src/env_parallel.sh

   1 #!/usr/bin/env sh
   2
   3 # This file must be sourced in sh:
   4 #
   5 #   . `which env_parallel.sh`
   6 #
   7 # after which 'env_parallel' works
   8 #
   9 #
  10 # Copyright (C) 2016-2021 Ole Tange, http://ole.tange.dk and Free
  11 # Software Foundation, Inc.
  12 #
  13 # This program is free software; you can redistribute it and/or modify
  14 # it under the terms of the GNU General Public License as published by
  15 # the Free Software Foundation; either version 3 of the License, or
  16 # (at your option) any later version.
  17 #
  18 # This program is distributed in the hope that it will be useful, but
  19 # WITHOUT ANY WARRANTY; without even the implied warranty of
  20 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  21 # General Public License for more details.
  22 #
  23 # You should have received a copy of the GNU General Public License
  24 # along with this program; if not, see <http://www.gnu.org/licenses/>
  25 # or write to the Free Software Foundation, Inc., 51 Franklin St,
  26 # Fifth Floor, Boston, MA 02110-1301 USA
  27 #
  28 # SPDX-FileCopyrightText: 2021 Ole Tange, http://ole.tange.dk and Free Software and Foundation, Inc.
  29 # SPDX-License-Identifier: GPL-3.0-or-later
  30 # shellcheck disable=SC2006
  31
  32 env_parallel() {
  33     # env_parallel.sh
  34
  35     _names_of_ALIASES() {
  36         # alias fails on Unixware 5
  37         for _i in `alias 2>/dev/null | perl -ne 's/^alias //;s/^(\S+)=.*/$1/ && print' 2>/dev/null`; do
  38             # Check if this name really is an alias
  39             # or just part of a multiline alias definition
  40             if alias "$_i" >/dev/null 2>/dev/null; then
  41                 echo "$_i"
  42             fi
  43         done
  44     }
  45     _bodies_of_ALIASES() {
  46         # alias may return:
  47         #   myalias='definition' (GNU/Linux ash)
  48         #   alias myalias='definition' (FreeBSD ash)
  49         # so remove 'alias ' from first line
  50         for _i in "$@"; do
  51                 echo 'alias '"`alias "$_i" | perl -pe '1..1 and s/^alias //'`"
  52         done
  53     }
  54     _names_of_maybe_FUNCTIONS() {
  55         set | perl -ne '/^([A-Z_0-9]+)\s*\(\)\s*\{?$/i and print "$1\n"'
  56     }
  57     _names_of_FUNCTIONS() {
  58         # myfunc is a function
  59         # shellcheck disable=SC2046
  60         LANG=C type `_names_of_maybe_FUNCTIONS` |
  61             perl -ne '/^(\S+) is a function$/ and not $seen{$1}++ and print "$1\n"'
  62     }
  63     _bodies_of_FUNCTIONS() {
  64         LANG=C type "$@" | perl -ne '/^(\S+) is a function$/ or print'
  65     }
  66     _names_of_VARIABLES() {
  67         # This may screw up if variables contain \n and =
  68         set | perl -ne 's/^(\S+?)=.*/$1/ and print;'
  69     }
  70     _bodies_of_VARIABLES() {
  71         # Crappy typeset -p
  72         for _i in "$@"
  73         do
  74             perl -e 'print @ARGV' "$_i="
  75             eval echo "\"\$$_i\"" | perl -e '$/=undef; $a=<>; chop($a); print $a' |
  76                 perl -pe 's/[\002-\011\013-\032\\\#\?\`\(\)\{\}\[\]\^\*\<\=\>\~\|\; \"\!\$\&\202-\377]/\\$&/go;'"s/'/\\\'/g; s/[\n]/'\\n'/go;";
  77             echo
  78         done
  79     }
  80     _ignore_HARDCODED() {
  81         # These names cannot be detected
  82         echo '(_|TIMEOUT)'
  83     }
  84     _ignore_READONLY() {
  85         # shellcheck disable=SC1078,SC1079,SC2026
  86         readonly | perl -e '@r = map {
  87                 chomp;
  88                 # sh on UnixWare: readonly TIMEOUT
  89                 # ash: readonly var='val'
  90                 # ksh: var='val'
  91                 s/^(readonly )?([^= ]*)(=.*|)$/$2/ or
