrap.conf

@@ -232,6 +232,7 @@ gnulib_modules="
   verror
   version-etc-fsf
   wcwidth
+  wcswidth
   winsz-ioctl
   winsz-termios
   write-any-file
@@ -299,7 +300,6 @@ if test $gettext_external = 1; then
   excluded_files='
       m4/glibc2.m4
       m4/intdiv0.m4
-      m4/lcmessage.m4
       m4/uintmax_t.m4
       m4/ulonglong.m4
       m4/visibility.m4

cfg.mk

@@ -27,7 +27,7 @@ bootstrap-tools = autoconf,automake,gnulib,bison
 # Now that we have better tests, make this the default.
 export VERBOSE = yes
 
-old_NEWS_hash = f2ebf9f1f16209f7a4b9927a755956fa
+old_NEWS_hash = 99b06e7bb289add96b10127fceced4e9
 
 # Add an exemption for sc_makefile_at_at_check.
 _makefile_at_at_check_exceptions = ' && !/^cu_install_program =/'
@@ -39,8 +39,8 @@ _hv_file ?= $(srcdir)/tests/misc/help-version
 dd = $(srcdir)/src/dd.c
 sc_dd_O_FLAGS:
 	@rm -f $@.1 $@.2
-	@{ echo O_FULLBLOCK; perl -nle '/^ +\| (O_\w*)$$/ and print $$1' \
-	  $(dd); } | sort > $@.1
+	@{ echo O_FULLBLOCK; echo O_NOCACHE;				\
+	  perl -nle '/^ +\| (O_\w*)$$/ and print $$1' $(dd); } | sort > $@.1
 	@{ echo O_NOFOLLOW; perl -nle '/{"[a-z]+",\s*(O_\w+)},/ and print $$1' \
 	  $(dd); } | sort > $@.2
 	@diff -u $@.1 $@.2 || diff=1 || diff=;				\
@@ -93,19 +93,6 @@ sc_root_tests:
 	       exit 1; } || :;						\
 	fi
 
-# Ensure that the syntax_check_exceptions file list in Makefile.am
-# stays in sync with corresponding files in the repository.
-sce = syntax_check_exceptions
-sc_x_sc_dist_check:
-	@test "$$( ($(VC_LIST) | sed -n '/\.x-sc_/p'			\
-		     | sed 's|^$(_dot_escaped_srcdir)/||';		\
-		   sed -n '/^$(sce) =[	 ]*\\$$/,/[^\]$$/p'		\
-		     $(srcdir)/Makefile.am				\
-		       | sed 's/^  *//;/^$(sce) =/d'			\
-		       | tr -s '\012\\' '  ' | fmt -1			\
-		   ) | sort | uniq -u)"					\
-	  && { echo 'Makefile.am: $(sce) mismatch' >&2; exit 1; } || :;
-
 # Create a list of regular expressions matching the names
 # of files included from system.h.  Exclude a couple.
 .re-list:
@@ -186,7 +173,7 @@ sc_NEWS_two_empty_lines:
 	@sed -n 4,/Noteworthy/p $(srcdir)/NEWS				\
 	    | perl -n0e '/(^|\n)\n\n\* Noteworthy/ or exit 1'		\
 	  || { echo '$(ME): use two empty lines to separate NEWS sections' \
-		 1>&2; exit 1; } || :;					\
+		 1>&2; exit 1; } || :
 
 # Perl-based tests used to exec perl from a #!/bin/sh script.
 # Now they all start with #!/usr/bin/perl and the portability
@@ -328,6 +315,15 @@ sc_space_before_open_paren:
 	else :;								\
 	fi
 
+# Similar to the gnulib maint.mk rule for sc_prohibit_strcmp
+# Use STREQ_LEN or STRPREFIX rather than comparing strncmp == 0, or != 0.
+sc_prohibit_strncmp:
+	@grep -nE '! *str''ncmp *\(|\<str''ncmp *\(.+\) *[!=]='		\
+	    $$($(VC_LIST_EXCEPT))					\
+	  | grep -vE ':# *define STR(N?EQ_LEN|PREFIX)\(' &&		\
+	  { echo '$(ME): use STREQ_LEN or STRPREFIX instead of str''ncmp' \
+		1>&2; exit 1; } || :
+
 # Override the default Cc: used in generating an announcement.
 announcement_Cc_ = $(translation_project_), \
   coreutils@gnu.org, coreutils-announce@gnu.org
@@ -337,3 +333,36 @@ include $(srcdir)/dist-check.mk
 update-copyright-env = \
   UPDATE_COPYRIGHT_USE_INTERVALS=1 \
   UPDATE_COPYRIGHT_MAX_LINE_LENGTH=79
+
+# List syntax-check exemptions.
+exclude_file_name_regexp--sc_space_tab = \
+  ^(tests/pr/|tests/misc/nl$$|gl/.*\.diff$$)
+exclude_file_name_regexp--sc_bindtextdomain = ^(gl/.*|lib/euidaccess-stat)\.c$$
+exclude_file_name_regexp--sc_unmarked_diagnostics =    ^build-aux/cvsu$$
+exclude_file_name_regexp--sc_error_message_uppercase = ^build-aux/cvsu$$
+exclude_file_name_regexp--sc_trailing_blank = ^tests/pr/
+exclude_file_name_regexp--sc_system_h_headers = \
+  ^src/((system|copy)\.h|libstdbuf\.c)$$
+
+_src = (false|lbracket|ls-(dir|ls|vdir)|tac-pipe|uname-(arch|uname))
+exclude_file_name_regexp--sc_require_config_h_first = \
+  (^lib/buffer-lcm\.c|src/$(_src)\.c)$$
+exclude_file_name_regexp--sc_require_config_h = \
+  $(exclude_file_name_regexp--sc_require_config_h_first)
+
+exclude_file_name_regexp--sc_po_check = ^gl/
+exclude_file_name_regexp--sc_prohibit_always-defined_macros = ^src/seq\.c$$
+exclude_file_name_regexp--sc_prohibit_empty_lines_at_EOF = ^tests/pr/
+exclude_file_name_regexp--sc_program_name = ^(gl/.*|lib/euidaccess-stat)\.c$$
+exclude_file_name_regexp--sc_file_system = \
+  NEWS|^(tests/init\.cfg|src/df\.c|tests/misc/df-P)$$
+exclude_file_name_regexp--sc_prohibit_always_true_header_tests = \
+  ^m4/stat-prog\.m4$$
+exclude_file_name_regexp--sc_prohibit_fail_0 = \
+  (^tests/init\.sh|Makefile\.am|\.mk)$$
+exclude_file_name_regexp--sc_prohibit_atoi_atof = ^lib/euidaccess-stat\.c$$
+exclude_file_name_regexp--sc_prohibit_tab_based_indentation = \
+  ^tests/pr/|(^gl/lib/reg.*\.c\.diff|Makefile(\.am)?|\.mk|^man/help2man)$$
+
+exclude_file_name_regexp--sc_prohibit_stat_st_blocks = \
+  ^(src/system\.h|tests/du/2g)$$

configure.ac

@@ -162,18 +162,18 @@ int main()
   time_t now = time ((time_t *) 0);
   int hour_GMT0, hour_unset;
   if (putenv ("TZ=GMT0") != 0)
-    exit (1);
+    return 1;
   hour_GMT0 = localtime (&now)->tm_hour;
   unset_TZ ();
   hour_unset = localtime (&now)->tm_hour;
   if (putenv ("TZ=PST8") != 0)
-    exit (1);
+    return 1;
   if (localtime (&now)->tm_hour == hour_GMT0)
-    exit (1);
+    return 1;
   unset_TZ ();
   if (localtime (&now)->tm_hour != hour_unset)
-    exit (1);
-  exit (0);
+    return 1;
+  return 0;
 }]])],
 [utils_cv_localtime_cache=no],
 [utils_cv_localtime_cache=yes],

doc/coreutils.texi

@@ -1455,10 +1455,11 @@ The @sc{gnu} utilities normally conform to the version of @acronym{POSIX}
 that is standard for your system.  To cause them to conform to a
 different version of @acronym{POSIX}, define the @env{_POSIX2_VERSION}
 environment variable to a value of the form @var{yyyymm} specifying
-the year and month the standard was adopted.  Two values are currently
+the year and month the standard was adopted.  Three values are currently
 supported for @env{_POSIX2_VERSION}: @samp{199209} stands for
-@acronym{POSIX} 1003.2-1992, and @samp{200112} stands for @acronym{POSIX}
-1003.1-2001.  For example, if you have a newer system but are running software
+@acronym{POSIX} 1003.2-1992, @samp{200112} stands for @acronym{POSIX}
+1003.1-2001, and @samp{200809} stands for @acronym{POSIX} 1003.1-2008.
+For example, if you have a newer system but are running software
 that assumes an older version of @acronym{POSIX} and uses @samp{sort +1}
 or @samp{tail +10}, you can work around any compatibility problems by setting
 @samp{_POSIX2_VERSION=199209} in your environment.
@@ -7971,8 +7972,8 @@ Set both input and output block sizes to @var{bytes}.
 This makes @command{dd} read and write @var{bytes} per block,
 overriding any @samp{ibs} and @samp{obs} settings.
 In addition, if no data-transforming @option{conv} option is specified,
-each input block is copied to the output as a single block,
-without aggregating short reads.
+input is copied to the output as soon as it's read,
+even if it is smaller than the block size.
 
 @item cbs=@var{bytes}
 @opindex cbs
@@ -8062,22 +8063,29 @@ Swap every pair of input bytes.  @sc{gnu} @command{dd}, unlike others, works
 when an odd number of bytes are read---the last byte is simply copied
 (since there is nothing to swap it with).
 
-@item noerror
-@opindex noerror
-@cindex read errors, ignoring
-Continue after read errors.
+@item sync
+@opindex sync @r{(padding with @acronym{ASCII} @sc{nul}s)}
+Pad every input block to size of @samp{ibs} with trailing zero bytes.
+When used with @samp{block} or @samp{unblock}, pad with spaces instead of
+zero bytes.
 
-@item nocreat
-@opindex nocreat
-@cindex creating output file, avoiding
-Do not create the output file; the output file must already exist.
+@end table
 
+The following ``conversions'' are really file flags
+and don't affect internal processing:
+
+@table @samp
 @item excl
 @opindex excl
 @cindex creating output file, requiring
 Fail if the output file already exists; @command{dd} must create the
 output file itself.
 
+@item nocreat
+@opindex nocreat
+@cindex creating output file, avoiding
+Do not create the output file; the output file must already exist.
+
 The @samp{excl} and @samp{nocreat} conversions are mutually exclusive.
 
 @item notrunc
@@ -8085,11 +8093,10 @@ The @samp{excl} and @samp{nocreat} conversions are mutually exclusive.
 @cindex truncating output file, avoiding
 Do not truncate the output file.
 
-@item sync
-@opindex sync @r{(padding with @acronym{ASCII} @sc{nul}s)}
-Pad every input block to size of @samp{ibs} with trailing zero bytes.
-When used with @samp{block} or @samp{unblock}, pad with spaces instead of
-zero bytes.
+@item noerror
+@opindex noerror
+@cindex read errors, ignoring
+Continue after read errors.
 
 @item fdatasync
 @opindex fdatasync
@@ -8168,6 +8175,31 @@ last-access and last-modified time) is not necessarily synchronized.
 @cindex synchronized data and metadata I/O
 Use synchronized I/O for both data and metadata.
 
+@item nocache
+@opindex nocache
+@cindex discarding file cache
+Discard the data cache for a file.
+When count=0 all cache is discarded,
+otherwise the cache is dropped for the processed
+portion of the file.  Also when count=0
+failure to discard the cache is diagnosed
+and reflected in the exit status.
+Here as some usage examples:
+
+@example
+# Advise to drop cache for whole file
+dd if=ifile iflag=nocache count=0
+
+# Ensure drop cache for the whole file
+dd of=ofile oflag=nocache conv=notrunc,fdatasync count=0
+
+# Drop cache for part of file
+dd if=ifile iflag=nocache skip=10 count=10 of=/dev/null
+
+# Stream data using just the read-ahead cache
+dd if=ifile of=ofile iflag=nocache oflag=nocache
+@end example
+
 @item nonblock
 @opindex nonblock
 @cindex nonblocking I/O
@@ -11504,7 +11536,7 @@ Exit status:
 * File type tests::             -[bcdfhLpSt]
 * Access permission tests::     -[gkruwxOG]
 * File characteristic tests::   -e -s -nt -ot -ef
-* String tests::                -z -n = !=
+* String tests::                -z -n = == !=
 * Numeric tests::               -eq -ne -lt -le -gt -ge
 * Connectives for test::        ! -a -o
 @end menu
@@ -11695,6 +11727,11 @@ True if the length of @var{string} is nonzero.
 @cindex equal string check
 True if the strings are equal.
 
+@item @var{string1} == @var{string2}
+@opindex ==
+@cindex equal string check
+True if the strings are equal (synonym for =).
+
 @item @var{string1} != @var{string2}
 @opindex !=
 @cindex not-equal string check

gl/lib/di-set.c

@@ -1,237 +0,0 @@
-/* Set operations for device-inode pairs stored in a space-efficient manner.
-
-   Copyright 2009-2011 Free Software Foundation, Inc.
-
-   This program is free software: you can redistribute it and/or modify
-   it under the terms of the GNU General Public License as published by
-   the Free Software Foundation; either version 3 of the License, or
-   (at your option) any later version.
-
-   This program is distributed in the hope that it will be useful,
-   but WITHOUT ANY WARRANTY; without even the implied warranty of
-   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-   GNU General Public License for more details.
-
-   You should have received a copy of the GNU General Public License
-   along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
-
-/* written by Paul Eggert and Jim Meyering */
-
-#include <config.h>
-#include "di-set.h"
-
-#include "hash.h"
-#include "ino-map.h"
-
-#include <limits.h>
-#include <stdlib.h>
-
-/* The hash package hashes "void *", but this package wants to hash
-   integers.  Use integers that are as large as possible, but no
-   larger than void *, so that they can be cast to void * and back
-   without losing information.  */
-typedef size_t hashint;
-#define HASHINT_MAX ((hashint) -1)
-
-/* Integers represent inode numbers.  Integers in the range
-   1..(LARGE_INO_MIN-1) represent inode numbers directly.  (The hash
-   package does not work with null pointers, so inode 0 cannot be used
-   as a key.)  To find the representations of other inode numbers, map
-   them through INO_MAP.  */
-#define LARGE_INO_MIN (HASHINT_MAX / 2)
-
-/* Set operations for device-inode pairs stored in a space-efficient
-   manner.  Use a two-level hash table.  The top level hashes by
-   device number, as there are typically a small number of devices.
-   The lower level hashes by mapped inode numbers.  In the typical
-   case where the inode number is positive and small, the inode number
-   maps to itself, masquerading as a void * value; otherwise, its
-   value is the result of hashing the inode value through INO_MAP.  */
-
-/* A pair that maps a device number to a set of inode numbers.  */
-struct di_ent
-{
-  dev_t dev;
-  struct hash_table *ino_set;
-};
-
-/* A two-level hash table that manages and indexes these pairs.  */
-struct di_set
-{
-  /* Map device numbers to sets of inode number representatives.  */
-  struct hash_table *dev_map;
-
-  /* If nonnull, map large inode numbers to their small
-     representatives.  If null, there are no large inode numbers in
-     this set.  */
-  struct ino_map *ino_map;
-
-  /* Cache of the most recently allocated and otherwise-unused storage
-     for probing this table.  */
-  struct di_ent *probe;
-};
-
-/* Hash a device-inode-set entry.  */
-static size_t
-di_ent_hash (void const *x, size_t table_size)
-{
-  struct di_ent const *p = x;
-  dev_t dev = p->dev;
-
-  /* When DEV is wider than size_t, exclusive-OR the words of DEV into H.
-     This avoids loss of info, without applying % to the wider type,
-     which could be quite slow on some systems.  */
-  size_t h = dev;
-  unsigned int i;
-  unsigned int n_words = sizeof dev / sizeof h + (sizeof dev % sizeof h != 0);
-  for (i = 1; i < n_words; i++)
-    h ^= dev >> CHAR_BIT * sizeof h * i;
-
-  return h % table_size;
-}
-
-/* Return true if two device-inode-set entries are the same.  */
-static bool
-di_ent_compare (void const *x, void const *y)
-{
-  struct di_ent const *a = x;
-  struct di_ent const *b = y;
-  return a->dev == b->dev;
-}
-
-/* Free a device-inode-set entry.  */
-static void
-di_ent_free (void *v)
-{
-  struct di_ent *a = v;
-  hash_free (a->ino_set);
-  free (a);
-}
-
-/* Create a set of device-inode pairs.  Return NULL on allocation failure.  */
-struct di_set *
-di_set_alloc (void)
-{
-  struct di_set *dis = malloc (sizeof *dis);
-  if (dis)
-    {
-      enum { INITIAL_DEV_MAP_SIZE = 11 };
-      dis->dev_map = hash_initialize (INITIAL_DEV_MAP_SIZE, NULL,
-                                      di_ent_hash, di_ent_compare,
-                                      di_ent_free);
-      if (! dis->dev_map)
-        {
-          free (dis);
-          return NULL;
-        }
-      dis->ino_map = NULL;
-      dis->probe = NULL;
-    }
-
-  return dis;
-}
-
-/* Free a set of device-inode pairs.  */
-void
-di_set_free (struct di_set *dis)
-{
-  hash_free (dis->dev_map);
-  free (dis->ino_map);
-  free (dis->probe);
-  free (dis);
-}
-
-/* Hash an encoded inode number I.  */
-static size_t
-di_ino_hash (void const *i, size_t table_size)
-{
-  return (hashint) i % table_size;
-}
-
-/* Using the DIS table, map a device to a hash table that represents
-   a set of inode numbers.  Return NULL on error.  */
-static struct hash_table *
-map_device (struct di_set *dis, dev_t dev)
-{
-  /* Find space for the probe, reusing the cache if available.  */
-  struct di_ent *ent;
-  struct di_ent *probe = dis->probe;
-  if (probe)
-    {
-      /* If repeating a recent query, return the cached result.   */
-      if (probe->dev == dev)
-        return probe->ino_set;
-    }
-  else
-    {
-      dis->probe = probe = malloc (sizeof *probe);
-      if (! probe)
-        return NULL;
-    }
-
-  /* Probe for the device.  */
-  probe->dev = dev;
-  ent = hash_insert (dis->dev_map, probe);
-  if (! ent)
-    return NULL;
-
-  if (ent != probe)
-    {
-      /* Use the existing entry.  */
-      probe->ino_set = ent->ino_set;
-    }
-  else
-    {
-      enum { INITIAL_INO_SET_SIZE = 1021 };
-
-      /* Prepare to allocate a new probe next time; this one is in use.  */
-      dis->probe = NULL;
-
-      /* DEV is new; allocate an inode set for it.  */
-      probe->ino_set = hash_initialize (INITIAL_INO_SET_SIZE, NULL,
-                                        di_ino_hash, NULL, NULL);
-    }
-
-  return probe->ino_set;
-}
-
-/* Using the DIS table, map an inode number to a mapped value.
-   Return INO_MAP_INSERT_FAILURE on error.  */
-static hashint
-map_inode_number (struct di_set *dis, ino_t ino)
-{
-  if (0 < ino && ino < LARGE_INO_MIN)
-    return ino;
-
-  if (! dis->ino_map)
-    {
-      dis->ino_map = ino_map_alloc (LARGE_INO_MIN);
-      if (! dis->ino_map)
-        return INO_MAP_INSERT_FAILURE;
-    }
-
-  return ino_map_insert (dis->ino_map, ino);
-}
-
-/* Attempt to insert the DEV,INO pair into the set DIS.
-   If it matches a pair already in DIS, keep that pair and return 0.
-   Otherwise, if insertion is successful, return 1.
-   Upon any failure return -1.  */
-int
-di_set_insert (struct di_set *dis, dev_t dev, ino_t ino)
-{
-  hashint i;
-
-  /* Map the device number to a set of inodes.  */
-  struct hash_table *ino_set = map_device (dis, dev);
-  if (! ino_set)
-    return -1;
-
-  /* Map the inode number to a small representative I.  */
-  i = map_inode_number (dis, ino);
-  if (i == INO_MAP_INSERT_FAILURE)
-    return -1;
-
-  /* Put I into the inode set.  */
-  return hash_insert0 (ino_set, (void *) i, NULL);
-}

gl/lib/di-set.h

@@ -1,12 +0,0 @@
-#include <sys/types.h>
-
-#undef _ATTRIBUTE_NONNULL_
-#if __GNUC__ == 3 && __GNUC_MINOR__ >= 3 || 3 < __GNUC__
-# define _ATTRIBUTE_NONNULL_(m) __attribute__ ((__nonnull__ (m)))
-#else
-# define _ATTRIBUTE_NONNULL_(m)
-#endif
-
-struct di_set *di_set_alloc (void);
-int di_set_insert (struct di_set *, dev_t, ino_t) _ATTRIBUTE_NONNULL_ (1);
-void di_set_free (struct di_set *) _ATTRIBUTE_NONNULL_ (1);