  92                 # bash: declare -ar BASH_VERSINFO=([0]="4" [1]="4")
  93                 # zsh: typeset -r var='val'
  94                   s/^\S+\s+\S+\s+(\S[^=]*)(=.*|$)/$1/;
  95                 $_ } <>;
  96             $vars = join "|",map { quotemeta $_ } @r;
  97             print $vars ? "($vars)" : "(,,nO,,VaRs,,)";
  98             '
  99     }
 100     _remove_bad_NAMES() {
 101         # Do not transfer vars and funcs from env_parallel
 102         _ignore_RO="`_ignore_READONLY`"
 103         _ignore_HARD="`_ignore_HARDCODED`"
 104         # Macos-grep does not like long patterns
 105         # Old Solaris grep does not support -E
 106         # Perl Version of:
 107         # grep -Ev '^(...)$' |
 108         perl -ne '/^(
 109                      PARALLEL_ENV|
 110                      PARALLEL_TMP|
 111                      _alias_NAMES|
 112                      _bodies_of_ALIASES|
 113                      _bodies_of_FUNCTIONS|
 114                      _bodies_of_VARIABLES|
 115                      _error_PAR|
 116                      _function_NAMES|
 117                      _get_ignored_VARS|
 118                      _grep_REGEXP|
 119                      _ignore_HARD|
 120                      _ignore_HARDCODED|
 121                      _ignore_READONLY|
 122                      _ignore_RO|
 123                      _ignore_UNDERSCORE|
 124                      _list_alias_BODIES|
 125                      _list_function_BODIES|
 126                      _list_variable_VALUES|
 127                      _make_grep_REGEXP|
 128                      _names_of_ALIASES|
 129                      _names_of_FUNCTIONS|
 130                      _names_of_VARIABLES|
 131                      _names_of_maybe_FUNCTIONS|
 132                      _parallel_exit_CODE|
 133                      _prefix_PARALLEL_ENV|
 134                      _prefix_PARALLEL_ENV|
 135                      _remove_bad_NAMES|
 136                      _remove_readonly|
 137                      _variable_NAMES|
 138                      _warning_PAR|
 139                      _which_PAR)$/x and next;
 140             # Filter names matching --env
 141             /^'"$_grep_REGEXP"'$/ or next;
 142             /^'"$_ignore_UNDERSCORE"'$/ and next;
 143             # Remove readonly variables
 144             /^'"$_ignore_RO"'$/ and next;
 145             /^'"$_ignore_HARD"'$/ and next;
 146             print;'
 147     }
 148     _get_ignored_VARS() {
 149         perl -e '
 150             for(@ARGV){
 151                 $next_is_env and push @envvar, split/,/, $_;
 152                 $next_is_env=/^--env$/;
 153             }
 154             if(grep { /^_$/ } @envvar) {
 155                 if(not open(IN, "<", "$ENV{HOME}/.parallel/ignored_vars")) {
 156                     print STDERR "parallel: Error: ",
 157                     "Run \"parallel --record-env\" in a clean environment first.\n";
 158                 } else {
 159                     chomp(@ignored_vars = <IN>);
 160                 }
 161             }
 162             if($ENV{PARALLEL_IGNORED_NAMES}) {
 163                 push @ignored_vars, split/\s+/, $ENV{PARALLEL_IGNORED_NAMES};
 164                 chomp @ignored_vars;
 165             }
 166             $vars = join "|",map { quotemeta $_ } @ignored_vars;
 167             print $vars ? "($vars)" : "(,,nO,,VaRs,,)";
 168             ' -- "$@"
 169     }
 170
 171     # Get the --env variables if set
 172     # --env _ should be ignored
 173     # and convert  a b c  to (a|b|c)
 174     # If --env not set: Match everything (.*)
 175     _make_grep_REGEXP() {
 176         perl -e '
 177             for(@ARGV){