gl/lib/ino-map.c

@@ -1,164 +0,0 @@
-/* Map an ino_t inode number to a small integer.
-
-   Copyright 2009-2011 Free Software Foundation, Inc.
-
-   This program is free software: you can redistribute it and/or modify
-   it under the terms of the GNU General Public License as published by
-   the Free Software Foundation; either version 3 of the License, or
-   (at your option) any later version.
-
-   This program is distributed in the hope that it will be useful,
-   but WITHOUT ANY WARRANTY; without even the implied warranty of
-   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-   GNU General Public License for more details.
-
-   You should have received a copy of the GNU General Public License
-   along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
-
-/* written by Paul Eggert and Jim Meyering */
-
-#include <config.h>
-#include "ino-map.h"
-
-#include "hash.h"
-#include "verify.h"
-
-#include <limits.h>
-#include <stdlib.h>
-
-/* A pair that maps an inode number to a mapped inode number; the
-   latter is a small unique ID for the former.  */
-struct ino_map_ent
-{
-  ino_t ino;
-  size_t mapped_ino;
-};
-
-/* A table that manages and indexes these pairs.  */
-struct ino_map
-{
-  /* A table of KEY,VAL pairs, where KEY is the raw ino_t value and
-     VAL is the small number that it maps to.  */
-  struct hash_table *map;
-
-  /* The next mapped inode number to hand out.  */
-  size_t next_mapped_ino;
-
-  /* Cache of the most recently allocated and otherwise-unused storage
-     for probing the table.  */
-  struct ino_map_ent *probe;
-};
-
-/* Hash an inode map entry.  */
-static size_t
-ino_hash (void const *x, size_t table_size)
-{
-  struct ino_map_ent const *p = x;
-  ino_t ino = p->ino;
-
-  /* When INO is wider than size_t, exclusive-OR the words of INO into H.
-     This avoids loss of info, without applying % to the wider type,
-     which could be quite slow on some systems.  */
-  size_t h = ino;
-  unsigned int i;
-  unsigned int n_words = sizeof ino / sizeof h + (sizeof ino % sizeof h != 0);
-  for (i = 1; i < n_words; i++)
-    h ^= ino >> CHAR_BIT * sizeof h * i;
-
-  return h % table_size;
-}
-
-/* Return true if two inode map entries are the same.  */
-static bool
-ino_compare (void const *x, void const *y)
-{
-  struct ino_map_ent const *a = x;
-  struct ino_map_ent const *b = y;
-  return a->ino == b->ino;
-}
-
-/* Allocate an inode map that will hand out integers starting with
-   NEXT_MAPPED_INO.  Return NULL if memory is exhausted.  */
-struct ino_map *
-ino_map_alloc (size_t next_mapped_ino)
-{
-  struct ino_map *im = malloc (sizeof *im);
-
-  if (im)
-    {
-      enum { INITIAL_INO_MAP_TABLE_SIZE = 1021 };
-      im->map = hash_initialize (INITIAL_INO_MAP_TABLE_SIZE, NULL,
-                                 ino_hash, ino_compare, free);
-      if (! im->map)
-        {
-          free (im);
-          return NULL;
-        }
-      im->next_mapped_ino = next_mapped_ino;
-      im->probe = NULL;
-    }
-
-  return im;
-}
-
-/* Free an inode map.  */
-void
-ino_map_free (struct ino_map *map)
-{
-  hash_free (map->map);
-  free (map->probe);
-  free (map);
-}
-
-
-/* Insert into MAP the inode number INO if it's not there already,
-   and return its nonnegative mapped inode number.
-   If INO is already in MAP, return the existing mapped inode number.
-   Return INO_MAP_INSERT_FAILURE on memory or counter exhaustion.  */
-size_t
-ino_map_insert (struct ino_map *im, ino_t ino)
-{
-  struct ino_map_ent *ent;
-
-  /* Find space for the probe, reusing the cache if available.  */
-  struct ino_map_ent *probe = im->probe;
-  if (probe)
-    {
-      /* If repeating a recent query, return the cached result.   */
-      if (probe->ino == ino)
-        return probe->mapped_ino;
-    }
-  else
-    {
-      im->probe = probe = malloc (sizeof *probe);
-      if (! probe)
-        return INO_MAP_INSERT_FAILURE;
-    }
-
-  probe->ino = ino;
-  ent = hash_insert (im->map, probe);
-  if (! ent)
-    return INO_MAP_INSERT_FAILURE;
-
-  if (ent != probe)
-    {
-      /* Use the existing entry.  */
-      probe->mapped_ino = ent->mapped_ino;
-    }
-  else
-    {
-      /* If adding 1 to map->next_mapped_ino would cause it to
-         overflow to zero, then it must equal INO_MAP_INSERT_FAILURE,
-         which is the value that should be returned in that case.
-         Verify that this works.  */
-      verify (INO_MAP_INSERT_FAILURE + 1 == 0);
-
-      /* Prepare to allocate a new probe next time; this one is in use.  */
-      im->probe = NULL;
-
-      /* INO is new; allocate a mapped inode number for it.  */
-      probe->mapped_ino = im->next_mapped_ino++;
-    }
-
-  return probe->mapped_ino;
-}

gl/lib/ino-map.h

@@ -1,14 +0,0 @@
-#include <sys/types.h>
-
-#undef _ATTRIBUTE_NONNULL_
-#if __GNUC__ == 3 && __GNUC_MINOR__ >= 3 || 3 < __GNUC__
-# define _ATTRIBUTE_NONNULL_(m) __attribute__ ((__nonnull__ (m)))
-#else
-# define _ATTRIBUTE_NONNULL_(m)
-#endif
-
-#define INO_MAP_INSERT_FAILURE ((size_t) -1)
-
-struct ino_map *ino_map_alloc (size_t);
-void ino_map_free (struct ino_map *) _ATTRIBUTE_NONNULL_ (1);
-size_t ino_map_insert (struct ino_map *, ino_t) _ATTRIBUTE_NONNULL_ (1);

gl/lib/mbsalign.c

@@ -79,27 +79,6 @@ wc_truncate (wchar_t *wc, size_t width)
   return cells;
 }
 
-/* FIXME: move this function to gnulib as it's missing on:
-   OpenBSD 3.8, IRIX 5.3, Solaris 2.5.1, mingw, BeOS  */
-
-static int
-rpl_wcswidth (const wchar_t *s, size_t n)
-{
-  int ret = 0;
-
-  while (n-- > 0 && *s != L'\0')
-    {
-      int nwidth = wcwidth (*s++);
-      if (nwidth == -1)             /* non printable */
-        return -1;
-      if (ret > (INT_MAX - nwidth)) /* overflow */
-        return -1;
-      ret += nwidth;
-    }
-
-  return ret;
-}
-
 /* Write N_SPACES space characters to DEST while ensuring
    nothing is written beyond DEST_END. A terminating NUL
    is always added to DEST.
@@ -171,7 +150,7 @@ mbsalign (const char *src, char *dest, size_t dest_size,
           str_wc[src_chars - 1] = L'\0';
           wc_enabled = true;
           conversion = wc_ensure_printable (str_wc);
-          n_cols = rpl_wcswidth (str_wc, src_chars);
+          n_cols = wcswidth (str_wc, src_chars);
         }
     }
 

gl/lib/mbsalign.h

@@ -19,7 +19,7 @@
 typedef enum { MBS_ALIGN_LEFT, MBS_ALIGN_RIGHT, MBS_ALIGN_CENTER } mbs_align_t;
 
 enum {
-  /* Use unibyte mode for invalid multibyte strings or
+  /* Use unibyte mode for invalid multibyte strings
      or when heap memory is exhausted.  */
   MBA_UNIBYTE_FALLBACK = 0x0001
 

gl/modules/di-set

@@ -1,24 +0,0 @@
-Description:
-manipulate sets of device-inode pairs efficiently
-
-Files:
-lib/di-set.c
-lib/di-set.h
-
-Depends-on:
-ino-map
-hash
-
-configure.ac:
-
-Makefile.am:
-lib_SOURCES += di-set.c di-set.h
-
-Include:
-"di-set.h"
-
-License
-GPL
-
-Maintainer:
-Jim Meyering

gl/modules/di-set-tests

@@ -1,10 +0,0 @@
-Files:
-tests/test-di-set.c
-
-Depends-on:
-
-configure.ac:
-
-Makefile.am:
-TESTS += test-di-set
-check_PROGRAMS += test-di-set

gl/modules/ino-map

@@ -1,24 +0,0 @@
-Description:
-maintain a mapping of ino_t numbers to small integers
-
-Files:
-lib/ino-map.c
-lib/ino-map.h
-
-Depends-on:
-hash
-verify
-
-configure.ac:
-
-Makefile.am:
-lib_SOURCES += ino-map.c ino-map.h
-
-Include:
-"ino-map.h"
-
-License
-GPL
-
-Maintainer:
-Jim Meyering

gl/modules/ino-map-tests

@@ -1,10 +0,0 @@
-Files:
-tests/test-ino-map.c
-
-Depends-on:
-
-configure.ac:
-
-Makefile.am:
-TESTS += test-ino-map
-check_PROGRAMS += test-ino-map

gl/modules/randread-tests

@@ -1,5 +1,6 @@
 Files:
 tests/test-rand-isaac.c
+tests/macros.h
 
 Depends-on:
 

gl/tests/test-di-set.c

@@ -1,63 +0,0 @@
-/* Test the di-set module.
-   Copyright (C) 2010-2011 Free Software Foundation, Inc.
-
-   This program is free software: you can redistribute it and/or modify
-   it under the terms of the GNU General Public License as published by
-   the Free Software Foundation; either version 3 of the License, or
-   (at your option) any later version.
-
-   This program is distributed in the hope that it will be useful,
-   but WITHOUT ANY WARRANTY; without even the implied warranty of
-   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-   GNU General Public License for more details.
-
-   You should have received a copy of the GNU General Public License
-   along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
-
-/* Written by Jim Meyering.  */
-
-#include <config.h>
-#include <stdlib.h>
-#include <stdio.h>
-#include <string.h>
-#include <stdint.h>
-
-#define ASSERT(expr) \
-  do                                                                         \
-    {                                                                        \
-      if (!(expr))                                                           \
-        {                                                                    \
-          fprintf (stderr, "%s:%d: assertion failed\n", __FILE__, __LINE__); \
-          fflush (stderr);                                                   \
-          abort ();                                                          \
-        }                                                                    \
-    }                                                                        \
-  while (0)
-
-#include "di-set.h"
-
-int
-main (void)
-{
-  struct di_set *dis = di_set_alloc ();
-  ASSERT (dis);
-
-  ASSERT (di_set_insert (dis, 2, 5) == 1); /* first insertion succeeds */
-  ASSERT (di_set_insert (dis, 2, 5) == 0); /* duplicate fails */
-  ASSERT (di_set_insert (dis, 3, 5) == 1); /* diff dev, duplicate inode is ok */
-  ASSERT (di_set_insert (dis, 2, 8) == 1); /* same dev, different inode is ok */
-
-  /* very large (or negative) inode number */
-  ASSERT (di_set_insert (dis, 5, (ino_t) -1) == 1);
-  ASSERT (di_set_insert (dis, 5, (ino_t) -1) == 0); /* dup */
-
-  unsigned int i;
-  for (i = 0; i < 3000; i++)
-    ASSERT (di_set_insert (dis, 9, i) == 1);
-  for (i = 0; i < 3000; i++)
-    ASSERT (di_set_insert (dis, 9, i) == 0); /* duplicate fails */
-
-  di_set_free (dis);
-
-  return 0;
-}

gl/tests/test-ino-map.c

@@ -1,62 +0,0 @@
-/* Test the ino-map module.
-   Copyright (C) 2010-2011 Free Software Foundation, Inc.
-
-   This program is free software: you can redistribute it and/or modify
-   it under the terms of the GNU General Public License as published by
-   the Free Software Foundation; either version 3 of the License, or
-   (at your option) any later version.
-
-   This program is distributed in the hope that it will be useful,
-   but WITHOUT ANY WARRANTY; without even the implied warranty of
-   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-   GNU General Public License for more details.
-
-   You should have received a copy of the GNU General Public License
-   along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
-
-/* Written by Jim Meyering.  */
-
-#include <config.h>
-#include <stdlib.h>
-#include <stdio.h>
-#include <string.h>
-
-/* FIXME: once/if in gnulib, use #include "macros.h" in place of this */
-#define ASSERT(expr) \
-  do                                                                         \
-    {                                                                        \
-      if (!(expr))                                                           \
-        {                                                                    \
-          fprintf (stderr, "%s:%d: assertion failed\n", __FILE__, __LINE__); \
-          fflush (stderr);                                                   \
-          abort ();                                                          \
-        }                                                                    \
-    }                                                                        \
-  while (0)
-
-#include "ino-map.h"
-
-int
-main ()
-{
-  enum { INO_MAP_INIT = 123 };
-  struct ino_map *ino_map = ino_map_alloc (INO_MAP_INIT);
-  ASSERT (ino_map != NULL);
-
-  ASSERT (ino_map_insert (ino_map, 42) == INO_MAP_INIT);
-  ASSERT (ino_map_insert (ino_map, 42) == INO_MAP_INIT);
-  ASSERT (ino_map_insert (ino_map, 398) == INO_MAP_INIT + 1);
-  ASSERT (ino_map_insert (ino_map, 398) == INO_MAP_INIT + 1);
-  ASSERT (ino_map_insert (ino_map, 0) == INO_MAP_INIT + 2);
-  ASSERT (ino_map_insert (ino_map, 0) == INO_MAP_INIT + 2);
-
-  int i;
-  for (i = 0; i < 100; i++)
-    {
-      ASSERT (ino_map_insert (ino_map, 10000 + i) == INO_MAP_INIT + 3 + i);
-    }
-
-  ino_map_free (ino_map);
-
-  return 0;
-}

gl/tests/test-rand-isaac.c

@@ -23,22 +23,9 @@
 
 #include <fcntl.h>
 #include <stdint.h>
-#include <stdlib.h>
-#include <stdio.h>
 #include <string.h>
 
-/* FIXME: once/if in gnulib, use #include "macros.h" in place of this */
-#define ASSERT(expr) \
-  do                                                                         \
-    {                                                                        \
-      if (!(expr))                                                           \
-        {                                                                    \
-          fprintf (stderr, "%s:%d: assertion failed\n", __FILE__, __LINE__); \
-          fflush (stderr);                                                   \
-          abort ();                                                          \
-        }                                                                    \
-    }                                                                        \
-  while (0)
+#include "macros.h"
 
 /* This expected output was generated by running the programs in
    <http://burtleburtle.net/bob/rand/isaacafa.html>, as last modified

src/Makefile.am

@@ -707,8 +707,8 @@ sc_check-AUTHORS: $(all_programs)
 
 # Most functions in src/*.c should have static scope.
 # Any that don't must be marked with `extern', but `main'
-# and `usage' are exceptions.  They're always extern, but
-# don't need to be marked. Also functions starting with __
+# and `usage' are exceptions: they're always extern, but
+# do not need to be marked.  Also functions starting with __
 # are exempted due to possibly being added by the compiler
 # (when compiled as a shared library for example).
 #
@@ -717,9 +717,9 @@ sc_check-AUTHORS: $(all_programs)
 sc_tight_scope: $(bin_PROGRAMS)
 	@t=exceptions-$$$$;						\
 	trap 's=$$?; rm -f $$t; exit $$s' 0;				\
-	am__exit='(exit $s); exit $s';					\
-	trap "s=129; $$am__exit" 1; trap "s=130; $$am__exit" 2;		\
-	trap "s=141; $$am__exit" 13; trap "s=143; $$am__exit" 15;	\
+	for sig in 1 2 3 13 15; do					\
+	  eval "trap 'v=`expr $$sig + 128`; (exit $$v); exit $$v' $$sig"; \
+	done;								\
 	src=`for f in $(SOURCES); do					\
 	       test -f $$f && d= || d=$(srcdir)/; echo $$d$$f; done`;	\
 	hdr=`for f in $(noinst_HEADERS); do				\
@@ -727,7 +727,8 @@ sc_tight_scope: $(bin_PROGRAMS)
 	( printf 'main\nusage\n_.*\n';					\
 	  grep -h -A1 '^extern .*[^;]$$' $$src				\
 	    | grep -vE '^(extern |--)' | sed 's/ .*//';			\
-	  perl -ne '/^extern (?:enum )?\S+ (\S*) \(/ and print "$$1\n"' $$hdr; \
+	  perl -lne '/^extern (?:enum )?\S+ \**(\S*) \(/ and print $$1' \
+		$$hdr;							\
 	) | $(ASSORT) -u | sed 's/^/^/;s/$$/$$/' > $$t;			\
 	nm -e *.$(OBJEXT)						\
 	    | sed -n 's/.* T //p'					\
@@ -736,7 +737,7 @@ sc_tight_scope: $(bin_PROGRAMS)
 	  { echo 'the above functions should have static scope' 1>&2;	\
 	    exit 1; } || : ;						\
 	( printf '^program_name$$\n';					\
-	  perl -ne '/^extern .*?\**(\w+);/ and print "^$$1\$$\n"'	\
+	  perl -lne '/^extern .*?\**(\w+);/ and print "^$$1\$$"'	\
 	    $$hdr *.h ) | $(ASSORT) -u > $$t;				\
 	nm -e *.$(OBJEXT)						\
 	    | sed -n 's/.* [BD] //p'					\

src/copy.c

@@ -39,6 +39,7 @@
 #include "extent-scan.h"
 #include "error.h"
 #include "fcntl--.h"
+#include "fiemap.h"
 #include "file-set.h"
 #include "filemode.h"
 #include "filenamecat.h"
@@ -233,21 +234,6 @@ sparse_copy (int src_fd, int dest_fd, char *buf, size_t buf_size,
   return true;
 }
 
-/* If the file ends with a `hole' (i.e., if sparse_copy set wrote_hole_at_eof),
-   call this function to record the length of the output file.  */
-static bool
-sparse_copy_finalize (int dest_fd, char const *dst_name)
-{
-  off_t len = lseek (dest_fd, 0, SEEK_CUR);
-  if (0 <= len && ftruncate (dest_fd, len) < 0)
-    {
-      error (0, errno, _("truncating %s"), quote (dst_name));
-      return false;
-    }
-
-  return true;
-}
-
 /* Perform the O(1) btrfs clone operation, if possible.
    Upon success, return 0.  Otherwise, return -1 and set errno.  */
 static inline int
@@ -309,7 +295,7 @@ write_zeros (int fd, uint64_t n_bytes)
    return false.  */
 static bool
 extent_copy (int src_fd, int dest_fd, char *buf, size_t buf_size,
-             off_t src_total_size, bool make_holes,
+             off_t src_total_size, enum Sparse_type sparse_mode,
              char const *src_name, char const *dst_name,
              bool *require_normal_copy)
 {
@@ -345,56 +331,105 @@ extent_copy (int src_fd, int dest_fd, char *buf, size_t buf_size,
         }
 
       unsigned int i;
-      for (i = 0; i < scan.ei_count; i++)
+      bool empty_extent = false;
+      for (i = 0; i < scan.ei_count || empty_extent; i++)
         {
-          off_t ext_start = scan.ext_info[i].ext_logical;
-          uint64_t ext_len = scan.ext_info[i].ext_length;
+          off_t ext_start;
+          uint64_t ext_len;
+          uint64_t hole_size;
 
-          if (lseek (src_fd, ext_start, SEEK_SET) < 0)
+          if (i < scan.ei_count)
             {
-              error (0, errno, _("cannot lseek %s"), quote (src_name));
-            fail:
-              extent_scan_free (&scan);
-              return false;
+              ext_start = scan.ext_info[i].ext_logical;
+              ext_len = scan.ext_info[i].ext_length;
+            }
+          else /* empty extent at EOF.  */
+            {
+              i--;
+              ext_start = last_ext_start + scan.ext_info[i].ext_length;
+              ext_len = 0;
             }
 
-          if (make_holes)
+          hole_size = ext_start - last_ext_start - last_ext_len;
+
+          wrote_hole_at_eof = false;
+
+          if (hole_size)
             {
-              if (lseek (dest_fd, ext_start, SEEK_SET) < 0)
+              if (lseek (src_fd, ext_start, SEEK_SET) < 0)
                 {
-                  error (0, errno, _("cannot lseek %s"), quote (dst_name));
-                  goto fail;
+                  error (0, errno, _("cannot lseek %s"), quote (src_name));
+                fail:
+                  extent_scan_free (&scan);
+                  return false;
                 }
-            }
-          else
-            {
-              /* When not inducing holes and when there is a hole between
-                 the end of the previous extent and the beginning of the
-                 current one, write zeros to the destination file.  */
-              if (last_ext_start + last_ext_len < ext_start)
+
+              if ((empty_extent && sparse_mode == SPARSE_ALWAYS)
+                  || (!empty_extent && sparse_mode != SPARSE_NEVER))
                 {
-                  uint64_t hole_size = (ext_start
-                                        - last_ext_start
-                                        - last_ext_len);
-                  if (! write_zeros (dest_fd, hole_size))
+                  if (lseek (dest_fd, ext_start, SEEK_SET) < 0)
+                    {
+                      error (0, errno, _("cannot lseek %s"), quote (dst_name));
+                      goto fail;
+                    }
+                  wrote_hole_at_eof = true;
+                }
+              else
+                {
+                  /* When not inducing holes and when there is a hole between
+                     the end of the previous extent and the beginning of the
+                     current one, write zeros to the destination file.  */
+                  off_t nzeros = hole_size;
+                  if (empty_extent)
+                    nzeros = MIN (src_total_size - dest_pos, hole_size);
+
+                  if (! write_zeros (dest_fd, nzeros))
                     {
                       error (0, errno, _("%s: write failed"), quote (dst_name));
                       goto fail;
                     }
+
+                  dest_pos = MIN (src_total_size, ext_start);
                 }
             }
 
           last_ext_start = ext_start;
-          last_ext_len = ext_len;
 