 178                 /^_$/ and $next_is_env = 0;
 179                 $next_is_env and push @envvar, split/,/, $_;
 180                 $next_is_env = /^--env$/;
 181             }
 182             $vars = join "|",map { quotemeta $_ } @envvar;
 183             print $vars ? "($vars)" : "(.*)";
 184             ' -- "$@"
 185     }
 186     _which_PAR() {
 187         # type returns:
 188         #   ll is an alias for ls -l (in ash)
 189         #   bash is a tracked alias for /bin/bash
 190         #   true is a shell builtin (in bash)
 191         #   myfunc is a function (in bash)
 192         #   myfunc is a shell function (in zsh)
 193         #   which is /usr/bin/which (in sh, bash)
 194         #   which is hashed (/usr/bin/which)
 195         #   gi is aliased to `grep -i' (in bash)
 196         #   aliased to `alias | /usr/bin/which --tty-only --read-alias --show-dot --show-tilde'
 197         # Return 0 if found, 1 otherwise
 198         LANG=C type "$@" |
 199             perl -pe '$exit += (s/ is an alias for .*// ||
 200                                 s/ is aliased to .*// ||
 201                                 s/ is a function// ||
 202                                 s/ is a shell function// ||
 203                                 s/ is a shell builtin// ||
 204                                 s/.* is hashed .(\S+).$/$1/ ||
 205                                 s/.* is (a tracked alias for )?//);
 206                       END { exit not $exit }'
 207     }
 208     _warning_PAR() {
 209         echo "env_parallel: Warning: $*" >&2
 210     }
 211     _error_PAR() {
 212         echo "env_parallel: Error: $*" >&2
 213     }
 214
 215     if _which_PAR parallel >/dev/null; then
 216         true parallel found in path
 217     else
 218         # shellcheck disable=SC2016
 219         _error_PAR 'parallel must be in $PATH.'
 220         return 255
 221     fi
 222
 223     # Grep regexp for vars given by --env
 224     _grep_REGEXP="`_make_grep_REGEXP \"$@\"`"
 225     unset _make_grep_REGEXP
 226
 227     # Deal with --env _
 228     _ignore_UNDERSCORE="`_get_ignored_VARS \"$@\"`"
 229     unset _get_ignored_VARS
 230
 231     # --record-env
 232     if perl -e 'exit grep { /^--record-env$/ } @ARGV' -- "$@"; then
 233         true skip
 234     else
 235         (_names_of_ALIASES;
 236          _names_of_FUNCTIONS;
 237          _names_of_VARIABLES) |
 238             cat > "$HOME"/.parallel/ignored_vars
 239         return 0
 240     fi
 241
 242     # --session
 243     if perl -e 'exit grep { /^--session$/ } @ARGV' -- "$@"; then
 244         true skip
 245     else
 246         # Insert ::: between each level of session
 247         # so you can pop off the last ::: at --end-session
 248         PARALLEL_IGNORED_NAMES="`echo \"$PARALLEL_IGNORED_NAMES\";
 249           echo :::;
 250           (_names_of_ALIASES;
 251            _names_of_FUNCTIONS;
 252            _names_of_VARIABLES) | perl -ne '
 253             BEGIN{
 254               map { $ignored_vars{$_}++ }
 255                 split/\s+/, $ENV{PARALLEL_IGNORED_NAMES};
 256             }
 257             chomp;
 258             for(split/\s+/) {
 259               if(not $ignored_vars{$_}) {
 260                 print $_,\"\\n\";
 261               }
 262             }
 263             '`"
 264         export PARALLEL_IGNORED_NAMES
 265         return 0
 266     fi
 267     if perl -e 'exit grep { /^--end.?session$/ } @ARGV' -- "$@"; then
 268         true skip
 269     else
 270         # Pop off last ::: from PARALLEL_IGNORED_NAMES
 271         PARALLEL_IGNORED_NAMES="`perl -e '
 272           $ENV{PARALLEL_IGNORED_NAMES} =~ s/(.*):::.*?$/$1/s;
 273           print $ENV{PARALLEL_IGNORED_NAMES}
 274         '`"