-          off_t n_read;
-          if ( ! sparse_copy (src_fd, dest_fd, buf, buf_size,
-                              make_holes, src_name, dst_name,
-                              ext_len, &n_read,
-                              &wrote_hole_at_eof))
-            return false;
+          /* Treat an unwritten but allocated extent much like a hole.
+             I.E. don't read, but don't convert to a hole in the destination,
+             unless SPARSE_ALWAYS.  */
+          if (scan.ext_info[i].ext_flags & FIEMAP_EXTENT_UNWRITTEN)
+            {
+              empty_extent = true;
+              last_ext_len = 0;
+              if (ext_len == 0) /* The last extent is empty and processed.  */
+                empty_extent = false;
+            }
+          else
+            {
+              off_t n_read;
+              empty_extent = false;
+              last_ext_len = ext_len;
 
-          dest_pos = ext_start + n_read;
+              if ( ! sparse_copy (src_fd, dest_fd, buf, buf_size,
+                                  sparse_mode == SPARSE_ALWAYS,
+                                  src_name, dst_name, ext_len, &n_read,
+                                  &wrote_hole_at_eof))
+                goto fail;
+
+              dest_pos = ext_start + n_read;
+            }
+
+          /* If the file ends with unwritten extents not accounted for in the
+             size, then skip processing them, and the associated redundant
+             read() calls which will always return 0.  We will need to
+             remove this when we add fallocate() so that we can maintain
+             extents beyond the apparent size.  */
+          if (dest_pos == src_total_size)
+            {
+              scan.hit_final_extent = true;
+              break;
+            }
         }
 
       /* Release the space allocated to scan->ext_info.  */
@@ -412,7 +447,7 @@ extent_copy (int src_fd, int dest_fd, char *buf, size_t buf_size,
      just converted them to a hole in the destination, we must call ftruncate
      here in order to record the proper length in the destination.  */
   if ((dest_pos < src_total_size || wrote_hole_at_eof)
-      && (make_holes
+      && (sparse_mode != SPARSE_NEVER
           ? ftruncate (dest_fd, src_total_size)
           : ! write_zeros (dest_fd, src_total_size - dest_pos)))
     {
@@ -531,8 +566,11 @@ copy_attr (char const *src_path ATTRIBUTE_UNUSED,
    DST_NAME_IN is a directory that was created previously in the
    recursion.   SRC_SB and ANCESTORS describe SRC_NAME_IN.
    Set *COPY_INTO_SELF if SRC_NAME_IN is a parent of
-   FIRST_DIR_CREATED_PER_COMMAND_LINE_ARG  FIXME
    (or the same as) DST_NAME_IN; otherwise, clear it.
+   Propagate *FIRST_DIR_CREATED_PER_COMMAND_LINE_ARG from
+   caller to each invocation of copy_internal.  Be careful to
+   pass the address of a temporary, and to update
+   *FIRST_DIR_CREATED_PER_COMMAND_LINE_ARG only upon completion.
    Return true if successful.  */
 
 static bool
@@ -561,16 +599,18 @@ copy_dir (char const *src_name_in, char const *dst_name_in, bool new_dst,
   if (x->dereference == DEREF_COMMAND_LINE_ARGUMENTS)
     non_command_line_options.dereference = DEREF_NEVER;
 
+  bool new_first_dir_created = false;
   namep = name_space;
   while (*namep != '\0')
     {
       bool local_copy_into_self;
       char *src_name = file_name_concat (src_name_in, namep, NULL);
       char *dst_name = file_name_concat (dst_name_in, namep, NULL);
+      bool first_dir_created = *first_dir_created_per_command_line_arg;
 
       ok &= copy_internal (src_name, dst_name, new_dst, src_sb->st_dev,
                            ancestors, &non_command_line_options, false,
-                           first_dir_created_per_command_line_arg,
+                           &first_dir_created,
                            &local_copy_into_self, NULL);
       *copy_into_self |= local_copy_into_self;
 
@@ -583,9 +623,12 @@ copy_dir (char const *src_name_in, char const *dst_name_in, bool new_dst,
       if (local_copy_into_self)
         break;
 
+      new_first_dir_created |= first_dir_created;
       namep += strlen (namep) + 1;
     }
   free (name_space);
+  *first_dir_created_per_command_line_arg = new_first_dir_created;
+
   return ok;
 }
 
@@ -983,7 +1026,8 @@ copy_reg (char const *src_name, char const *dst_name,
          '--sparse=never' option is specified, write all data but use
          any extents to read more efficiently.  */
       if (extent_copy (source_desc, dest_desc, buf, buf_size,
-                       src_open_sb.st_size, make_holes,
+                       src_open_sb.st_size,
+                       S_ISREG (sb.st_mode) ? x->sparse_mode : SPARSE_NEVER,
                        src_name, dst_name, &normal_copy_required))
         goto preserve_metadata;
 
@@ -1000,8 +1044,9 @@ copy_reg (char const *src_name, char const *dst_name,
                           UINTMAX_MAX, &n_read,
                           &wrote_hole_at_eof)
            || (wrote_hole_at_eof &&
-               ! sparse_copy_finalize (dest_desc, dst_name)))
+               ftruncate (dest_desc, n_read) < 0))
         {
+          error (0, errno, _("failed to extend %s"), quote (dst_name));
           return_val = false;
           goto close_src_and_dst_desc;
         }
@@ -2184,13 +2229,24 @@ copy_internal (char const *src_name, char const *dst_name,
         }
     }
 
-  /* cp, invoked with `--link --no-dereference', should not follow the
-     link; we guarantee this with gnulib's linkat module (on systems
-     where link(2) follows the link, gnulib creates a symlink with
-     identical contents, which is good enough for our purposes).  */
+  /* POSIX 2008 states that it is implementation-defined whether
+     link() on a symlink creates a hard-link to the symlink, or only
+     to the referent (effectively dereferencing the symlink) (POSIX
+     2001 required the latter behavior, although many systems provided
+     the former).  Yet cp, invoked with `--link --no-dereference',
+     should not follow the link.  We can approximate the desired
+     behavior by skipping this hard-link creating block and instead
+     copying the symlink, via the `S_ISLNK'- copying code below.
+     LINK_FOLLOWS_SYMLINKS is tri-state; if it is -1, we don't know
+     how link() behaves, so we use the fallback case for safety.
+
+     Note gnulib's linkat module, guarantees that the symlink is not
+     dereferenced.  However its emulation currently doesn't maintain
+     timestamps or ownership so we only call it when we know the
+     emulation will not be needed.  */
   else if (x->hard_link
-           && (!S_ISLNK (src_mode)
-               || x->dereference != DEREF_NEVER))
+           && !(LINK_FOLLOWS_SYMLINKS && S_ISLNK (src_mode)
+                && x->dereference == DEREF_NEVER))
     {
        if (linkat (AT_FDCWD, src_name, AT_FDCWD, dst_name, 0))
         {
@@ -2313,7 +2369,9 @@ copy_internal (char const *src_name, char const *dst_name,
 
   /* If we've just created a hard-link due to cp's --link option,
      we're done.  */
-  if (x->hard_link && ! S_ISDIR (src_mode))
+  if (x->hard_link && ! S_ISDIR (src_mode)
+      && !(LINK_FOLLOWS_SYMLINKS && S_ISLNK (src_mode)
+           && x->dereference == DEREF_NEVER))
     return delayed_ok;
 
   if (copied_as_regular)

src/cut.c

@@ -496,6 +496,8 @@ set_fields (const char *fieldstr)
       if (rp[i].hi > max_range_endpoint)
         max_range_endpoint = rp[i].hi;
     }
+  if (max_range_endpoint < eol_range_start)
+    max_range_endpoint = eol_range_start;
 
   /* Allocate an array large enough so that it may be indexed by
      the field numbers corresponding to all finite ranges

src/dd.c

@@ -199,6 +199,9 @@ static int input_seek_errno;
 static uintmax_t input_offset;
 static bool input_offset_overflow;
 
+/* True if a partial read should be diagnosed.  */
+static bool warn_partial_read;
+
 /* Records truncated by conv=block. */
 static uintmax_t r_truncate = 0;
 
@@ -225,6 +228,9 @@ static sig_atomic_t volatile interrupt_signal;
 /* A count of the number of pending info signals that have been received.  */
 static sig_atomic_t volatile info_signal_count;
 
+/* Whether to discard cache for input or output.  */
+static bool i_nocache, o_nocache;
+
 /* Function used for read (to handle iflag=fullblock parameter).  */
 static ssize_t (*iread_fnc) (int fd, char *buf, size_t size);
 
@@ -259,6 +265,7 @@ static struct symbol_value const conversions[] =
   {"", 0}
 };
 
+#define FFS_MASK(x) ((x) ^ ((x) & ((x) - 1)))
 enum
   {
     /* Compute a value that's bitwise disjoint from the union
@@ -278,17 +285,23 @@ enum
           | O_SYNC
           | O_TEXT
           ),
-    /* Use its lowest bit.  */
-    O_FULLBLOCK = v ^ (v & (v - 1))
+
+    /* Use its lowest bits for private flags.  */
+    O_FULLBLOCK = FFS_MASK (v),
+    v2 = v ^ O_FULLBLOCK,
+
+    O_NOCACHE = FFS_MASK (v2)
   };
 
 /* Ensure that we got something.  */
 verify (O_FULLBLOCK != 0);
+verify (O_NOCACHE != 0);
 
 #define MULTIPLE_BITS_SET(i) (((i) & ((i) - 1)) != 0)
 
 /* Ensure that this is a single-bit value.  */
 verify ( ! MULTIPLE_BITS_SET (O_FULLBLOCK));
+verify ( ! MULTIPLE_BITS_SET (O_NOCACHE));
 
 /* Flags, for iflag="..." and oflag="...".  */
 static struct symbol_value const flags[] =
@@ -300,6 +313,7 @@ static struct symbol_value const flags[] =
   {"directory",	O_DIRECTORY},
   {"dsync",	O_DSYNC},
   {"noatime",	O_NOATIME},
+  {"nocache",	O_NOCACHE},   /* Discard cache.  */
   {"noctty",	O_NOCTTY},
   {"nofollow",	HAVE_WORKING_O_NOFOLLOW ? O_NOFOLLOW : 0},
   {"nolinks",	O_NOLINKS},
@@ -467,11 +481,11 @@ Usage: %s [OPERAND]...\n\
       fputs (_("\
 Copy a file, converting and formatting according to the operands.\n\
 \n\
-  bs=BYTES        read and write BYTES bytes at a time (also see ibs=,obs=)\n\
+  bs=BYTES        read and write up to BYTES bytes at a time\n\
   cbs=BYTES       convert BYTES bytes at a time\n\
   conv=CONVS      convert the file as per the comma separated symbol list\n\
   count=BLOCKS    copy only BLOCKS input blocks\n\
-  ibs=BYTES       read BYTES bytes at a time (default: 512)\n\
+  ibs=BYTES       read up to BYTES bytes at a time (default: 512)\n\
 "), stdout);
       fputs (_("\
   if=FILE         read from FILE instead of stdin\n\
@@ -499,18 +513,16 @@ Each CONV symbol may be:\n\
   block     pad newline-terminated records with spaces to cbs-size\n\
   unblock   replace trailing spaces in cbs-size records with newline\n\
   lcase     change upper case to lower case\n\
-"), stdout);
-      fputs (_("\
-  nocreat   do not create the output file\n\
-  excl      fail if the output file already exists\n\
-  notrunc   do not truncate the output file\n\
   ucase     change lower case to upper case\n\
   swab      swap every pair of input bytes\n\
-"), stdout);
-      fputs (_("\
-  noerror   continue after read errors\n\
   sync      pad every input block with NULs to ibs-size; when used\n\
             with block or unblock, pad with spaces rather than NULs\n\
+"), stdout);
+      fputs (_("\
+  excl      fail if the output file already exists\n\
+  nocreat   do not create the output file\n\
+  notrunc   do not truncate the output file\n\
+  noerror   continue after read errors\n\
   fdatasync  physically write output file data before finishing\n\
   fsync     likewise, but also write metadata\n\
 "), stdout);
@@ -536,6 +548,10 @@ Each FLAG symbol may be:\n\
         fputs (_("  nonblock  use non-blocking I/O\n"), stdout);
       if (O_NOATIME)
         fputs (_("  noatime   do not update access time\n"), stdout);
+#if HAVE_POSIX_FADVISE
+      if (O_NOCACHE)
+        fputs (_("  nocache   discard cached data\n"), stdout);
+#endif
       if (O_NOCTTY)
         fputs (_("  noctty    do not assign controlling terminal from file\n"),
                stdout);
@@ -789,6 +805,91 @@ process_signals (void)
     }
 }
 
+/* Return LEN rounded down to a multiple of PAGE_SIZE
+   while storing the remainder internally per FD.
+   Pass LEN == 0 to get the current remainder.  */
+
+static off_t
+cache_round (int fd, off_t len)
+{
+  static off_t i_pending, o_pending;
+  off_t *pending = (fd == STDIN_FILENO ? &i_pending : &o_pending);
+
+  if (len)
+    {
+      off_t c_pending = *pending + len;
+      *pending = c_pending % page_size;
+      if (c_pending > *pending)
+        len = c_pending - *pending;
+      else
+        len = 0;
+    }
+  else
+    len = *pending;
+
+  return len;
+}
+
+/* Discard the cache from the current offset of either
+   STDIN_FILENO or STDOUT_FILENO.
+   Return true on success.  */
+
+static bool
+invalidate_cache (int fd, off_t len)
+{
+  int adv_ret = -1;
+
+  /* Minimize syscalls.  */
+  off_t clen = cache_round (fd, len);
+  if (len && !clen)
+    return true; /* Don't advise this time.  */
+  if (!len && !clen && max_records)
+    return true; /* Nothing pending.  */
+  off_t pending = len ? cache_round (fd, 0) : 0;
+
+  if (fd == STDIN_FILENO)
+    {
+      if (input_seekable)
+        {
+          /* Note we're being careful here to only invalidate what
+             we've read, so as not to dump any read ahead cache.  */
+#if HAVE_POSIX_FADVISE
+            adv_ret = posix_fadvise (fd, input_offset - clen - pending, clen,
+                                     POSIX_FADV_DONTNEED);
+#else
+            errno = ENOTSUP;
+#endif
+        }
+      else
+        errno = ESPIPE;
+    }
+  else if (fd == STDOUT_FILENO)
+    {
+      static off_t output_offset = -2;
+
+      if (output_offset != -1)
+        {
+          if (0 > output_offset)
+            {
+              output_offset = lseek (fd, 0, SEEK_CUR);
+              output_offset -= clen + pending;
+            }
+          if (0 <= output_offset)
+            {
+#if HAVE_POSIX_FADVISE
+              adv_ret = posix_fadvise (fd, output_offset, clen,
+                                       POSIX_FADV_DONTNEED);
+#else
+              errno = ENOTSUP;
+#endif
+              output_offset += clen + pending;
+            }
+        }
+    }
+
+  return adv_ret != -1 ? true : false;
+}
+
 /* Read from FD into the buffer BUF of size SIZE, processing any
    signals that arrive before bytes are read.  Return the number of
    bytes read if successful, -1 (setting errno) on failure.  */
@@ -796,14 +897,35 @@ process_signals (void)
 static ssize_t
 iread (int fd, char *buf, size_t size)
 {
-  while (true)
+  ssize_t nread;
+
+  do
     {
-      ssize_t nread;
       process_signals ();
       nread = read (fd, buf, size);
-      if (! (nread < 0 && errno == EINTR))
-        return nread;
     }
+  while (nread < 0 && errno == EINTR);
+
+  if (0 < nread && warn_partial_read)
+    {
+      static ssize_t prev_nread;
+
+      if (0 < prev_nread && prev_nread < size)
+        {
+          uintmax_t prev = prev_nread;
+          error (0, 0, ngettext (("warning: partial read (%"PRIuMAX" byte); "
+                                  "suggest iflag=fullblock"),
+                                 ("warning: partial read (%"PRIuMAX" bytes); "
+                                  "suggest iflag=fullblock"),
+                                 select_plural (prev)),
+                 prev);
+          warn_partial_read = false;
+        }
+
+      prev_nread = nread;
+    }
+
+  return nread;
 }
 
 /* Wrapper around iread function to accumulate full blocks.  */
@@ -849,9 +971,7 @@ iwrite (int fd, char const *buf, size_t size)
          posix_fadvise to tell the system not to pollute the buffer
          cache with this data.  Don't bother to diagnose lseek or
          posix_fadvise failure. */
-      off_t off = lseek (STDOUT_FILENO, 0, SEEK_CUR);
-      if (0 <= off)
-        fdadvise (STDOUT_FILENO, off, 0, FADVISE_DONTNEED);
+      invalidate_cache (STDOUT_FILENO, 0);
 
       /* Attempt to ensure that that final block is committed
          to disk as quickly as possible.  */
@@ -880,6 +1000,9 @@ iwrite (int fd, char const *buf, size_t size)
         total_written += nwritten;
     }
 
+  if (o_nocache && total_written)
+    invalidate_cache (fd, total_written);
+
   return total_written;
 }
 
@@ -1070,7 +1193,7 @@ scanargs (int argc, char *const *argv)
     input_blocksize = output_blocksize = blocksize;
   else
     {
-      /* POSIX says dd aggregates short reads into
+      /* POSIX says dd aggregates partial reads into
          output_blocksize if bs= is not specified.  */
       conversions_mask |= C_TWOBUFS;
     }
@@ -1090,6 +1213,17 @@ scanargs (int argc, char *const *argv)
       error (0, 0, "%s: %s", _("invalid output flag"), "'fullblock'");
       usage (EXIT_FAILURE);
     }
+
+  /* Warn about partial reads if bs=SIZE is given and iflag=fullblock
+     is not, and if counting or skipping bytes or using direct I/O.
+     This helps to avoid confusion with miscounts, and to avoid issues
+     with direct I/O on GNU/Linux.  */
+  warn_partial_read =
+    (! (conversions_mask & C_TWOBUFS) && ! (input_flags & O_FULLBLOCK)
+     && (skip_records
+         || (0 < max_records && max_records < (uintmax_t) -1)
+         || (input_flags | output_flags) & O_DIRECT));
+
   iread_fnc = ((input_flags & O_FULLBLOCK)
                ? iread_fullblock
                : iread);
@@ -1103,6 +1237,20 @@ scanargs (int argc, char *const *argv)
     error (EXIT_FAILURE, 0, _("cannot combine lcase and ucase"));
   if (multiple_bits_set (conversions_mask & (C_EXCL | C_NOCREAT)))
     error (EXIT_FAILURE, 0, _("cannot combine excl and nocreat"));
+  if (multiple_bits_set (input_flags & (O_DIRECT | O_NOCACHE))
+      || multiple_bits_set (output_flags & (O_DIRECT | O_NOCACHE)))
+    error (EXIT_FAILURE, 0, _("cannot combine direct and nocache"));
+
+  if (input_flags & O_NOCACHE)
+    {
+      i_nocache = true;
+      input_flags &= ~O_NOCACHE;
+    }
+  if (output_flags & O_NOCACHE)
+    {
+      o_nocache = true;
+      output_flags &= ~O_NOCACHE;
+    }
 }
 
 /* Fix up translation table. */
@@ -1639,7 +1787,7 @@ dd_copy (void)
          There are 3 reasons why there might be unskipped blocks/bytes:
              1. file is too small
              2. pipe has not enough data
-             3. short reads  */
+             3. partial reads  */
       if (us_blocks || (!input_offset_overflow && us_bytes))
         {
           error (0, 0,
@@ -1685,6 +1833,9 @@ dd_copy (void)
 
       nread = iread_fnc (STDIN_FILENO, ibuf, input_blocksize);
 
+      if (nread >= 0 && i_nocache)
+        invalidate_cache (STDIN_FILENO, nread);
+
       if (nread == 0)
         break;			/* EOF.  */
 