 275         return 0
 276     fi
 277     # Grep alias names
 278     _alias_NAMES="`_names_of_ALIASES | _remove_bad_NAMES | xargs echo`"
 279     _list_alias_BODIES="_bodies_of_ALIASES $_alias_NAMES"
 280     if [ "$_alias_NAMES" = "" ] ; then
 281         # no aliases selected
 282         _list_alias_BODIES="true"
 283     fi
 284     unset _alias_NAMES
 285
 286     # Grep function names
 287     _function_NAMES="`_names_of_FUNCTIONS | _remove_bad_NAMES | xargs echo`"
 288     _list_function_BODIES="_bodies_of_FUNCTIONS $_function_NAMES"
 289     if [ "$_function_NAMES" = "" ] ; then
 290         # no functions selected
 291         _list_function_BODIES="true"
 292     fi
 293     unset _function_NAMES
 294
 295     # Grep variable names
 296     _variable_NAMES="`_names_of_VARIABLES | _remove_bad_NAMES | xargs echo`"
 297     _list_variable_VALUES="_bodies_of_VARIABLES $_variable_NAMES"
 298     if [ "$_variable_NAMES" = "" ] ; then
 299         # no variables selected
 300         _list_variable_VALUES="true"
 301     fi
 302     unset _variable_NAMES
 303
 304     PARALLEL_ENV="`
 305         $_list_alias_BODIES;
 306         $_list_function_BODIES;
 307         $_list_variable_VALUES;
 308     `"
 309     export PARALLEL_ENV
 310     unset _list_alias_BODIES _list_variable_VALUES _list_function_BODIES
 311     unset _bodies_of_ALIASES _bodies_of_VARIABLES _bodies_of_FUNCTIONS
 312     unset _names_of_ALIASES _names_of_VARIABLES _names_of_FUNCTIONS
 313     unset _ignore_HARDCODED _ignore_READONLY _ignore_UNDERSCORE
 314     unset _remove_bad_NAMES _grep_REGEXP
 315     unset _prefix_PARALLEL_ENV
 316     # Test if environment is too big
 317     # shellcheck disable=SC2092
 318     if `_which_PAR true` >/dev/null 2>/dev/null ; then
 319         parallel "$@"
 320         _parallel_exit_CODE=$?
 321         # Clean up variables/functions
 322         unset PARALLEL_ENV
 323         unset _which_PAR _which_TRUE
 324         unset _warning_PAR _error_PAR
 325         # Unset _parallel_exit_CODE before return
 326         eval "unset _parallel_exit_CODE; return $_parallel_exit_CODE"
 327     else
 328         unset PARALLEL_ENV;
 329         _error_PAR "Your environment is too big."
 330         _error_PAR "You can try 3 different approaches:"
 331         _error_PAR "1. Run 'env_parallel --session' before you set"
 332         _error_PAR "   variables or define functions."
 333         _error_PAR "2. Use --env and only mention the names to copy."
 334         _error_PAR "3. Try running this in a clean environment once:"
 335         _error_PAR "     env_parallel --record-env"
 336         _error_PAR "   And then use '--env _'"
 337         _error_PAR "For details see: man env_parallel"
 338         return 255
 339     fi
 340 }
 341
 342 parset() {
 343     _parset_PARALLEL_PRG=parallel
 344     _parset_main "$@"
 345 }
 346
 347 env_parset() {
 348     _parset_PARALLEL_PRG=env_parallel
 349     _parset_main "$@"
 350 }
 351
 352 _parset_main() {
 353     # If $1 contains ',' or space:
 354     #   Split on , to get the destination variable names
 355     # If $1 is a single destination variable name:
 356     #   Treat it as the name of an array
 357     #
 358     #   # Create array named myvar
 359     #   parset myvar echo ::: {1..10}
 360     #   echo ${myvar[5]}
 361     #
 362     #   # Put output into $var_a $var_b $var_c
 363     #   varnames=(var_a var_b var_c)
 364     #   parset "${varnames[*]}" echo ::: {1..3}
 365     #   echo $var_c
 366     #
 367     #   # Put output into $var_a4 $var_b4 $var_c4
 368     #   parset "var_a4 var_b4 var_c4" echo ::: {1..3}
 369     #   echo $var_c4
 370
 371     _make_TEMP() {
 372         # mktemp does not exist on some OS