@@ -1694,8 +1845,14 @@ dd_copy (void)
           if (conversions_mask & C_NOERROR)
             {
               print_stats ();
+              size_t bad_portion = input_blocksize - partread;
+
+              /* We already know this data is not cached,
+                 but call this so that correct offsets are maintained.  */
+              invalidate_cache (STDIN_FILENO, bad_portion);
+
               /* Seek past the bad block if possible. */
-              if (!advance_input_after_read_error (input_blocksize - partread))
+              if (!advance_input_after_read_error (bad_portion))
                 {
                   exit_status = EXIT_FAILURE;
 
@@ -1950,5 +2107,31 @@ main (int argc, char **argv)
 
   exit_status = dd_copy ();
 
+  if (max_records == 0)
+    {
+      /* Special case to invalidate cache to end of file.  */
+      if (i_nocache && !invalidate_cache (STDIN_FILENO, 0))
+        {
+          error (0, errno, _("failed to discard cache for: %s"),
+                 quote (input_file));
+          exit_status = EXIT_FAILURE;
+        }
+      if (o_nocache && !invalidate_cache (STDOUT_FILENO, 0))
+        {
+          error (0, errno, _("failed to discard cache for: %s"),
+                 quote (output_file));
+          exit_status = EXIT_FAILURE;
+        }
+    }
+  else if (max_records != (uintmax_t) -1)
+    {
+      /* Invalidate any pending region less that page size,
+         in case the kernel might round up.  */
+      if (i_nocache)
+        invalidate_cache (STDIN_FILENO, 0);
+      if (o_nocache)
+        invalidate_cache (STDOUT_FILENO, 0);
+    }
+
   quit (exit_status);
 }

src/df.c

@@ -22,11 +22,14 @@
 #include <stdio.h>
 #include <sys/types.h>
 #include <getopt.h>
+#include <assert.h>
 
 #include "system.h"
 #include "error.h"
 #include "fsusage.h"
 #include "human.h"
+#include "mbsalign.h"
+#include "mbswidth.h"
 #include "mountlist.h"
 #include "quote.h"
 #include "find-mount-point.h"
@@ -112,6 +115,52 @@ static bool print_grand_total;
 /* Grand total data. */
 static struct fs_usage grand_fsu;
 
+/* Display modes.  */
+enum { DEFAULT_MODE, INODES_MODE, HUMAN_MODE, POSIX_MODE, NMODES };
+static int header_mode = DEFAULT_MODE;
+
+/* Displayable fields.  */
+enum
+{
+  DEV_FIELD,   /* file system */
+  TYPE_FIELD,  /* FS type */
+  TOTAL_FIELD, /* blocks or inodes */
+  USED_FIELD,  /* ditto */
+  FREE_FIELD,  /* ditto */
+  PCENT_FIELD, /* percent used */
+  MNT_FIELD,   /* mount point */
+  NFIELDS
+};
+
+/* Header strings for the above fields in each mode.
+   NULL means to use the header for the default mode.  */
+static const char *headers[NFIELDS][NMODES] = {
+/*  DEFAULT_MODE	INODES_MODE	HUMAN_MODE	POSIX_MODE  */
+  { N_("Filesystem"),   NULL,           NULL,           NULL },
+  { N_("Type"),         NULL,           NULL,           NULL },
+  { N_("blocks"),       N_("Inodes"),   N_("Size"),     NULL },
+  { N_("Used"),         N_("IUsed"),    NULL,           NULL },
+  { N_("Available"),    N_("IFree"),    N_("Avail"),    NULL },
+  { N_("Use%"),         N_("IUse%"),    NULL,           N_("Capacity") },
+  { N_("Mounted on"),   NULL,           NULL,           NULL }
+};
+
+/* Alignments for the 3 textual and 4 numeric fields.  */
+static mbs_align_t alignments[NFIELDS] = {
+  MBS_ALIGN_LEFT, MBS_ALIGN_LEFT,
+  MBS_ALIGN_RIGHT, MBS_ALIGN_RIGHT, MBS_ALIGN_RIGHT, MBS_ALIGN_RIGHT,
+  MBS_ALIGN_LEFT
+};
+
+/* Auto adjusted (up) widths used to align columns.  */
+static size_t widths[NFIELDS] = { 14, 4, 5, 5, 5, 4, 0 };
+
+/* Storage for pointers for each string (cell of table).  */
+static char ***table;
+
+/* The current number of processed rows (including header).  */
+static size_t nrows;
+
 /* For long options that have no equivalent short option, use a
    non-character as a pseudo short option, starting with CHAR_MAX + 1.  */
 enum
@@ -141,70 +190,133 @@ static struct option const long_options[] =
   {NULL, 0, NULL, 0}
 };
 
+/* Dynamically allocate a row of pointers in TABLE, which
+   can then be accessed with standard 2D array notation.  */
+
 static void
-print_header (void)
+alloc_table_row (void)
 {
-  char buf[MAX (LONGEST_HUMAN_READABLE + 1, INT_BUFSIZE_BOUND (uintmax_t))];
+  nrows++;
+  table = xnrealloc (table, nrows, sizeof (char *));
+  table[nrows-1] = xnmalloc (NFIELDS, sizeof (char *));
+}
 
-  if (print_type)
-    /* TRANSLATORS:
-       For best results (df header/column alignment), ensure that
-       your translation has the same length as the original.  */
-    fputs (_("Filesystem    Type"), stdout);
-  else
-    fputs (_("Filesystem        "), stdout);
+/* Output each cell in the table, accounting for the
+   alignment and max width of each column.  */
 
-  if (inode_format)
-    /* TRANSLATORS:
-       For best results (df header/column alignment), ensure that
-       your translation has the same length as the original.
-       Also, each column name translation should end at the same
-       column as the corresponding original.  */
-    fputs (_("    Inodes   IUsed   IFree IUse%"), stdout);
-  else if (human_output_opts & human_autoscale)
+static void
+print_table (void)
+{
+  size_t field, row;
+
+  for (row = 0; row < nrows; row ++)
     {
-      if (human_output_opts & human_base_1024)
-        fputs (_("    Size  Used Avail Use%"), stdout);
-      else
-        fputs (_("     Size   Used  Avail Use%"), stdout);
-    }
-  else if (posix_format)
-    printf (_(" %s-blocks      Used Available Capacity"),
-            umaxtostr (output_block_size, buf));
-  else
-    {
-      int opts = (human_suppress_point_zero
-                  | human_autoscale | human_SI
-                  | (human_output_opts
-                     & (human_group_digits | human_base_1024 | human_B)));
-
-      /* Prefer the base that makes the human-readable value more exact,
-         if there is a difference.  */
-
-      uintmax_t q1000 = output_block_size;
-      uintmax_t q1024 = output_block_size;
-      bool divisible_by_1000;
-      bool divisible_by_1024;
-
-      do
+      for (field = 0; field < NFIELDS; field++)
         {
-          divisible_by_1000 = q1000 % 1000 == 0;  q1000 /= 1000;
-          divisible_by_1024 = q1024 % 1024 == 0;  q1024 /= 1024;
+          size_t width = widths[field];
+          char *cell = table[row][field];
+          if (!cell)
+            continue;
+
+          /* Note the DEV_FIELD used to be displayed on it's own line
+             if (!posix_format && mbswidth (cell) > 20), but that
+             functionality is probably more problematic than helpful.  */
+          if (field != 0)
+            putchar (' ');
+          if (field == MNT_FIELD) /* The last one.  */
+            fputs (cell, stdout);
+          else
+            {
+              cell = ambsalign (table[row][field], &width,
+                                alignments[field], MBA_UNIBYTE_FALLBACK);
+              fputs (cell, stdout);
+              free (cell);
+            }
+          IF_LINT (free (table[row][field]));
         }
-      while (divisible_by_1000 & divisible_by_1024);
-
-      if (divisible_by_1000 < divisible_by_1024)
-        opts |= human_base_1024;
-      if (divisible_by_1024 < divisible_by_1000)
-        opts &= ~human_base_1024;
-      if (! (opts & human_base_1024))
-        opts |= human_B;
-
-      printf (_(" %4s-blocks      Used Available Use%%"),
-              human_readable (output_block_size, buf, opts, 1, 1));
+      putchar ('\n');
+      IF_LINT (free (table[row]));
     }
 
-  fputs (_(" Mounted on\n"), stdout);
+  IF_LINT (free (table));
+}
+
+/* Optain the appropriate header entries.  */
+
+static void
+get_header (void)
+{
+  size_t field;
+
+  alloc_table_row ();
+
+  for (field = 0; field < NFIELDS; field++)
+    {
+      if (field == TYPE_FIELD && !print_type)
+        {
+          table[nrows-1][field] = NULL;
+          continue;
+        }
+
+      char *cell = NULL;
+      char const *header = _(headers[field][header_mode]);
+      if (!header)
+        header = _(headers[field][DEFAULT_MODE]);
+
+      if (header_mode == DEFAULT_MODE && field == TOTAL_FIELD)
+        {
+          char buf[LONGEST_HUMAN_READABLE + 1];
+
+          int opts = (human_suppress_point_zero
+                      | human_autoscale | human_SI
+                      | (human_output_opts
+                         & (human_group_digits | human_base_1024 | human_B)));
+
+          /* Prefer the base that makes the human-readable value more exact,
+             if there is a difference.  */
+
+          uintmax_t q1000 = output_block_size;
+          uintmax_t q1024 = output_block_size;
+          bool divisible_by_1000;
+          bool divisible_by_1024;
+
+          do
+            {
+              divisible_by_1000 = q1000 % 1000 == 0;  q1000 /= 1000;
+              divisible_by_1024 = q1024 % 1024 == 0;  q1024 /= 1024;
+            }
+          while (divisible_by_1000 & divisible_by_1024);
+
+          if (divisible_by_1000 < divisible_by_1024)
+            opts |= human_base_1024;
+          if (divisible_by_1024 < divisible_by_1000)
+            opts &= ~human_base_1024;
+          if (! (opts & human_base_1024))
+            opts |= human_B;
+
+          char *num = human_readable (output_block_size, buf, opts, 1, 1);
+
+          if (asprintf (&cell, "%s-%s", num, header) == -1)
+            cell = NULL;
+        }
+      else if (header_mode == POSIX_MODE && field == TOTAL_FIELD)
+        {
+          char buf[INT_BUFSIZE_BOUND (uintmax_t)];
+          char *num = umaxtostr (output_block_size, buf);
+
+          if (asprintf (&cell, "%s-%s", num, header) == -1)
+            cell = NULL;
+        }
+      else
+        cell = strdup (header);
+
+      if (!cell)
+        xalloc_die ();
+
+      table[nrows-1][field] = cell;
+
+      widths[field] = MAX (widths[field], mbswidth (cell, 0));
+    }
 }
 
 /* Is FSTYPE a type of file system that should be listed?  */
@@ -305,7 +417,7 @@ add_uint_with_neg_flag (uintmax_t *dest, bool *dest_neg,
     *dest = -*dest;
 }
 
-/* Display a space listing for the disk device with absolute file name DISK.
+/* Optain a space listing for the disk device with absolute file name DISK.
    If MOUNT_POINT is non-NULL, it is the name of the root of the
    file system on DISK.
    If STAT_FILE is non-null, it is the name of a file within the file
@@ -319,16 +431,13 @@ add_uint_with_neg_flag (uintmax_t *dest, bool *dest_neg,
    ME_DUMMY and ME_REMOTE are the mount entry flags.  */
 
 static void
-show_dev (char const *disk, char const *mount_point,
-          char const *stat_file, char const *fstype,
-          bool me_dummy, bool me_remote,
-          const struct fs_usage *force_fsu)
+get_dev (char const *disk, char const *mount_point,
+         char const *stat_file, char const *fstype,
+         bool me_dummy, bool me_remote,
+         const struct fs_usage *force_fsu)
 {
   struct fs_usage fsu;
-  char buf[3][LONGEST_HUMAN_READABLE + 2];
-  int width;
-  int col1_adjustment = 0;
-  int use_width;
+  char buf[LONGEST_HUMAN_READABLE + 2];
   uintmax_t input_units;
   uintmax_t output_units;
   uintmax_t total;
@@ -338,6 +447,8 @@ show_dev (char const *disk, char const *mount_point,
   uintmax_t used;
   bool negate_used;
   double pct = -1;
+  char* cell;
+  size_t field;
 
   if (me_remote && show_local_fs)
     return;
@@ -370,39 +481,18 @@ show_dev (char const *disk, char const *mount_point,
   if (! file_systems_processed)
     {
       file_systems_processed = true;
-      print_header ();
+      get_header ();
     }
 
+  alloc_table_row ();
+
   if (! disk)
     disk = "-";			/* unknown */
   if (! fstype)
     fstype = "-";		/* unknown */
 
-  /* df.c reserved 5 positions for fstype,
-     but that does not suffice for type iso9660 */
-  if (print_type)
-    {
-      size_t disk_name_len = strlen (disk);
-      size_t fstype_len = strlen (fstype);
-      if (disk_name_len + fstype_len < 18)
-        printf ("%s%*s  ", disk, 18 - (int) disk_name_len, fstype);
-      else if (!posix_format)
-        printf ("%s\n%18s  ", disk, fstype);
-      else
-        printf ("%s %s", disk, fstype);
-    }
-  else
-    {
-      if (strlen (disk) > 20 && !posix_format)
-        printf ("%s\n%20s", disk, "");
-      else
-        printf ("%-20s", disk);
-    }
-
   if (inode_format)
     {
-      width = 7;
-      use_width = 5;
       input_units = output_units = 1;
       total = fsu.fsu_files;
       available = fsu.fsu_ffree;
@@ -416,22 +506,6 @@ show_dev (char const *disk, char const *mount_point,
     }
   else
     {
-      if (human_output_opts & human_autoscale)
-        width = 5 + ! (human_output_opts & human_base_1024);
-      else
-        {
-          width = 9;
-          if (posix_format)
-            {
-              uintmax_t b;
-              col1_adjustment = -3;
-              for (b = output_block_size; 9 < b; b /= 10)
-                col1_adjustment++;
-            }
-        }
-      use_width = ((posix_format
-                    && ! (human_output_opts & human_autoscale))
-                   ? 8 : 4);
       input_units = fsu.fsu_blocksize;
       output_units = output_block_size;
       total = fsu.fsu_blocks;
@@ -458,73 +532,110 @@ show_dev (char const *disk, char const *mount_point,
       negate_used = (total < available_to_root);
     }
 
-  printf (" %*s %*s %*s ",
-          width + col1_adjustment,
-          df_readable (false, total,
-                       buf[0], input_units, output_units),
-          width, df_readable (negate_used, used,
-                              buf[1], input_units, output_units),
-          width, df_readable (negate_available, available,
-                              buf[2], input_units, output_units));
-
-  if (! known_value (used) || ! known_value (available))
-    ;
-  else if (!negate_used
-           && used <= TYPE_MAXIMUM (uintmax_t) / 100
-           && used + available != 0
-           && (used + available < used) == negate_available)
+  for (field = 0; field < NFIELDS; field++)
     {
-      uintmax_t u100 = used * 100;
-      uintmax_t nonroot_total = used + available;
-      pct = u100 / nonroot_total + (u100 % nonroot_total != 0);
-    }
-  else
-    {
-      /* The calculation cannot be done easily with integer
-         arithmetic.  Fall back on floating point.  This can suffer
-         from minor rounding errors, but doing it exactly requires
-         multiple precision arithmetic, and it's not worth the
-         aggravation.  */
-      double u = negate_used ? - (double) - used : used;
-      double a = negate_available ? - (double) - available : available;
-      double nonroot_total = u + a;
-      if (nonroot_total)
+      switch (field)
         {
-          long int lipct = pct = u * 100 / nonroot_total;
-          double ipct = lipct;
+        case DEV_FIELD:
+          cell = xstrdup (disk);
+          break;
 
-          /* Like `pct = ceil (dpct);', but avoid ceil so that
-             the math library needn't be linked.  */
-          if (ipct - 1 < pct && pct <= ipct + 1)
-            pct = ipct + (ipct < pct);
-        }
-    }
+        case TYPE_FIELD:
+          cell = print_type ? xstrdup (fstype) : NULL;
+          break;
 
-  if (0 <= pct)
-    printf ("%*.0f%%", use_width - 1, pct);
-  else
-    printf ("%*s", use_width, "- ");
+        case TOTAL_FIELD:
+          cell = xstrdup (df_readable (false, total, buf,
+                                       input_units, output_units));
+          break;
+        case USED_FIELD:
+          cell = xstrdup (df_readable (negate_used, used, buf,
+                                       input_units, output_units));
+          break;
+        case FREE_FIELD:
+          cell = xstrdup (df_readable (negate_available, available, buf,
+                                       input_units, output_units));
+          break;
 
-  if (mount_point)
-    {
+        case PCENT_FIELD:
+          if (! known_value (used) || ! known_value (available))
+            ;
+          else if (!negate_used
+                   && used <= TYPE_MAXIMUM (uintmax_t) / 100
+                   && used + available != 0
+                   && (used + available < used) == negate_available)
+            {
+              uintmax_t u100 = used * 100;
+              uintmax_t nonroot_total = used + available;
+              pct = u100 / nonroot_total + (u100 % nonroot_total != 0);
+            }
+          else
+            {
+              /* The calculation cannot be done easily with integer
+                 arithmetic.  Fall back on floating point.  This can suffer
+                 from minor rounding errors, but doing it exactly requires
+                 multiple precision arithmetic, and it's not worth the
+                 aggravation.  */
+              double u = negate_used ? - (double) - used : used;
+              double a = negate_available ? - (double) - available : available;
+              double nonroot_total = u + a;
+              if (nonroot_total)
+                {
+                  long int lipct = pct = u * 100 / nonroot_total;
+                  double ipct = lipct;
+
+                  /* Like `pct = ceil (dpct);', but avoid ceil so that
+                     the math library needn't be linked.  */
+                  if (ipct - 1 < pct && pct <= ipct + 1)
+                    pct = ipct + (ipct < pct);
+                }
+            }
+
+          if (0 <= pct)
+            {
+              if (asprintf (&cell, "%.0f%%", pct) == -1)
+                cell = NULL;
+            }
+          else
+            cell = strdup ("-");
+
+          if (!cell)
+            xalloc_die ();
+
+          break;
+
+        case MNT_FIELD:
+          if (mount_point)
+            {
 #ifdef HIDE_AUTOMOUNT_PREFIX
-      /* Don't print the first directory name in MOUNT_POINT if it's an
-         artifact of an automounter.  This is a bit too aggressive to be
-         the default.  */
-      if (strncmp ("/auto/", mount_point, 6) == 0)
-        mount_point += 5;
-      else if (strncmp ("/tmp_mnt/", mount_point, 9) == 0)
-        mount_point += 8;
+              /* Don't print the first directory name in MOUNT_POINT if it's an
+                 artifact of an automounter.  This is a bit too aggressive to be
+                 the default.  */
+              if (STRNCMP_LIT (mount_point, "/auto/") == 0)
+                mount_point += 5;
+              else if (STRNCMP_LIT (mount_point, "/tmp_mnt/") == 0)
+                mount_point += 8;
 #endif
-      printf (" %s", mount_point);
+              cell = xstrdup (mount_point);
+            }
+          else
+            cell = NULL;
+          break;
+
+        default:
+          assert (!"unhandled field");
+        }
+
+      if (cell)
+        widths[field] = MAX (widths[field], mbswidth (cell, 0));
+      table[nrows-1][field] = cell;
     }
-  putchar ('\n');
 }
 
 /* If DISK corresponds to a mount point, show its usage
    and return true.  Otherwise, return false.  */
 static bool
-show_disk (char const *disk)
+get_disk (char const *disk)
 {
   struct mount_entry const *me;
   struct mount_entry const *best_match = NULL;
@@ -535,9 +646,9 @@ show_disk (char const *disk)
 
   if (best_match)
     {
-      show_dev (best_match->me_devname, best_match->me_mountdir, NULL,
-                best_match->me_type, best_match->me_dummy,
-                best_match->me_remote, NULL);
+      get_dev (best_match->me_devname, best_match->me_mountdir, NULL,
+               best_match->me_type, best_match->me_dummy,
+               best_match->me_remote, NULL);
       return true;
     }
 
@@ -548,7 +659,7 @@ show_disk (char const *disk)
    and show its disk usage.
    STATP must be the result of `stat (POINT, STATP)'.  */
 static void
-show_point (const char *point, const struct stat *statp)
+get_point (const char *point, const struct stat *statp)
 {
   struct stat disk_stats;
   struct mount_entry *me;
@@ -571,7 +682,7 @@ show_point (const char *point, const struct stat *statp)
           if (best_match_len <= len && len <= resolved_len
               && (len == 1 /* root file system */
                   || ((len == resolved_len || resolved[len] == '/')
-                      && strncmp (me->me_mountdir, resolved, len) == 0)))
+                      && STREQ_LEN (me->me_mountdir, resolved, len))))
             {
               best_match = me;
               best_match_len = len;
@@ -621,9 +732,9 @@ show_point (const char *point, const struct stat *statp)
       }
 
   if (best_match)
-    show_dev (best_match->me_devname, best_match->me_mountdir, point,
-              best_match->me_type, best_match->me_dummy, best_match->me_remote,
-              NULL);
+    get_dev (best_match->me_devname, best_match->me_mountdir, point,
+             best_match->me_type, best_match->me_dummy, best_match->me_remote,
+             NULL);
   else
     {
       /* We couldn't find the mount entry corresponding to POINT.  Go ahead and
@@ -634,7 +745,7 @@ show_point (const char *point, const struct stat *statp)
       char *mp = find_mount_point (point, statp);
       if (mp)
         {
-          show_dev (NULL, mp, NULL, NULL, false, false, NULL);
+          get_dev (NULL, mp, NULL, NULL, false, false, NULL);
           free (mp);
         }
     }
@@ -644,25 +755,25 @@ show_point (const char *point, const struct stat *statp)
    for it.  STATP is the results of `stat' on NAME.  */
 
 static void
-show_entry (char const *name, struct stat const *statp)
+get_entry (char const *name, struct stat const *statp)
 {
   if ((S_ISBLK (statp->st_mode) || S_ISCHR (statp->st_mode))
-      && show_disk (name))
+      && get_disk (name))
     return;
 
-  show_point (name, statp);
+  get_point (name, statp);
 }
 
 /* Show all mounted file systems, except perhaps those that are of
    an unselected type or are empty. */
 
 static void
-show_all_entries (void)
+get_all_entries (void)
 {
   struct mount_entry *me;
 
   for (me = mount_list; me; me = me->me_next)
-    show_dev (me->me_devname, me->me_mountdir, NULL, me->me_type,
+    get_dev (me->me_devname, me->me_mountdir, NULL, me->me_type,
               me->me_dummy, me->me_remote, NULL);
 }
 
@@ -862,6 +973,13 @@ main (int argc, char **argv)
                        &human_output_opts, &output_block_size);
     }
 
+  if (inode_format)
+    header_mode = INODES_MODE;
+  else if (human_output_opts & human_autoscale)
+    header_mode = HUMAN_MODE;
+  else if (posix_format)
+    header_mode = POSIX_MODE;
+
   /* Fail if the same file system type was both selected and excluded.  */
   {
     bool match = false;
@@ -939,18 +1057,20 @@ main (int argc, char **argv)
 
       for (i = optind; i < argc; ++i)
         if (argv[i])
-          show_entry (argv[i], &stats[i - optind]);
+          get_entry (argv[i], &stats[i - optind]);
     }
   else
-    show_all_entries ();
+    get_all_entries ();
 
   if (print_grand_total)
     {
       if (inode_format)
         grand_fsu.fsu_blocks = 1;
-      show_dev ("total", NULL, NULL, NULL, false, false, &grand_fsu);
+      get_dev ("total", NULL, NULL, NULL, false, false, &grand_fsu);
     }
 
+  print_table ();
+
   if (! file_systems_processed)
     error (EXIT_FAILURE, 0, _("no file systems processed"));
 

src/du.c

@@ -926,19 +926,19 @@ main (int argc, char **argv)
       bool skip_file = false;
       enum argv_iter_err ai_err;
       char *file_name = argv_iter (ai, &ai_err);
-      if (ai_err == AI_ERR_EOF)
-        break;
       if (!file_name)
         {
           switch (ai_err)
             {
+            case AI_ERR_EOF:
+              goto argv_iter_done;
             case AI_ERR_READ:
-              error (0, errno, _("%s: read error"), quote (files_from));
-              continue;
-
+              error (0, errno, _("%s: read error"),
+                     quotearg_colon (files_from));
+              ok = false;
+              goto argv_iter_done;
             case AI_ERR_MEM:
               xalloc_die ();
-
             default:
               assert (!"unexpected error code from argv_iter");
             }
@@ -985,11 +985,12 @@ main (int argc, char **argv)
           ok &= du_files (temp_argv, bit_flags);
         }
     }
+ argv_iter_done:
 
   argv_iter_free (ai);
   di_set_free (di_set);
 
-  if (files_from && (ferror (stdin) || fclose (stdin) != 0))
+  if (files_from && (ferror (stdin) || fclose (stdin) != 0) && ok)
     error (EXIT_FAILURE, 0, _("error reading %s"), quote (files_from));
 
   if (print_grand_total)

src/extent-scan.c

@@ -17,18 +17,48 @@
    Written by Jie Liu (jeff.liu@oracle.com).  */
 
 #include <config.h>
-#include <stdio.h>
 #include <sys/types.h>
 #include <sys/ioctl.h>
+#include <sys/utsname.h>
 #include <assert.h>
 
 #include "system.h"
 #include "extent-scan.h"
+#include "xstrtol.h"
 
 #ifndef HAVE_FIEMAP
 # include "fiemap.h"
 #endif
 
+/* Work around Linux kernel issues on BTRFS and EXT4 before 2.6.39.
+   FIXME: remove in 2013, or whenever we're pretty confident
+   that the offending, unpatched kernels are no longer in use.  */
+static bool
+extent_need_sync (void)
+{
+  static int need_sync = -1;
+
+  if (need_sync == -1)
+    {
+      struct utsname name;
+      need_sync = 0; /* No workaround by default.  */
+
+#ifdef __linux__
+      if (uname (&name) != -1 && STRNCMP_LIT (name.release, "2.6.") == 0)
+        {
+           unsigned long val;
+           if (xstrtoul (name.release + 4, NULL, 10, &val, NULL) == LONGINT_OK)
+             {
+               if (val < 39)
+                 need_sync = 1;
+             }
+        }
+#endif
+    }
+
+  return need_sync;
+}
+
 /* Allocate space for struct extent_scan, initialize the entries if
    necessary and return it as the input argument of extent_scan_read().  */
 extern void
@@ -36,9 +66,11 @@ extent_scan_init (int src_fd, struct extent_scan *scan)
 {
   scan->fd = src_fd;
   scan->ei_count = 0;
+  scan->ext_info = NULL;
   scan->scan_start = 0;
   scan->initial_scan_failed = false;
   scan->hit_final_extent = false;
+  scan->fm_flags = extent_need_sync () ? FIEMAP_FLAG_SYNC : 0;
 }
 
 #ifdef __linux__
@@ -50,60 +82,126 @@ extent_scan_init (int src_fd, struct extent_scan *scan)
 extern bool
 extent_scan_read (struct extent_scan *scan)
 {
-  union { struct fiemap f; char c[4096]; } fiemap_buf;
-  struct fiemap *fiemap = &fiemap_buf.f;
-  struct fiemap_extent *fm_extents = &fiemap->fm_extents[0];
-  enum { count = (sizeof fiemap_buf - sizeof *fiemap) / sizeof *fm_extents };
-  verify (count != 0);
+  unsigned int si = 0;
+  struct extent_info *last_ei IF_LINT ( = scan->ext_info);
 
-  /* This is required at least to initialize fiemap->fm_start,
-     but also serves (in mid 2010) to appease valgrind, which
-     appears not to know the semantics of the FIEMAP ioctl. */
-  memset (&fiemap_buf, 0, sizeof fiemap_buf);
-
-  fiemap->fm_start = scan->scan_start;
-  fiemap->fm_flags = FIEMAP_FLAG_SYNC;
-  fiemap->fm_extent_count = count;
-  fiemap->fm_length = FIEMAP_MAX_OFFSET - scan->scan_start;
-
-  /* Fall back to the standard copy if call ioctl(2) failed for the
-     the first time.  */
-  if (ioctl (scan->fd, FS_IOC_FIEMAP, fiemap) < 0)
+  while (true)
     {
-      if (scan->scan_start == 0)
-        scan->initial_scan_failed = true;
-      return false;
+      union { struct fiemap f; char c[4096]; } fiemap_buf;
+      struct fiemap *fiemap = &fiemap_buf.f;
+      struct fiemap_extent *fm_extents = &fiemap->fm_extents[0];
+      enum { count = (sizeof fiemap_buf - sizeof *fiemap)/sizeof *fm_extents };
+      verify (count > 1);
+
+      /* This is required at least to initialize fiemap->fm_start,
+         but also serves (in mid 2010) to appease valgrind, which
+         appears not to know the semantics of the FIEMAP ioctl. */
+      memset (&fiemap_buf, 0, sizeof fiemap_buf);
+
+      fiemap->fm_start = scan->scan_start;
+      fiemap->fm_flags = scan->fm_flags;
+      fiemap->fm_extent_count = count;
+      fiemap->fm_length = FIEMAP_MAX_OFFSET - scan->scan_start;
+
+      /* Fall back to the standard copy if call ioctl(2) failed for
+         the first time.  */
+      if (ioctl (scan->fd, FS_IOC_FIEMAP, fiemap) < 0)
+        {
+          if (scan->scan_start == 0)
+            scan->initial_scan_failed = true;
+          return false;
+        }
+
+      /* If 0 extents are returned, then no more scans are needed.  */
+      if (fiemap->fm_mapped_extents == 0)
+        {
+          scan->hit_final_extent = true;
+          return scan->scan_start != 0;
+        }
+
+      assert (scan->ei_count <= SIZE_MAX - fiemap->fm_mapped_extents);
+      scan->ei_count += fiemap->fm_mapped_extents;
+      scan->ext_info = xnrealloc (scan->ext_info, scan->ei_count,
+                                  sizeof (struct extent_info));
+
+      unsigned int i = 0;
+      for (i = 0; i < fiemap->fm_mapped_extents; i++)
+        {
+          assert (fm_extents[i].fe_logical <=
+                  OFF_T_MAX - fm_extents[i].fe_length);
+
+          if (si && last_ei->ext_flags ==
+              (fm_extents[i].fe_flags & ~FIEMAP_EXTENT_LAST)
+              && (last_ei->ext_logical + last_ei->ext_length
+                  == fm_extents[i].fe_logical))
+            {
+              /* Merge previous with last.  */
+              last_ei->ext_length += fm_extents[i].fe_length;
+              /* Copy flags in case different.  */
+              last_ei->ext_flags = fm_extents[i].fe_flags;
+            }
+          else if ((si == 0 && scan->scan_start > fm_extents[i].fe_logical)
+                   || (si && last_ei->ext_logical + last_ei->ext_length >
+                       fm_extents[i].fe_logical))
+            {
+              /* BTRFS before 2.6.38 could return overlapping extents
+                 for sparse files.  We adjust the returned extents
+                 rather than failing, as otherwise it would be inefficient
+                 to detect this on the initial scan.  */
+              uint64_t new_logical;
+              uint64_t length_adjust;
+              if (si == 0)
+                new_logical = scan->scan_start;
+              else
+                {
+                  /* We could return here if scan->scan_start == 0
+                     but don't so as to minimize special cases.  */
+                  new_logical = last_ei->ext_logical + last_ei->ext_length;
+                }
+              length_adjust = new_logical - fm_extents[i].fe_logical;
+              /* If an extent is contained within the previous one, fail.  */
+              if (length_adjust < fm_extents[i].fe_length)
+                {
+                  if (scan->scan_start == 0)
+                    scan->initial_scan_failed = true;
+                  return false;
+                }
+              fm_extents[i].fe_logical = new_logical;
+              fm_extents[i].fe_length -= length_adjust;
+              /* Process the adjusted extent again.  */
+              i--;
+              continue;
+            }
+          else
+            {
+              last_ei = scan->ext_info + si;
+              last_ei->ext_logical = fm_extents[i].fe_logical;
+              last_ei->ext_length = fm_extents[i].fe_length;
+              last_ei->ext_flags = fm_extents[i].fe_flags;
+              si++;
+            }
+        }
+
+      if (last_ei->ext_flags & FIEMAP_EXTENT_LAST)
+        scan->hit_final_extent = true;
+
+      /* If we have enough extents, discard the last as it might
+         be merged with one from the next scan.  */
+      if (si > count && !scan->hit_final_extent)
+        last_ei = scan->ext_info + --si - 1;
+
+      /* We don't bother reallocating any trailing slots.  */
+      scan->ei_count = si;
+
+      if (scan->hit_final_extent)
+        break;
+      else
+        scan->scan_start = last_ei->ext_logical + last_ei->ext_length;
+
+      if (si >= count)
+        break;
     }
 
-  /* If 0 extents are returned, then more get_extent_table() are not needed.  */
-  if (fiemap->fm_mapped_extents == 0)
-    {
-      scan->hit_final_extent = true;
-      return false;
-    }
-
-  scan->ei_count = fiemap->fm_mapped_extents;
-  scan->ext_info = xnmalloc (scan->ei_count, sizeof (struct extent_info));
-
-  unsigned int i;
-  for (i = 0; i < scan->ei_count; i++)
-    {
-      assert (fm_extents[i].fe_logical <= OFF_T_MAX);
-
-      scan->ext_info[i].ext_logical = fm_extents[i].fe_logical;
-      scan->ext_info[i].ext_length = fm_extents[i].fe_length;
-      scan->ext_info[i].ext_flags = fm_extents[i].fe_flags;
-    }
-
-  i--;
-  if (scan->ext_info[i].ext_flags & FIEMAP_EXTENT_LAST)
-    {
-      scan->hit_final_extent = true;
-      return true;
-    }
-
-  scan->scan_start = fm_extents[i].fe_logical + fm_extents[i].fe_length;
-
   return true;
 }
 #else

src/extent-scan.h

@@ -41,6 +41,9 @@ struct extent_scan
   /* Next scan start offset.  */
   off_t scan_start;
 
+  /* Flags to use for scan.  */
+  uint32_t fm_flags;
+
   /* How many extent info returned for a scan.  */
   uint32_t ei_count;
 
@@ -63,6 +66,8 @@ static inline void
 extent_scan_free (struct extent_scan *scan)
 {
   free (scan->ext_info);
+  scan->ext_info = NULL;
+  scan->ei_count = 0;
 }
 
 #endif /* EXTENT_SCAN_H */

src/fiemap.h

@@ -75,7 +75,7 @@ struct fiemap
 /* Location still pending, Sets EXTENT_UNKNOWN.  */
 # define FIEMAP_EXTENT_DELALLOC          0x00000004
 
-/* Data can not be read while fs is unmounted.  */
+/* Data cannot be read while fs is unmounted.  */
 # define FIEMAP_EXTENT_ENCODED           0x00000008
 
 /* Data is encrypted by fs.  Sets EXTENT_NO_BYPASS.  */

src/find-mount-point.c

@@ -29,11 +29,11 @@
    Give a diagnostic and return NULL if unable to determine the mount point.
    Exit if unable to restore current working directory.  */
 extern char *
-find_mount_point (const char *file, const struct stat *file_stat)
+find_mount_point (char const *file, struct stat const *file_stat)
 {
   struct saved_cwd cwd;
   struct stat last_stat;
-  char *mp = NULL;		/* The malloced mount point.  */
+  char *mp = NULL;		/* The malloc'd mount point.  */
 
   if (save_cwd (&cwd) != 0)
     {

src/find-mount-point.h

@@ -14,4 +14,4 @@
    You should have received a copy of the GNU General Public License
    along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
 
-extern char* find_mount_point (const char *, const struct stat *);
+extern char *find_mount_point (char const *, struct stat const *);

src/install.c

@@ -68,23 +68,6 @@ static bool use_default_selinux_context = true;
 # define matchpathcon_init_prefix(a, p) /* empty */
 #endif
 
-static bool change_timestamps (struct stat const *from_sb, char const *to);
-static bool change_attributes (char const *name);
-static bool copy_file (const char *from, const char *to,
-                       const struct cp_options *x);
-static bool install_file_in_file_parents (char const *from, char *to,
-                                          struct cp_options *x);
-static bool install_file_in_dir (const char *from, const char *to_dir,
-                                 const struct cp_options *x);
-static bool install_file_in_file (const char *from, const char *to,
-                                  const struct cp_options *x);
-static void get_ids (void);
-static void strip (char const *name);
-static void announce_mkdir (char const *dir, void *options);
-static int make_ancestor (char const *dir, char const *component,
-                          void *options);
-void usage (int status);
-
 /* The user name that will own the files, or NULL to make the owner
    the current user ID. */
 static char *owner_name;
@@ -131,7 +114,6 @@ static char const *strip_program = "strip";
 enum
 {
   PRESERVE_CONTEXT_OPTION = CHAR_MAX + 1,
-  PRESERVE_CONTEXT_OPTION_DEPRECATED,
   STRIP_PROGRAM_OPTION
 };
 
@@ -147,9 +129,6 @@ static struct option const long_options[] =
   {"owner", required_argument, NULL, 'o'},
   {"preserve-timestamps", no_argument, NULL, 'p'},
   {"preserve-context", no_argument, NULL, PRESERVE_CONTEXT_OPTION},
-  /* --preserve_context was silently supported until Apr 2009.
-     FIXME: disable altogether in a year or so.  */
-  {"preserve_context", no_argument, NULL, PRESERVE_CONTEXT_OPTION_DEPRECATED},
   {"strip", no_argument, NULL, 's'},
   {"strip-program", required_argument, NULL, STRIP_PROGRAM_OPTION},
   {"suffix", required_argument, NULL, 'S'},
@@ -407,6 +386,27 @@ target_directory_operand (char const *file)
   return is_a_dir;
 }
 
+/* Report that directory DIR was made, if OPTIONS requests this.  */
+static void
+announce_mkdir (char const *dir, void *options)
+{
+  struct cp_options const *x = options;
+  if (x->verbose)
+    prog_fprintf (stdout, _("creating directory %s"), quote (dir));
+}
+
+/* Make ancestor directory DIR, whose last file name component is
+   COMPONENT, with options OPTIONS.  Assume the working directory is
+   COMPONENT's parent.  */
+static int
+make_ancestor (char const *dir, char const *component, void *options)
+{
+  int r = mkdir (component, DEFAULT_MODE);
+  if (r == 0)
+    announce_mkdir (dir, options);
+  return r;
+}
+
 /* Process a command-line file name, for the -d option.  */
 static int
 process_dir (char *dir, struct savewd *wd, void *options)
@@ -419,6 +419,312 @@ process_dir (char *dir, struct savewd *wd, void *options)
           : EXIT_FAILURE);
 }
 