 373         perl -e 'use File::Temp qw(tempfile);
 374                  $ENV{"TMPDIR"} ||= "/tmp";
 375                  print((tempfile(DIR=>$ENV{"TMPDIR"}, TEMPLATE => "parXXXXX"))[1])'
 376     }
 377
 378     _parset_NAME="$1"
 379     if [ "$_parset_NAME" = "" ] ; then
 380         echo parset: Error: No destination variable given. >&2
 381         echo parset: Error: Try: >&2
 382         echo parset: Error: ' ' parset myarray echo ::: foo bar >&2
 383         return 255
 384     fi
 385     if [ "$_parset_NAME" = "--help" ] ; then
 386         echo parset: Error: Usage: >&2
 387         echo parset: Error: ' ' parset varname GNU Parallel options and command >&2
 388         echo
 389         parallel --help
 390         return 255
 391     fi
 392     if [ "$_parset_NAME" = "--version" ] ; then
 393         echo "parset 20211122 (GNU parallel `parallel --minversion 1`)"
 394         echo "Copyright (C) 2007-2021 Ole Tange, http://ole.tange.dk and Free Software"
 395         echo "Foundation, Inc."
 396         echo "License GPLv3+: GNU GPL version 3 or later <https://gnu.org/licenses/gpl.html>"
 397         echo "This is free software: you are free to change and redistribute it."
 398         echo "GNU parallel comes with no warranty."
 399         echo
 400         echo "Web site: https://www.gnu.org/software/parallel"
 401         echo
 402         echo "When using programs that use GNU Parallel to process data for publication"
 403         echo "please cite as described in 'parallel --citation'."
 404         echo
 405         return 255
 406     fi
 407     shift
 408     echo "$_parset_NAME" |
 409         perl -ne 'chomp;for (split /[, ]/) {
 410             # Allow: var_32 var[3]
 411             if(not /^[a-zA-Z_][a-zA-Z_0-9]*(\[\d+\])?$/) {
 412                 print STDERR "parset: Error: $_ is an invalid variable name.\n";
 413                 print STDERR "parset: Error: Variable names must be letter followed by letters or digits.\n";
 414                 print STDERR "parset: Error: Usage:\n";
 415                 print STDERR "parset: Error:   parset varname GNU Parallel options and command\n";
 416                 $exitval = 255;
 417             }
 418         }
 419         exit $exitval;
 420         ' || return 255
 421     _exit_FILE=`_make_TEMP`
 422     if perl -e 'exit not grep /,| /, @ARGV' "$_parset_NAME" ; then
 423         # $_parset_NAME contains , or space
 424         # Split on , or space to get the names
 425         # shellcheck disable=SC2016,SC2046
 426         eval "`
 427             # Compute results into files
 428             ($_parset_PARALLEL_PRG --files -k "$@"; echo $? > "$_exit_FILE") |
 429                 # var1= cat tmpfile1; rm tmpfile1
 430                 # var2= cat tmpfile2; rm tmpfile2
 431                 parallel --plain -q echo '{2}=\`cat {1}; rm {1}\`' :::: - :::+ \`
 432                     echo "$_parset_NAME" |
 433                         perl -pe 's/,/ /g'
 434                          \`
 435             `"
 436     else
 437         # $_parset_NAME does not contain , or space
 438         # => $_parset_NAME is the name of the array to put data into
 439         # Supported in: bash zsh ksh mksh
 440         # Arrays do not work in: sh ash dash
 441         eval "$_parset_NAME=( $(
 442             # Compute results into files. Save exit value
 443             ($_parset_PARALLEL_PRG --files -k "$@"; echo $? > "$_exit_FILE") |
 444                 perl -pe 'chop;$_="\"\`cat $_; rm $_\`\" "'
 445             ) )"
 446     fi
 447     unset _parset_NAME _parset_PARALLEL_PRG _parallel_exit_CODE
 448     # Unset _exit_FILE before return
 449     eval "unset _exit_FILE; return \`cat $_exit_FILE; rm $_exit_FILE\`"
 450 }