+/* Copy file FROM onto file TO, creating TO if necessary.
+   Return true if successful.  */
+
+static bool
+copy_file (const char *from, const char *to, const struct cp_options *x)
+{
+  bool copy_into_self;
+
+  if (copy_only_if_needed && !need_copy (from, to, x))
+    return true;
+
+  /* Allow installing from non-regular files like /dev/null.
+     Charles Karney reported that some Sun version of install allows that
+     and that sendmail's installation process relies on the behavior.
+     However, since !x->recursive, the call to "copy" will fail if FROM
+     is a directory.  */
+
+  return copy (from, to, false, x, &copy_into_self, NULL);
+}
+
+/* Set the attributes of file or directory NAME.
+   Return true if successful.  */
+
+static bool
+change_attributes (char const *name)
+{
+  bool ok = false;
+  /* chown must precede chmod because on some systems,
+     chown clears the set[ug]id bits for non-superusers,
+     resulting in incorrect permissions.
+     On System V, users can give away files with chown and then not
+     be able to chmod them.  So don't give files away.
+
+     We don't normally ignore errors from chown because the idea of
+     the install command is that the file is supposed to end up with
+     precisely the attributes that the user specified (or defaulted).
+     If the file doesn't end up with the group they asked for, they'll
+     want to know.  */
+
+  if (! (owner_id == (uid_t) -1 && group_id == (gid_t) -1)
+      && lchown (name, owner_id, group_id) != 0)
+    error (0, errno, _("cannot change ownership of %s"), quote (name));
+  else if (chmod (name, mode) != 0)
+    error (0, errno, _("cannot change permissions of %s"), quote (name));
+  else
+    ok = true;
+
+  if (use_default_selinux_context)
+    setdefaultfilecon (name);
+
+  return ok;
+}
+
+/* Set the timestamps of file DEST to match those of SRC_SB.
+   Return true if successful.  */
+
+static bool
+change_timestamps (struct stat const *src_sb, char const *dest)
+{
+  struct timespec timespec[2];
+  timespec[0] = get_stat_atime (src_sb);
+  timespec[1] = get_stat_mtime (src_sb);
+
+  if (utimens (dest, timespec))
+    {
+      error (0, errno, _("cannot set time stamps for %s"), quote (dest));
+      return false;
+    }
+  return true;
+}
+
+/* Strip the symbol table from the file NAME.
+   We could dig the magic number out of the file first to
+   determine whether to strip it, but the header files and
+   magic numbers vary so much from system to system that making
+   it portable would be very difficult.  Not worth the effort. */
+
+static void
+strip (char const *name)
+{
+  int status;
+  pid_t pid = fork ();
+
+  switch (pid)
+    {
+    case -1:
+      error (EXIT_FAILURE, errno, _("fork system call failed"));
+      break;
+    case 0:			/* Child. */
+      execlp (strip_program, strip_program, name, NULL);
+      error (EXIT_FAILURE, errno, _("cannot run %s"), strip_program);
+      break;
+    default:			/* Parent. */
+      if (waitpid (pid, &status, 0) < 0)
+        error (EXIT_FAILURE, errno, _("waiting for strip"));
+      else if (! WIFEXITED (status) || WEXITSTATUS (status))
+        error (EXIT_FAILURE, 0, _("strip process terminated abnormally"));
+      break;
+    }
+}
+
+/* Initialize the user and group ownership of the files to install. */
+
+static void
+get_ids (void)
+{
+  struct passwd *pw;
+  struct group *gr;
+
+  if (owner_name)
+    {
+      pw = getpwnam (owner_name);
+      if (pw == NULL)
+        {
+          unsigned long int tmp;
+          if (xstrtoul (owner_name, NULL, 0, &tmp, NULL) != LONGINT_OK
+              || UID_T_MAX < tmp)
+            error (EXIT_FAILURE, 0, _("invalid user %s"), quote (owner_name));
+          owner_id = tmp;
+        }
+      else
+        owner_id = pw->pw_uid;
+      endpwent ();
+    }
+  else
+    owner_id = (uid_t) -1;
+
+  if (group_name)
+    {
+      gr = getgrnam (group_name);
+      if (gr == NULL)
+        {
+          unsigned long int tmp;
+          if (xstrtoul (group_name, NULL, 0, &tmp, NULL) != LONGINT_OK
+              || GID_T_MAX < tmp)
+            error (EXIT_FAILURE, 0, _("invalid group %s"), quote (group_name));
+          group_id = tmp;
+        }
+      else
+        group_id = gr->gr_gid;
+      endgrent ();
+    }
+  else
+    group_id = (gid_t) -1;
+}
+
+void
+usage (int status)
+{
+  if (status != EXIT_SUCCESS)
+    fprintf (stderr, _("Try `%s --help' for more information.\n"),
+             program_name);
+  else
+    {
+      printf (_("\
+Usage: %s [OPTION]... [-T] SOURCE DEST\n\
+  or:  %s [OPTION]... SOURCE... DIRECTORY\n\
+  or:  %s [OPTION]... -t DIRECTORY SOURCE...\n\
+  or:  %s [OPTION]... -d DIRECTORY...\n\
+"),
+              program_name, program_name, program_name, program_name);
+      fputs (_("\
+\n\
+This install program copies files (often just compiled) into destination\n\
+locations you choose.  If you want to download and install a ready-to-use\n\
+package on a GNU/Linux system, you should instead be using a package manager\n\
+like yum(1) or apt-get(1).\n\
+\n\
+In the first three forms, copy SOURCE to DEST or multiple SOURCE(s) to\n\
+the existing DIRECTORY, while setting permission modes and owner/group.\n\
+In the 4th form, create all components of the given DIRECTORY(ies).\n\
+\n\
+"), stdout);
+      fputs (_("\
+Mandatory arguments to long options are mandatory for short options too.\n\
+"), stdout);
+      fputs (_("\
+      --backup[=CONTROL]  make a backup of each existing destination file\n\
+  -b                  like --backup but does not accept an argument\n\
+  -c                  (ignored)\n\
+  -C, --compare       compare each pair of source and destination files, and\n\
+                        in some cases, do not modify the destination at all\n\
+  -d, --directory     treat all arguments as directory names; create all\n\
+                        components of the specified directories\n\
+"), stdout);
+      fputs (_("\
+  -D                  create all leading components of DEST except the last,\n\
+                        then copy SOURCE to DEST\n\
+  -g, --group=GROUP   set group ownership, instead of process' current group\n\
+  -m, --mode=MODE     set permission mode (as in chmod), instead of rwxr-xr-x\n\
+  -o, --owner=OWNER   set ownership (super-user only)\n\
+"), stdout);
+      fputs (_("\
+  -p, --preserve-timestamps   apply access/modification times of SOURCE files\n\
+                        to corresponding destination files\n\
+  -s, --strip         strip symbol tables\n\
+      --strip-program=PROGRAM  program used to strip binaries\n\
+  -S, --suffix=SUFFIX  override the usual backup suffix\n\
+  -t, --target-directory=DIRECTORY  copy all SOURCE arguments into DIRECTORY\n\
+  -T, --no-target-directory  treat DEST as a normal file\n\
+  -v, --verbose       print the name of each directory as it is created\n\
+"), stdout);
+      fputs (_("\
+      --preserve-context  preserve SELinux security context\n\
+  -Z, --context=CONTEXT  set SELinux security context of files and directories\
+\n\
+"), stdout);
+
+      fputs (HELP_OPTION_DESCRIPTION, stdout);
+      fputs (VERSION_OPTION_DESCRIPTION, stdout);
+      fputs (_("\
+\n\
+The backup suffix is `~', unless set with --suffix or SIMPLE_BACKUP_SUFFIX.\n\
+The version control method may be selected via the --backup option or through\n\
+the VERSION_CONTROL environment variable.  Here are the values:\n\
+\n\
+"), stdout);
+      fputs (_("\
+  none, off       never make backups (even if --backup is given)\n\
+  numbered, t     make numbered backups\n\
+  existing, nil   numbered if numbered backups exist, simple otherwise\n\
+  simple, never   always make simple backups\n\
+"), stdout);
+      emit_ancillary_info ();
+    }
+  exit (status);
+}
+
+/* Copy file FROM onto file TO and give TO the appropriate
+   attributes.
+   Return true if successful.  */
+
+static bool
+install_file_in_file (const char *from, const char *to,
+                      const struct cp_options *x)
+{
+  struct stat from_sb;
+  if (x->preserve_timestamps && stat (from, &from_sb) != 0)
+    {
+      error (0, errno, _("cannot stat %s"), quote (from));
+      return false;
+    }
+  if (! copy_file (from, to, x))
+    return false;
+  if (strip_files)
+    strip (to);
+  if (x->preserve_timestamps && (strip_files || ! S_ISREG (from_sb.st_mode))
+      && ! change_timestamps (&from_sb, to))
+    return false;
+  return change_attributes (to);
+}
+
+/* Copy file FROM onto file TO, creating any missing parent directories of TO.
+   Return true if successful.  */
+
+static bool
+install_file_in_file_parents (char const *from, char *to,
+                              struct cp_options *x)
+{
+  bool save_working_directory =
+    ! (IS_ABSOLUTE_FILE_NAME (from) && IS_ABSOLUTE_FILE_NAME (to));
+  int status = EXIT_SUCCESS;
+
+  struct savewd wd;
+  savewd_init (&wd);
+  if (! save_working_directory)
+    savewd_finish (&wd);
+
+  if (mkancesdirs (to, &wd, make_ancestor, x) == -1)
+    {
+      error (0, errno, _("cannot create directory %s"), to);
+      status = EXIT_FAILURE;
+    }
+
+  if (save_working_directory)
+    {
+      int restore_result = savewd_restore (&wd, status);
+      int restore_errno = errno;
+      savewd_finish (&wd);
+      if (EXIT_SUCCESS < restore_result)
+        return false;
+      if (restore_result < 0 && status == EXIT_SUCCESS)
+        {
+          error (0, restore_errno, _("cannot create directory %s"), to);
+          return false;
+        }
+    }
+
+  return (status == EXIT_SUCCESS && install_file_in_file (from, to, x));
+}
+
+/* Copy file FROM into directory TO_DIR, keeping its same name,
+   and give the copy the appropriate attributes.
+   Return true if successful.  */
+
+static bool
+install_file_in_dir (const char *from, const char *to_dir,
+                     const struct cp_options *x)
+{
+  const char *from_base = last_component (from);
+  char *to = file_name_concat (to_dir, from_base, NULL);
+  bool ret = install_file_in_file (from, to, x);
+  free (to);
+  return ret;
+}
+
 int
 main (int argc, char **argv)
 {
@@ -529,10 +835,6 @@ main (int argc, char **argv)
           no_target_directory = true;
           break;
 
-        case PRESERVE_CONTEXT_OPTION_DEPRECATED:
-          error (0, 0, _("WARNING: --preserve_context is deprecated; "
-                         "use --preserve-context instead"));
-          /* fall through */
         case PRESERVE_CONTEXT_OPTION:
           if ( ! selinux_enabled)
             {
@@ -681,330 +983,3 @@ main (int argc, char **argv)
 
   exit (exit_status);
 }
-
-/* Copy file FROM onto file TO, creating any missing parent directories of TO.
-   Return true if successful.  */
-
-static bool
-install_file_in_file_parents (char const *from, char *to,
-                              struct cp_options *x)
-{
-  bool save_working_directory =
-    ! (IS_ABSOLUTE_FILE_NAME (from) && IS_ABSOLUTE_FILE_NAME (to));
-  int status = EXIT_SUCCESS;
-
-  struct savewd wd;
-  savewd_init (&wd);
-  if (! save_working_directory)
-    savewd_finish (&wd);
-
-  if (mkancesdirs (to, &wd, make_ancestor, x) == -1)
-    {
-      error (0, errno, _("cannot create directory %s"), to);
-      status = EXIT_FAILURE;
-    }
-
-  if (save_working_directory)
-    {
-      int restore_result = savewd_restore (&wd, status);
-      int restore_errno = errno;
-      savewd_finish (&wd);
-      if (EXIT_SUCCESS < restore_result)
-        return false;
-      if (restore_result < 0 && status == EXIT_SUCCESS)
-        {
-          error (0, restore_errno, _("cannot create directory %s"), to);
-          return false;
-        }
-    }
-
-  return (status == EXIT_SUCCESS && install_file_in_file (from, to, x));
-}
-
-/* Copy file FROM onto file TO and give TO the appropriate
-   attributes.
-   Return true if successful.  */
-
-static bool
-install_file_in_file (const char *from, const char *to,
-                      const struct cp_options *x)
-{
-  struct stat from_sb;
-  if (x->preserve_timestamps && stat (from, &from_sb) != 0)
-    {
-      error (0, errno, _("cannot stat %s"), quote (from));
-      return false;
-    }
-  if (! copy_file (from, to, x))
-    return false;
-  if (strip_files)
-    strip (to);
-  if (x->preserve_timestamps && (strip_files || ! S_ISREG (from_sb.st_mode))
-      && ! change_timestamps (&from_sb, to))
-    return false;
-  return change_attributes (to);
-}
-
-/* Copy file FROM into directory TO_DIR, keeping its same name,
-   and give the copy the appropriate attributes.
-   Return true if successful.  */
-
-static bool
-install_file_in_dir (const char *from, const char *to_dir,
-                     const struct cp_options *x)
-{
-  const char *from_base = last_component (from);
-  char *to = file_name_concat (to_dir, from_base, NULL);
-  bool ret = install_file_in_file (from, to, x);
-  free (to);
-  return ret;
-}
-
-/* Copy file FROM onto file TO, creating TO if necessary.
-   Return true if successful.  */
-
-static bool
-copy_file (const char *from, const char *to, const struct cp_options *x)
-{
-  bool copy_into_self;
-
-  if (copy_only_if_needed && !need_copy (from, to, x))
-    return true;
-
-  /* Allow installing from non-regular files like /dev/null.
-     Charles Karney reported that some Sun version of install allows that
-     and that sendmail's installation process relies on the behavior.
-     However, since !x->recursive, the call to "copy" will fail if FROM
-     is a directory.  */
-
-  return copy (from, to, false, x, &copy_into_self, NULL);
-}
-
-/* Set the attributes of file or directory NAME.
-   Return true if successful.  */
-
-static bool
-change_attributes (char const *name)
-{
-  bool ok = false;
-  /* chown must precede chmod because on some systems,
-     chown clears the set[ug]id bits for non-superusers,
-     resulting in incorrect permissions.
-     On System V, users can give away files with chown and then not
-     be able to chmod them.  So don't give files away.
-
-     We don't normally ignore errors from chown because the idea of
-     the install command is that the file is supposed to end up with
-     precisely the attributes that the user specified (or defaulted).
-     If the file doesn't end up with the group they asked for, they'll
-     want to know.  */
-
-  if (! (owner_id == (uid_t) -1 && group_id == (gid_t) -1)
-      && lchown (name, owner_id, group_id) != 0)
-    error (0, errno, _("cannot change ownership of %s"), quote (name));
-  else if (chmod (name, mode) != 0)
-    error (0, errno, _("cannot change permissions of %s"), quote (name));
-  else
-    ok = true;
-
-  if (use_default_selinux_context)
-    setdefaultfilecon (name);
-
-  return ok;
-}
-
-/* Set the timestamps of file TO to match those of file FROM.
-   Return true if successful.  */
-
-static bool
-change_timestamps (struct stat const *from_sb, char const *to)
-{
-  struct timespec timespec[2];
-  timespec[0] = get_stat_atime (from_sb);
-  timespec[1] = get_stat_mtime (from_sb);
-
-  if (utimens (to, timespec))
-    {
-      error (0, errno, _("cannot set time stamps for %s"), quote (to));
-      return false;
-    }
-  return true;
-}
-
-/* Strip the symbol table from the file NAME.
-   We could dig the magic number out of the file first to
-   determine whether to strip it, but the header files and
-   magic numbers vary so much from system to system that making
-   it portable would be very difficult.  Not worth the effort. */
-
-static void
-strip (char const *name)
-{
-  int status;
-  pid_t pid = fork ();
-
-  switch (pid)
-    {
-    case -1:
-      error (EXIT_FAILURE, errno, _("fork system call failed"));
-      break;
-    case 0:			/* Child. */
-      execlp (strip_program, strip_program, name, NULL);
-      error (EXIT_FAILURE, errno, _("cannot run %s"), strip_program);
-      break;
-    default:			/* Parent. */
-      if (waitpid (pid, &status, 0) < 0)
-        error (EXIT_FAILURE, errno, _("waiting for strip"));
-      else if (! WIFEXITED (status) || WEXITSTATUS (status))
-        error (EXIT_FAILURE, 0, _("strip process terminated abnormally"));
-      break;
-    }
-}
-
-/* Initialize the user and group ownership of the files to install. */
-
-static void
-get_ids (void)
-{
-  struct passwd *pw;
-  struct group *gr;
-
-  if (owner_name)
-    {
-      pw = getpwnam (owner_name);
-      if (pw == NULL)
-        {
-          unsigned long int tmp;
-          if (xstrtoul (owner_name, NULL, 0, &tmp, NULL) != LONGINT_OK
-              || UID_T_MAX < tmp)
-            error (EXIT_FAILURE, 0, _("invalid user %s"), quote (owner_name));
-          owner_id = tmp;
-        }
-      else
-        owner_id = pw->pw_uid;
-      endpwent ();
-    }
-  else
-    owner_id = (uid_t) -1;
-
-  if (group_name)
-    {
-      gr = getgrnam (group_name);
-      if (gr == NULL)
-        {
-          unsigned long int tmp;
-          if (xstrtoul (group_name, NULL, 0, &tmp, NULL) != LONGINT_OK
-              || GID_T_MAX < tmp)
-            error (EXIT_FAILURE, 0, _("invalid group %s"), quote (group_name));
-          group_id = tmp;
-        }
-      else
-        group_id = gr->gr_gid;
-      endgrent ();
-    }
-  else
-    group_id = (gid_t) -1;
-}
-
-/* Report that directory DIR was made, if OPTIONS requests this.  */
-static void
-announce_mkdir (char const *dir, void *options)
-{
-  struct cp_options const *x = options;
-  if (x->verbose)
-    prog_fprintf (stdout, _("creating directory %s"), quote (dir));
-}
-
-/* Make ancestor directory DIR, whose last file name component is
-   COMPONENT, with options OPTIONS.  Assume the working directory is
-   COMPONENT's parent.  */
-static int
-make_ancestor (char const *dir, char const *component, void *options)
-{
-  int r = mkdir (component, DEFAULT_MODE);
-  if (r == 0)
-    announce_mkdir (dir, options);
-  return r;
-}
-
-void
-usage (int status)
-{
-  if (status != EXIT_SUCCESS)
-    fprintf (stderr, _("Try `%s --help' for more information.\n"),
-             program_name);
-  else
-    {
-      printf (_("\
-Usage: %s [OPTION]... [-T] SOURCE DEST\n\
-  or:  %s [OPTION]... SOURCE... DIRECTORY\n\
-  or:  %s [OPTION]... -t DIRECTORY SOURCE...\n\
-  or:  %s [OPTION]... -d DIRECTORY...\n\
-"),
-              program_name, program_name, program_name, program_name);
-      fputs (_("\
-\n\
-This install program copies files (often just compiled) into destination\n\
-locations you choose.  If you want to download and install a ready-to-use\n\
-package on a GNU/Linux system, you should instead be using a package manager\n\
-like yum(1) or apt-get(1).\n\
-\n\
-In the first three forms, copy SOURCE to DEST or multiple SOURCE(s) to\n\
-the existing DIRECTORY, while setting permission modes and owner/group.\n\
-In the 4th form, create all components of the given DIRECTORY(ies).\n\
-\n\
-"), stdout);
-      fputs (_("\
-Mandatory arguments to long options are mandatory for short options too.\n\
-"), stdout);
-      fputs (_("\
-      --backup[=CONTROL]  make a backup of each existing destination file\n\
-  -b                  like --backup but does not accept an argument\n\
-  -c                  (ignored)\n\
-  -C, --compare       compare each pair of source and destination files, and\n\
-                        in some cases, do not modify the destination at all\n\
-  -d, --directory     treat all arguments as directory names; create all\n\
-                        components of the specified directories\n\
-"), stdout);
-      fputs (_("\
-  -D                  create all leading components of DEST except the last,\n\
-                        then copy SOURCE to DEST\n\
-  -g, --group=GROUP   set group ownership, instead of process' current group\n\
-  -m, --mode=MODE     set permission mode (as in chmod), instead of rwxr-xr-x\n\
-  -o, --owner=OWNER   set ownership (super-user only)\n\
-"), stdout);
-      fputs (_("\
-  -p, --preserve-timestamps   apply access/modification times of SOURCE files\n\
-                        to corresponding destination files\n\
-  -s, --strip         strip symbol tables\n\
-      --strip-program=PROGRAM  program used to strip binaries\n\
-  -S, --suffix=SUFFIX  override the usual backup suffix\n\
-  -t, --target-directory=DIRECTORY  copy all SOURCE arguments into DIRECTORY\n\
-  -T, --no-target-directory  treat DEST as a normal file\n\
-  -v, --verbose       print the name of each directory as it is created\n\
-"), stdout);
-      fputs (_("\
-      --preserve-context  preserve SELinux security context\n\
-  -Z, --context=CONTEXT  set SELinux security context of files and directories\
-\n\
-"), stdout);
-
-      fputs (HELP_OPTION_DESCRIPTION, stdout);
-      fputs (VERSION_OPTION_DESCRIPTION, stdout);
-      fputs (_("\
-\n\
-The backup suffix is `~', unless set with --suffix or SIMPLE_BACKUP_SUFFIX.\n\
-The version control method may be selected via the --backup option or through\n\
-the VERSION_CONTROL environment variable.  Here are the values:\n\
-\n\
-"), stdout);
-      fputs (_("\
-  none, off       never make backups (even if --backup is given)\n\
-  numbered, t     make numbered backups\n\
-  existing, nil   numbered if numbered backups exist, simple otherwise\n\
-  simple, never   always make simple backups\n\
-"), stdout);
-      emit_ancillary_info ();
-    }
-  exit (status);
-}

src/libstdbuf.c

@@ -22,7 +22,7 @@
 #include "system.h"
 #include "verify.h"
 
-/* Note currently for glibc (2.3.5) the following call does not change the
+/* Note currently for glibc (2.3.5) the following call does not change
    the buffer size, and more problematically does not give any indication
    that the new size request was ignored:
 

src/ls.c

@@ -1131,8 +1131,8 @@ is_colored (enum indicator_no type)
   size_t len = color_indicator[type].len;
   char const *s = color_indicator[type].string;
   return ! (len == 0
-            || (len == 1 && strncmp (s, "0", 1) == 0)
-            || (len == 2 && strncmp (s, "00", 2) == 0));
+            || (len == 1 && STRNCMP_LIT (s, "0") == 0)
+            || (len == 2 && STRNCMP_LIT (s, "00") == 0));
 }
 
 static void
@@ -1996,7 +1996,7 @@ decode_switches (int argc, char **argv)
         if (! (style = getenv ("TIME_STYLE")))
           style = bad_cast ("locale");
 
-      while (strncmp (style, posix_prefix, sizeof posix_prefix - 1) == 0)
+      while (STREQ_LEN (style, posix_prefix, sizeof posix_prefix - 1))
         {
           if (! hard_locale (LC_TIME))
             return optind;
@@ -2388,7 +2388,7 @@ parse_ls_color (void)
     }
 
   if (color_indicator[C_LINK].len == 6
-      && !strncmp (color_indicator[C_LINK].string, "target", 6))
+      && !STRNCMP_LIT (color_indicator[C_LINK].string, "target"))
     color_symlink_as_referent = true;
 }
 
@@ -4178,7 +4178,7 @@ print_color_indicator (const struct fileinfo *f, bool symlink_target)
         }
       else if (S_ISLNK (mode))
         type = ((!linkok
-                 && (!strncmp (color_indicator[C_LINK].string, "target", 6)
+                 && (!STRNCMP_LIT (color_indicator[C_LINK].string, "target")
                      || color_indicator[C_ORPHAN].string))
                 ? C_ORPHAN : C_LINK);
       else if (S_ISFIFO (mode))
@@ -4209,8 +4209,8 @@ print_color_indicator (const struct fileinfo *f, bool symlink_target)
       for (ext = color_ext_list; ext != NULL; ext = ext->next)
         {
           if (ext->ext.len <= len
-              && strncmp (name - ext->ext.len, ext->ext.string,
-                          ext->ext.len) == 0)
+              && STREQ_LEN (name - ext->ext.len, ext->ext.string,
+                            ext->ext.len))
             break;
         }
     }

src/md5sum.c

@@ -262,7 +262,7 @@ split_3 (char *s, size_t s_len,
 
   /* Check for BSD-style checksum line. */
   algo_name_len = strlen (DIGEST_TYPE_STRING);
-  if (strncmp (s + i, DIGEST_TYPE_STRING, algo_name_len) == 0)
+  if (STREQ_LEN (s + i, DIGEST_TYPE_STRING, algo_name_len))
     {
       if (s[i + algo_name_len] == ' ')
         ++i;

src/sort.c

@@ -109,12 +109,16 @@ struct rlimit { size_t rlim_cur; };
    and is responsible for merging TOTAL lines. */
 #define MAX_MERGE(total, level) (((total) >> (2 * ((level) + 1))) + 1)
 
-/* Heuristic value for the number of lines for which it is worth
-   creating a subthread, during an internal merge sort, on a machine
-   that has processors galore.  Currently this number is just a guess.
-   This value must be at least 4.  We don't know of any machine where
-   this number has any practical effect.  */
-enum { SUBTHREAD_LINES_HEURISTIC = 4 };
+/* Heuristic value for the number of lines for which it is worth creating
+   a subthread, during an internal merge sort.  I.e., it is a small number
+   of "average" lines for which sorting via two threads is faster than
+   sorting via one on an "average" system.  On an dual-core 2.0 GHz i686
+   system with 3GB of RAM and 2MB of L2 cache, a file containing 128K
+   lines of gensort -a output is sorted slightly faster with --parallel=2
+   than with --parallel=1.  By contrast, using --parallel=1 is about 10%
+   faster than using --parallel=2 with a 64K-line input.  */
+enum { SUBTHREAD_LINES_HEURISTIC = 128 * 1024 };
+verify (4 <= SUBTHREAD_LINES_HEURISTIC);
 
 /* The number of threads after which there are
    diminishing performance gains.  */
@@ -315,8 +319,12 @@ static size_t merge_buffer_size = MAX (MIN_MERGE_BUFFER_SIZE, 256 * 1024);
    specified by the user.  Zero if the user has not specified a size.  */
 static size_t sort_size;
 
-/* The guessed size for non-regular files.  */
-#define INPUT_FILE_SIZE_GUESS (1024 * 1024)
+/* The initial allocation factor for non-regular files.
+   This is used, e.g., when reading from a pipe.
+   Don't make it too big, since it is multiplied by ~130 to
+   obtain the size of the actual buffer sort will allocate.
+   Also, there may be 8 threads all doing this at the same time.  */
+#define INPUT_FILE_SIZE_GUESS (128 * 1024)
 
 /* Array of directory names in which any temporary files are to be created. */
 static char const **temp_dirs;

src/split.c

@@ -895,12 +895,12 @@ main (int argc, char **argv)
           /* skip any whitespace */
           while (isspace (to_uchar (*optarg)))
             optarg++;
-          if (strncmp (optarg, "r/", 2) == 0)
+          if (STRNCMP_LIT (optarg, "r/") == 0)
             {
               split_type = type_rr;
               optarg += 2;
             }
-          else if (strncmp (optarg, "l/", 2) == 0)
+          else if (STRNCMP_LIT (optarg, "l/") == 0)
             {
               split_type = type_chunk_lines;
               optarg += 2;

src/stdbuf.c

@@ -209,7 +209,7 @@ set_LD_PRELOAD (void)
   char const *const *path = search_path;
   char *libstdbuf;
 
-  do
+  while (true)
     {
       struct stat sb;
 
@@ -224,8 +224,11 @@ set_LD_PRELOAD (void)
       if (stat (libstdbuf, &sb) == 0)   /* file_exists  */
         break;
       free (libstdbuf);
+
+      ++path;
+      if ( ! *path)
+        error (EXIT_CANCELED, 0, _("failed to find %s"), quote (LIB_NAME));
     }
-  while (*++path);
 
   /* FIXME: Do we need to support libstdbuf.dll, c:, '\' separators etc?  */
 

src/system.h

@@ -257,6 +257,13 @@ select_plural (uintmax_t n)
 }
 
 #define STREQ(a, b) (strcmp (a, b) == 0)
+#define STREQ_LEN(a, b, n) (strncmp (a, b, n) == 0)
+#define STRPREFIX(a, b) (strncmp(a, b, strlen (b)) == 0)
+
+/* Just like strncmp, but the first argument must be a literal string
+   and you don't specify the length.  */
+#define STRNCMP_LIT(s, literal) \
+  strncmp (s, "" literal "", sizeof (literal) - 1)
 
 #if !HAVE_DECL_GETLOGIN
 char *getlogin ();
@@ -607,7 +614,7 @@ emit_ancillary_info (void)
   /* Don't output this redundant message for English locales.
      Note we still output for 'C' so that it gets included in the man page.  */
   const char *lc_messages = setlocale (LC_MESSAGES, NULL);
-  if (lc_messages && strncmp (lc_messages, "en_", 3))
+  if (lc_messages && STRNCMP_LIT (lc_messages, "en_"))
     {
       /* TRANSLATORS: Replace LANG_CODE in this URL with your language code
          <http://translationproject.org/team/LANG_CODE.html> to form one of

src/test.c

@@ -173,7 +173,8 @@ get_mtime (char const *filename, struct timespec *mtime)
 static bool
 binop (char const *s)
 {
-  return ((STREQ (s,   "=")) || (STREQ (s,  "!=")) || (STREQ (s, "-nt")) ||
+  return ((STREQ (s,   "=")) || (STREQ (s,  "!=")) || (STREQ (s, "==")) ||
+          (STREQ (s,   "-nt")) ||
           (STREQ (s, "-ot")) || (STREQ (s, "-ef")) || (STREQ (s, "-eq")) ||
           (STREQ (s, "-ne")) || (STREQ (s, "-lt")) || (STREQ (s, "-le")) ||
           (STREQ (s, "-gt")) || (STREQ (s, "-ge")));
@@ -360,7 +361,8 @@ binary_operator (bool l_is_l)
       test_syntax_error (_("unknown binary operator"), argv[op]);
     }
 
-  if (argv[op][0] == '=' && !argv[op][1])
+  if (argv[op][0] == '=' && (!argv[op][1] ||
+       ((argv[op][1] == '=') && !argv[op][2])))
     {
       bool value = STREQ (argv[pos], argv[pos + 2]);
       pos += 3;

src/tr.c

@@ -548,7 +548,7 @@ look_up_char_class (char const *class_str, size_t len)
   enum Char_class i;
 
   for (i = 0; i < ARRAY_CARDINALITY (char_class_name); i++)
-    if (strncmp (class_str, char_class_name[i], len) == 0
+    if (STREQ_LEN (class_str, char_class_name[i], len)
         && strlen (char_class_name[i]) == len)
       return i;
   return CC_NO_CLASS;

src/uname.c

@@ -333,7 +333,7 @@ main (int argc, char **argv)
 
               /* Hack "safely" around the ppc vs. powerpc return value. */
               if (cputype == CPU_TYPE_POWERPC
-                  && strncmp (element, "ppc", 3) == 0)
+                  && STRNCMP_LIT (element, "ppc") == 0)
                 element = "powerpc";
             }
 # endif

src/uptime.c

@@ -151,11 +151,7 @@ print_uptime (size_t n, const STRUCT_UTMP *this)
   printf (ngettext ("%lu user", "%lu users", entries),
           (unsigned long int) entries);
 
-#if defined HAVE_GETLOADAVG || defined C_GETLOADAVG
   loads = getloadavg (avg, 3);
-#else
-  loads = -1;
-#endif
 
   if (loads == -1)
     putchar ('\n');

src/wc.c

@@ -709,6 +709,9 @@ main (int argc, char **argv)
       ai = argv_iter_init_argv (files);
     }
 
+  if (!ai)
+    xalloc_die ();
+
   fstatus = get_input_fstatus (nfiles, files);
   number_width = compute_number_width (nfiles, fstatus);
 
@@ -719,16 +722,17 @@ main (int argc, char **argv)
       bool skip_file = false;
       enum argv_iter_err ai_err;
       char *file_name = argv_iter (ai, &ai_err);
-      if (ai_err == AI_ERR_EOF)
-        break;
       if (!file_name)
         {
           switch (ai_err)
             {
+            case AI_ERR_EOF:
+              goto argv_iter_done;
             case AI_ERR_READ:
-              error (EXIT_FAILURE, errno, _("%s: read error"),
-                     quote (files_from));
-              continue;
+              error (0, errno, _("%s: read error"),
+                     quotearg_colon (files_from));
+              ok = false;
+              goto argv_iter_done;
             case AI_ERR_MEM:
               xalloc_die ();
             default:
@@ -770,6 +774,7 @@ main (int argc, char **argv)
       else
         ok &= wc_file (file_name, &fstatus[nfiles ? i : 0]);
     }
+ argv_iter_done:
 
   /* No arguments on the command line is fine.  That means read from stdin.
      However, no arguments on the --files0-from input stream is an error

src/who.c

@@ -584,7 +584,7 @@ scan_entries (size_t n, const STRUCT_UTMP *utmp_buf)
       ttyname_b = ttyname (STDIN_FILENO);
       if (!ttyname_b)
         return;
-      if (strncmp (ttyname_b, DEV_DIR_WITH_TRAILING_SLASH, DEV_DIR_LEN) == 0)
+      if (STRNCMP_LIT (ttyname_b, DEV_DIR_WITH_TRAILING_SLASH) == 0)
         ttyname_b += DEV_DIR_LEN;	/* Discard /dev/ prefix.  */
     }
 

tests/Makefile.am

@@ -11,6 +11,7 @@ EXTRA_DIST =		\
   check.mk		\
   envvar-check		\
   filefrag-extent-compare \
+  fiemap-capable	\
   init.cfg		\
   init.sh		\
   lang-default		\
@@ -26,7 +27,7 @@ root_tests =					\
   cp/special-bits				\
   cp/cp-mv-enotsup-xattr			\
   cp/capability					\
-  cp/sparse-fiemap                              \
+  cp/sparse-fiemap				\
   dd/skip-seek-past-dev				\
   install/install-C-root			\
   ls/capability					\
@@ -320,6 +321,7 @@ TESTS =						\
   cp/dir-vs-file				\
   cp/existing-perm-race				\
   cp/fail-perm					\
+  cp/fiemap-empty                               \
   cp/fiemap-perf                                \
   cp/fiemap-2                                   \
   cp/file-perm-race				\
@@ -327,6 +329,7 @@ TESTS =						\
   cp/link					\
   cp/link-no-deref				\
   cp/link-preserve				\
+  cp/link-symlink				\
   cp/no-deref-link1				\
   cp/no-deref-link2				\
   cp/no-deref-link3				\
@@ -351,6 +354,7 @@ TESTS =						\
   df/unreadable					\
   dd/direct					\
   dd/misc					\
+  dd/nocache					\
   dd/not-rewound				\
   dd/reblock					\
   dd/skip-seek					\
@@ -369,6 +373,7 @@ TESTS =						\
   du/exclude					\
   du/fd-leak					\
   du/files0-from				\
+  du/files0-from-dir				\
   du/hard-link					\
   du/inacc-dest					\
   du/inacc-dir					\

tests/check.mk

@@ -40,8 +40,13 @@ vc_exe_in_TESTS: Makefile
 check: vc_exe_in_TESTS
 .PHONY: vc_exe_in_TESTS
 
-built_programs = \
-  (cd $(top_builddir)/src && MAKEFLAGS= $(MAKE) -s built_programs.list)
+CLEANFILES =
+CLEANFILES += .built-programs
+check-am: .built-programs
+.built-programs:
+	$(AM_V_GEN)(cd $(top_builddir)/src				\
+            && MAKEFLAGS= $(MAKE) -s built_programs.list)		\
+          > $@-t && mv $@-t $@
 
 # Note that the first lines are statements.  They ensure that environment
 # variables that can perturb tests are unset or set to expected values.
@@ -76,7 +81,7 @@ TESTS_ENVIRONMENT =				\
   abs_top_builddir='$(abs_top_builddir)'	\
   abs_top_srcdir='$(abs_top_srcdir)'		\
   abs_srcdir='$(abs_srcdir)'			\
-  built_programs="`$(built_programs)`"		\
+  built_programs="`cat .built-programs`"	\
   host_os=$(host_os)				\
   host_triplet='$(host_triplet)'		\
   srcdir='$(srcdir)'				\

tests/cp/fiemap-2

@@ -20,7 +20,8 @@
 print_ver_ cp
 
 # Require a fiemap-enabled FS.
-fiemap_capable_ . \
+touch fiemap_chk # check a file rather than current dir for best coverage
+fiemap_capable_ fiemap_chk \
   || skip_ "this file system lacks FIEMAP support"
 
 # Exercise the code that handles a file ending in a hole.

tests/cp/fiemap-empty

@@ -0,0 +1,96 @@
+#!/bin/sh
+# Test cp reads unwritten extents efficiently
+
+# Copyright (C) 2011 Free Software Foundation, Inc.
+
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+. "${srcdir=.}/init.sh"; path_prepend_ ../src
+print_ver_ cp
+
+touch fiemap_chk
+fiemap_capable_ fiemap_chk ||
+  skip_test_ 'this file system lacks FIEMAP support'
+rm fiemap_chk
+
+# TODO: rather than requiring `fallocate`, possible add
+# this functionality to truncate --alloc
+fallocate --help >/dev/null || skip_test_ 'The fallocate utility is required'
+fallocate -l 1 -n falloc.test ||
+  skip_test_ 'this file system lacks FALLOCATE support'
+rm falloc.test
+
+# Require more space than we'll actually use, so that
+# tests run in parallel do not run out of space.
+# Otherwise, with inadequate space, simply running the following
+# fallocate command would induce a temporary disk-full condition,
+# which would cause failure of unrelated tests run in parallel.
+require_file_system_bytes_free_ 800000000
+
+fallocate -l 600MiB space.test ||
+  skip_test_ 'this test needs at least 600MiB free space'
+
+# Disable this test on old BTRFS (e.g. Fedora 14)
+# which reports ordinary extents for unwritten ones.
+filefrag space.test || skip_test_ 'the `filefrag` utility is missing'
+filefrag -v space.test | grep -F 'unwritten' > /dev/null ||
+  skip_test_ 'this file system does not report empty extents as "unwritten"'
+
+rm space.test
+
+# Ensure we read a large empty file quickly
+fallocate -l 300MiB empty.big || framework_failure
+timeout 3 cp --sparse=always empty.big cp.test || fail=1
+test $(stat -c %s empty.big) = $(stat -c %s cp.test) || fail=1
+rm empty.big cp.test
+
+# Ensure we handle extents beyond file size correctly.
+# Note until we support fallocate, we will not maintain
+# the file allocation.  FIXME: amend this test when fallocate is supported.
+fallocate -l 10MiB -n unwritten.withdata || framework_failure
+dd count=10 if=/dev/urandom conv=notrunc iflag=fullblock of=unwritten.withdata
+cp unwritten.withdata cp.test || fail=1
+test $(stat -c %s unwritten.withdata) = $(stat -c %s cp.test) || fail=1
+cmp unwritten.withdata cp.test || fail=1
+rm unwritten.withdata cp.test
+
+# The following to generate unaccounted extents followed by a hole, is not
+# supported by ext4 at least. The ftruncate discards all extents not
+# accounted for in the size.
+#  fallocate -l 10MiB -n unacc.withholes
+#  dd count=10 if=/dev/urandom conv=notrunc iflag=fullblock of=unacc.withholes
+#  truncate -s20M unacc.withholes
+
+# Ensure we handle a hole after empty extents correctly.
+# Since all extents are accounted for in the size,
+# we can maintain the allocation independently from
+# fallocate() support.
+fallocate -l 10MiB empty.withholes
+truncate -s 20M empty.withholes
+sectors_per_block=$(expr $(stat -c %o .) / 512)
+cp empty.withholes cp.test || fail=1
+test $(stat -c %s empty.withholes) = $(stat -c %s cp.test) || fail=1
+# These are usually equal but can vary by an IO block due to alignment
+alloc_diff=$(expr $(stat -c %b empty.withholes) - $(stat -c %b cp.test))
+alloc_diff=$(echo $alloc_diff | tr -d -- -) # abs()
+test $alloc_diff -le $sectors_per_block || fail=1
+# Again with SPARSE_ALWAYS
+cp --sparse=always empty.withholes cp.test || fail=1
+test $(stat -c %s empty.withholes) = $(stat -c %s cp.test) || fail=1
+# cp.test should take 0 space, but allowing for some systems
+# that store default extended attributes in data blocks
+test $(stat -c %b cp.test) -le $sectors_per_block || fail=1
+rm empty.withholes cp.test
+
+Exit $fail

tests/cp/fiemap-perf

@@ -20,12 +20,24 @@
 print_ver_ cp
 
 # Require a fiemap-enabled FS.
-fiemap_capable_ . \
-  || skip_ "this file system lacks FIEMAP support"
+touch fiemap_chk
+fiemap_capable_ fiemap_chk ||
+  skip_ "this file system lacks FIEMAP support"
+
+# Exclude ext3 (or unknown fs types)
+# as the emulated extent scanning is slow
+df -t ext3 . >/dev/null &&
+  skip_ "ext3 has known slow FIEMAP scanning"
 
 # Create a large-but-sparse file.
 timeout 10 truncate -s1T f || framework_failure_
 
+# Disable this test on old BTRFS (e.g. Fedora 14)
+# which reports (unwritten) extents for holes.
+filefrag f || skip_test_ 'the `filefrag` utility is missing'
+filefrag f | grep -F ': 0 extents found' > /dev/null ||
+  skip_test_ 'this file system reports extents for holes'
+
 # Nothing can read (much less write) that many bytes in so little time.
 timeout 10 cp f f2 || fail=1
 

tests/cp/link-symlink

@@ -0,0 +1,41 @@
+#!/bin/sh
+# Ensure that cp -a --link maintains timestamps if possible
+
+# Copyright (C) 2011 Free Software Foundation, Inc.
+
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+. "${srcdir=.}/init.sh"; path_prepend_ ../src
+print_ver_ cp
+
+# Check that the timestamps of the symlink are copied
+# if we're using hardlink to symlink emulation.
+touch file
+ln -s file link || framework_failure
+touch -m -h -d 2011-01-01 link ||
+  skip_test_ "Your system doesn't support updating symlink timestamps"
+case `stat --format=%y link` in
+  2011-01-01*) ;;
+  *) skip_test_ "Your system doesn't support updating symlink timestamps" ;;
+esac
+
+# link.cp is probably a hardlink, but may also be a symlink
+# In either case the timestamp should match the original.
+cp -al link link.cp
+case `stat --format=%y link.cp` in
+  2011-01-01*) ;;
+  *) fail=1 ;;
+esac
+
+Exit $fail

tests/cp/preserve-gid

@@ -52,14 +52,6 @@ t0() {
   fi
 }
 
-t1() {
-  f=$1; shift
-  u=$1; shift
-  g=$1; shift
-  t0 "$f" "$u" "$g" setuidgid -g "$nameless_gid1,$nameless_gid2" \
-      "$nameless_uid" "$@"
-}
-
 nameless_uid=`$PERL -le '
   foreach my $i (1000..16*1024-1)
     {
@@ -105,6 +97,30 @@ t0 b1 "$nameless_uid" "$nameless_gid2" cp -p
 t0 c0 0 "$nameless_gid1" cp -p
 t0 c1 0 "$nameless_gid2" cp -p
 
+# For the remaining tests, we need a cp binary that is accessible to a user
+# with UID of $nameless_uid.  The build directory may not be accessible,
+# so create a temporary directory and copy cp into it, ensure that
+# $nameless_uid can access it and then make that directory the search path.
+tmp_path=
+cleanup_() { rm -rf "$tmp_path"; }
+
+# Cause mktemp to create a directory directly under /tmp.
+# Setting TMPDIR explicitly is required here, in case $TMPDIR
+# is not readable by our nameless IDs.
+test -d /tmp && TMPDIR=/tmp
+tmp_path=$(mktemp -d) || fail_ "failed to create temporary directory"
+cp "$abs_path_dir_/cp" "$tmp_path"
+chmod -R a+rx "$tmp_path"
+
+t1() {
+  f=$1; shift
+  u=$1; shift
+  g=$1; shift
+  t0 "$f" "$u" "$g" \
+      setuidgid -g "$nameless_gid1,$nameless_gid2" \
+      "$nameless_uid" env PATH="$tmp_path" "$@"
+}
+
 t1 a0 "$nameless_uid" "$nameless_gid1" cp
 t1 b0 "$nameless_uid" "$nameless_gid1" cp
 t1 b1 "$nameless_uid" "$nameless_gid1" cp

tests/cp/sparse-fiemap

@@ -18,9 +18,13 @@
 
 . "${srcdir=.}/init.sh"; path_prepend_ ../src
 print_ver_ cp
+$PERL -e 1 || skip_test_ 'you lack perl'
 
-if fiemap_capable_ . ; then
-  : # Current dir is on a partition with working extents.  Good!
+# The test was seen to fail on ext3 so exclude that type
+# (or any file system where the type can't be determined)
+touch fiemap_chk
+if fiemap_capable_ fiemap_chk && ! df -t ext3 . >/dev/null; then
+  : # Current partition has working extents.  Good!
 else
   # It's not;  we need to create one, hence we need root access.
   require_root_
@@ -49,8 +53,6 @@ fi
 # in which it calls ioctl (fd, FS_IOC_FIEMAP,...
 # This also verifies that non-trivial extents are preserved.
 
-$PERL -e 1 || skip_test_ 'skipping part of this test; you lack perl'
-
 # Extract logical block number and length pairs from filefrag -v output.
 # The initial sed is to remove the "eof" from the normally-empty "flags" field.
 # Similarly, remove flags values like "unknown,delalloc,eof".
@@ -66,39 +68,41 @@ for i in $(seq 1 2 21); do
     $PERL -e 'BEGIN { $n = '$i' * 1024; *F = *STDOUT }' \
           -e 'for (1..'$j') { sysseek (*F, $n, 1)' \
           -e '&& syswrite (*F, chr($_)x$n) or die "$!"}' > j1 || fail=1
-    # sync
+
+    # Note there is an implicit sync performed by cp on Linux kernels
+    # before 2.6.39 to work around bugs in EXT4 and BTRFS.
+    # Note also the -s parameter to the filefrag commands below
+    # for the same reasons.
     cp --sparse=always j1 j2 || fail=1
-    # sync
-    # Technically we may need the 'sync' uses above, but
-    # uncommenting them makes this test take much longer.
 
     cmp j1 j2 || fail=1
-    filefrag -v j1 | grep extent \
-      || skip_test_ 'skipping part of this test; you lack filefrag'
+    if ! filefrag -vs j1 | grep -F extent >/dev/null; then
+      test $skip != 1 && warn_ 'skipping part; you lack filefrag'
+      skip=1
+    else
+      # Here is sample filefrag output:
+      #   $ perl -e 'BEGIN{$n=16*1024; *F=*STDOUT}' \
+      #          -e 'for (1..5) { sysseek(*F,$n,1)' \
+      #          -e '&& syswrite *F,"."x$n or die "$!"}' > j
+      #   $ filefrag -v j
+      #   File system type is: ef53
+      #   File size of j is 163840 (40 blocks, blocksize 4096)
+      #    ext logical physical expected length flags
+      #      0       4  6258884               4
+      #      1      12  6258892  6258887      4
+      #      2      20  6258900  6258895      4
+      #      3      28  6258908  6258903      4
+      #      4      36  6258916  6258911      4 eof
+      #   j: 6 extents found
 
-    # Here is sample filefrag output:
-    #   $ perl -e 'BEGIN{$n=16*1024; *F=*STDOUT}' \
-    #          -e 'for (1..5) { sysseek(*F,$n,1)' \
-    #          -e '&& syswrite *F,"."x$n or die "$!"}' > j
-    #   $ filefrag -v j
-    #   File system type is: ef53
-    #   File size of j is 163840 (40 blocks, blocksize 4096)
-    #    ext logical physical expected length flags
-    #      0       4  6258884               4
-    #      1      12  6258892  6258887      4
-    #      2      20  6258900  6258895      4
-    #      3      28  6258908  6258903      4
-    #      4      36  6258916  6258911      4 eof
-    #   j: 6 extents found
-
-    # exclude the physical block numbers; they always differ
-    filefrag -v j1 > ff1 || fail=1
-    filefrag -v j2 > ff2 || fail=1
-    { f ff1; f ff2; } \
-      | $PERL $abs_top_srcdir/tests/filefrag-extent-compare \
-        || { fail=1; break; }
+      # exclude the physical block numbers; they always differ
+      filefrag -v j1 > ff1 || framework_failure
+      filefrag -vs j2 > ff2 || framework_failure
+      { f ff1; f ff2; } | $PERL $abs_top_srcdir/tests/filefrag-extent-compare ||
+        fail=1
+    fi
+    test $fail = 1 && break 2
   done
-  test $fail = 1 && break
 done
 
 Exit $fail

tests/cp/sparse-to-pipe

@@ -26,6 +26,10 @@ timeout 10 cat pipe > copy &
 
 truncate -s1M sparse || framework_failure_
 cp sparse pipe || fail=1
+
+# Ensure that the cat has completed before comparing.
+wait
+
 cmp sparse copy || fail=1
 
 Exit $fail

tests/dd/nocache

@@ -0,0 +1,58 @@
+#!/bin/sh
+# Ensure dd handles the 'nocache' flag
+
+# Copyright (C) 2011 Free Software Foundation, Inc.
+
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+. "${srcdir=.}/init.sh"; path_prepend_ ../src
+print_ver_ dd
+
+# This should not call posix_fadvise
+dd iflag=nocache oflag=nocache if=/dev/null of=/dev/null || fail=1
+
+# We should get an error for trying to process a pipe
+dd count=0 | dd iflag=nocache count=0 && fail=1
+
+# O_DIRECT is orthogonal to drop cache so mutually exclusive
+dd iflag=nocache,direct if=/dev/null && fail=1
+
+# The rest ensure that the documented uses cases
+# proceed without error
+for f in ifile ofile; do
+  dd if=/dev/zero of=$f conv=fdatasync count=100 || framework_failure
+done
+
+# Advise to drop cache for whole file
+if ! dd if=ifile iflag=nocache count=0 2>err; then
+  if grep -F 'Operation not supported' err >/dev/null; then
+     warn_ 'skipping part; this file system lacks support for posix_fadvise()'
+     skip=1
+  else
+    fail=1
+  fi
+fi
+
+if test "$skip" != 1; then
+  # Ensure drop cache for whole file
+  dd of=ofile oflag=nocache conv=notrunc,fdatasync count=0 || fail=1
+
+  # Drop cache for part of file
+  dd if=ifile iflag=nocache skip=10 count=10 of=/dev/null || fail=1
+
+  # Stream data just using readahead cache
+  dd if=ifile of=ofile iflag=nocache oflag=nocache || fail=1
+fi
+
+Exit $fail

tests/du/files0-from-dir

@@ -0,0 +1,39 @@
+#!/bin/sh
+# ensure that du and wc handle --files0-from=DIR
+
+# Copyright (C) 2011 Free Software Foundation, Inc.
+
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+. "${srcdir=.}/init.sh"; path_prepend_ ../src
+print_ver_ du wc
+
+mkdir dir
+
+# Skip this test if reading from a directory succeeds.
+# In that case, using --files0-from=dir would yield garbage,
+# interpreting the directory entry as a sequence of
+# NUL-separated file names.
+cat dir > /dev/null && skip_ "cat dir/ succeeds"
+
+for prog in du wc; do
+  $prog --files0-from=dir > /dev/null 2>err && fail=1
+  printf "$prog: dir:\n" > exp || fail=1
+  # The diagnostic string is usually "Is a directory" (ENOTDIR),
+  # but accept a different string or errno value.
+  sed 's/dir:.*/dir:/' err > k; mv k err
+  compare err exp || fail=1
+done
+
+Exit $fail

tests/fiemap-capable

@@ -0,0 +1,16 @@
+import struct, fcntl, sys, os
+
+def sizeof(t): return struct.calcsize(t)
+IOCPARM_MASK = 0x7f
+IOC_OUT = 0x40000000
+IOC_IN = 0x80000000
+IOC_INOUT = (IOC_IN|IOC_OUT)
+def _IOWR(x,y,t): return (IOC_INOUT|((sizeof(t)&IOCPARM_MASK)<<16)|((x)<<8)|y)
+
+try:
+    fd = os.open (len (sys.argv) == 2 and sys.argv[1] or '.', os.O_RDONLY)
+    struct_fiemap = '=qqllll'
+    FS_IOC_FIEMAP = _IOWR (ord ('f'), 11, struct_fiemap)
+    fcntl.ioctl (fd, FS_IOC_FIEMAP, struct.pack(struct_fiemap, 0,~0,0,0,0,0))
+except:
+    sys.exit (1)

tests/filefrag-extent-compare

@@ -28,30 +28,47 @@ my @b;
 foreach my $i (0..@A/2-1) { $a[$i] = { L_BLK => $A[2*$i], LEN => $A[2*$i+1] } };
 foreach my $i (0..@B/2-1) { $b[$i] = { L_BLK => $B[2*$i], LEN => $B[2*$i+1] } };
 
+# Merge adjacent extents in array E.
+sub merge_extents($)
+{
+  my ($e) = @_;
+
+  my $i = 0;
+  while (1)
+    {
+      !defined $e->[$i+1]
+        and last;
+      $e->[$i]->{L_BLK} + $e->[$i]->{LEN} != $e->[$i+1]->{L_BLK}
+        and ++$i, next;
+
+      $e->[$i]->{LEN} += $e->[$i+1]->{LEN};
+      # Remove $e->[$i+1]
+      splice @$e, $i+1, 1;
+    }
+}
+
+merge_extents \@a;
+merge_extents \@b;
+
+@a == @b
+  or die "$ME: extent counts differ, even after adjustment\n";
+
 my $i = 0;
-my $j = 0;
-while (1)
+while (defined $a[$i])
   {
-    !defined $a[$i] && !defined $b[$j]
-      and exit 0;
-    defined $a[$i] && defined $b[$j]
-      or die "\@a and \@b have different lengths, even after adjustment\n";
-    ($a[$i]->{L_BLK} == $b[$j]->{L_BLK}
-     && $a[$i]->{LEN} == $b[$j]->{LEN})
-      and next;
-    ($a[$i]->{LEN} < $b[$j]->{LEN}
-     && exists $a[$i+1] && $a[$i]->{LEN} + $a[$i+1]->{LEN} == $b[$j]->{LEN})
-      and ++$i, next;
-    exists $b[$j+1] && $a[$i]->{LEN} == $b[$i]->{LEN} + $b[$i+1]->{LEN}
-      and ++$j, next;
-    die "differing extent:\n"
-      . "  [$i]=$a[$i]->{L_BLK} $a[$i]->{LEN}\n"
-      . "  [$j]=$b[$j]->{L_BLK} $b[$j]->{LEN}\n"
-  }
-continue
-  {
-    ++$i;
-    ++$j;
+    my $start_match = $a[$i]->{L_BLK} == $b[$i]->{L_BLK};
+    my $len_match = $a[$i]->{LEN} == $b[$i]->{LEN};
+    if ( ! ($start_match && ($len_match || $i == (@a - 1))))
+      {
+        # On XFS on Linux kernel 2.6.38, it was seen that the size of the
+        # last extent can vary, and can extend beyond the length of the file.
+        # So we ignore the length of the last extent, because if the
+        # file is the wrong length we'll get failures elsewhere.
+        die "$ME: differing extent:\n"
+          . "  [$i]=$a[$i]->{L_BLK} $a[$i]->{LEN}\n"
+            . "  [$i]=$b[$i]->{L_BLK} $b[$i]->{LEN}\n";
+      }
+    $i++;
   }
 
 ### Setup "GNU" style for perl-mode and cperl-mode.

tests/init.cfg

@@ -151,6 +151,14 @@ require_built_()
   test $skip_ = yes && skip_test_ "required program(s) not built"
 }
 
+require_file_system_bytes_free_()
+{
+  local req=$1
+  local expr=$(stat -f --printf "$req / %S <= %a" .)
+  awk "BEGIN{ exit !($expr) }" \
+    || skip_test_ "this test needs at least $req bytes of free space"
+}
+
 uid_is_privileged_()
 {
   # Make sure id -u succeeds.
@@ -216,6 +224,13 @@ skip_if_()
 
 require_selinux_()
 {
+  # When in a chroot of an SELinux-enabled system, but with a mock-simulated
+  # SELinux-*disabled* system, recognize that SELinux is disabled system wide:
+  grep 'selinuxfs$' /proc/filesystems > /dev/null \
+    || skip_test_ "this system lacks SELinux support"
+
+  # Independent of whether SELinux is enabled system-wide,
+  # the current file system may lack SELinux support.
   case `ls -Zd .` in
     '? .'|'unlabeled .')
       skip_test_ "this system (or maybe just" \
@@ -295,13 +310,16 @@ require_proc_pid_status_()
     kill $pid
 }
 
-# Return nonzero if the specified directory is on a file system for
-# which FIEMAP support exists, and the file system type is new enough
-# (unlike ext2 and ext3) that it is hard to find an instance *without*
-# FIEMAP support.
+# Return nonzero if the specified path is on a file system for
+# which FIEMAP support exists.  Note some file systems (like ext3 and btrfs)
+# only support FIEMAP for files, not directories.
 fiemap_capable_()
 {
-  df -T -t btrfs -t xfs -t ext4 -t ocfs2 -t gfs2 "$@"
+  if ! python < /dev/null; then
+    warn_ 'fiemap_capable_: python missing: assuming not fiemap capable'
+    return 1
+  fi
+  python $abs_srcdir/fiemap-capable "$@"
 }
 
 # Does the current (working-dir) file system support sparse files?

tests/init.sh

@@ -74,10 +74,10 @@ Exit () { set +e; (exit $1); exit $1; }
 # the reason for skip/failure to console, rather than to the .log files.
 : ${stderr_fileno_=2}
 
-warn_() { echo "$@" 1>&$stderr_fileno_; }
-fail_() { warn_ "$ME_: failed test: $@"; Exit 1; }
-skip_() { warn_ "$ME_: skipped test: $@"; Exit 77; }
-framework_failure_() { warn_ "$ME_: set-up failure: $@"; Exit 99; }
+warn_ () { echo "$@" 1>&$stderr_fileno_; }
+fail_ () { warn_ "$ME_: failed test: $@"; Exit 1; }
+skip_ () { warn_ "$ME_: skipped test: $@"; Exit 77; }
+framework_failure_ () { warn_ "$ME_: set-up failure: $@"; Exit 99; }
 
 # Sanitize this shell to POSIX mode, if possible.
 DUALCASE=1; export DUALCASE
@@ -193,7 +193,7 @@ fi
 test -n "$EXEEXT" && shopt -s expand_aliases
 
 # Enable glibc's malloc-perturbing option.
-# This is cheap and useful for exposing code that depends on the fact that
+# This is useful for exposing code that depends on the fact that
 # malloc-related functions often return memory that is mostly zeroed.
 # If you have the time and cycles, use valgrind to do an even better job.
 : ${MALLOC_PERTURB_=87}
@@ -202,22 +202,22 @@ export MALLOC_PERTURB_
 # This is a stub function that is run upon trap (upon regular exit and
 # interrupt).  Override it with a per-test function, e.g., to unmount
 # a partition, or to undo any other global state changes.
-cleanup_() { :; }
+cleanup_ () { :; }
 
 if ( diff --version < /dev/null 2>&1 | grep GNU ) > /dev/null 2>&1; then
-  compare() { diff -u "$@"; }
+  compare () { diff -u "$@"; }
 elif ( cmp --version < /dev/null 2>&1 | grep GNU ) > /dev/null 2>&1; then
-  compare() { cmp -s "$@"; }
+  compare () { cmp -s "$@"; }
 else
-  compare() { cmp "$@"; }
+  compare () { cmp "$@"; }
 fi
 
 # An arbitrary prefix to help distinguish test directories.
-testdir_prefix_() { printf gt; }
+testdir_prefix_ () { printf gt; }
 
 # Run the user-overridable cleanup_ function, remove the temporary
 # directory and exit with the incoming value of $?.
-remove_tmp_()
+remove_tmp_ ()
 {
   __st=$?
   cleanup_
@@ -233,7 +233,7 @@ remove_tmp_()
 # contains only the specified bytes (see the case stmt below), then print
 # a space-separated list of those names and return 0.  Otherwise, don't
 # print anything and return 1.  Naming constraints apply also to DIR.
-find_exe_basenames_()
+find_exe_basenames_ ()
 {
   feb_dir_=$1
   feb_fail_=0
@@ -266,7 +266,7 @@ find_exe_basenames_()
 # PROG that simply invokes PROG.exe, then return 0.  If any selected
 # file name or the directory name, $1, contains an unexpected character,
 # define no alias and return 1.
-create_exe_shims_()
+create_exe_shims_ ()
 {
   case $EXEEXT in
     '') return 0 ;;
@@ -288,7 +288,7 @@ create_exe_shims_()
 
 # Use this function to prepend to PATH an absolute name for each
 # specified, possibly-$initial_cwd_-relative, directory.
-path_prepend_()
+path_prepend_ ()
 {
   while test $# != 0; do
     path_dir_=$1
@@ -311,7 +311,7 @@ path_prepend_()
   export PATH
 }
 
-setup_()
+setup_ ()
 {
   if test "$VERBOSE" = yes; then
     # Test whether set -x may cause the selected shell to corrupt an
@@ -364,7 +364,7 @@ setup_()
 #  - make only $MAX_TRIES_ attempts
 
 # Helper function.  Print $N pseudo-random bytes from a-zA-Z0-9.
-rand_bytes_()
+rand_bytes_ ()
 {
   n_=$1
 
@@ -396,7 +396,7 @@ rand_bytes_()
     | LC_ALL=C tr -c $chars_ 01234567$chars_$chars_$chars_
 }
 
-mktempd_()
+mktempd_ ()
 {
   case $# in
   2);;

tests/ls/color-norm

@@ -27,7 +27,7 @@ umask 22
 # Output time as something constant
 export TIME_STYLE="+norm"
 
-# helper to to strip ls columns up to "norm" time
+# helper to strip ls columns up to "norm" time
 qls() { sed 's/-r.*norm/norm/'; }
 
 touch exe || framework_failure

tests/misc/cut

@@ -150,6 +150,10 @@ my @Tests =
    {ERR=>$no_endpoint}],
   ['inval5', '-f', '1-,-', {IN=>''}, {OUT=>''}, {EXIT=>1}, {ERR=>$no_endpoint}],
   ['inval6', '-f', '-1,-', {IN=>''}, {OUT=>''}, {EXIT=>1}, {ERR=>$no_endpoint}],
+  # This would evoke a segfault from 5.3.0..6.10
+  ['big-unbounded-b', '--output-d=:', '-b1234567890-', {IN=>''}, {OUT=>''}],
+  ['big-unbounded-c', '--output-d=:', '-c1234567890-', {IN=>''}, {OUT=>''}],
+  ['big-unbounded-f', '--output-d=:', '-f1234567890-', {IN=>''}, {OUT=>''}],
  );
 
 @Tests = triple_test \@Tests;

tests/misc/help-version

@@ -226,7 +226,7 @@ id_setup () { args=-u; }
 
 # Use env to avoid invoking built-in sleep of Solaris 11's /bin/sh.
 kill_setup () {
-  env sleep 10m &
+  env sleep 31.5 &
   args=$!
 }
 

tests/misc/stdbuf

@@ -104,7 +104,7 @@ retry_delay_ stdbuf_unbuffer .1 6 || fail=1
 
 # Ensure block buffering stdout takes effect
 # We don't currently test block buffering failures as
-# this doesn't work on on GLIBC-2.7 or GLIBC-2.9 at least.
+# this doesn't work on GLIBC-2.7 or GLIBC-2.9 at least.
    # stdbuf_blockbuffer()
    # {
    #   local delay="$1"

tests/misc/test

@@ -132,6 +132,8 @@ my @Tests =
 
   ['streq-1', qw(t = t)],
   ['streq-2', qw(t = f), {EXIT=>1}],
+  ['streqeq-1', qw(t == t)],
+  ['streqeq-2', qw(t == f), {EXIT=>1}],
   ['streq-3', qw(! = !)],
   ['streq-4', qw(= = =)],
   ['streq-5', "'(' = '('"],

tests/mv/sticky-to-xpart

@@ -47,8 +47,7 @@ version=`
 `
 case $version in
   $PACKAGE_VERSION) ;;
-  *) echo "$0: cannot access just-built mv as user $NON_ROOT_USERNAME" 1>&2
-     fail=1 ;;
+  *) skip_ "cannot access just-built mv as user $NON_ROOT_USERNAME";;
 esac
 
 setuidgid $NON_ROOT_USERNAME env PATH="$PATH" \

tests/tail-2/inotify-rotate

@@ -24,12 +24,12 @@ fi
 . "${srcdir=.}/init.sh"; path_prepend_ ../src
 expensive_
 
-# Wait up to 10 seconds for grep REGEXP FILE to succeed.
+# Wait several seconds for grep REGEXP FILE to succeed.
 # Usage: grep_timeout REGEXP FILE
 grep_timeout()
 {
     local j
-    for j in $(seq 100); do
+    for j in $(seq 150); do
         grep $1 $2 > /dev/null && return 0
         sleep 0.1
     done