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) 2016-2020 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 env_parallel() {
  29     # env_parallel.sh
  30
  31     _names_of_ALIASES() {
  32         # alias fails on Unixware 5
  33         for _i in `alias 2>/dev/null | perl -ne 's/^alias //;s/^(\S+)=.*/$1/ && print' 2>/dev/null`; do
  34             # Check if this name really is an alias
  35             # or just part of a multiline alias definition
  36             if alias $_i >/dev/null 2>/dev/null; then
  37                 echo $_i
  38             fi
  39         done
  40     }
  41     _bodies_of_ALIASES() {
  42         # alias may return:
  43         #   myalias='definition' (GNU/Linux ash)
  44         #   alias myalias='definition' (FreeBSD ash)
  45         # so remove 'alias ' from first line
  46         for _i in "$@"; do
  47                 echo 'alias '"`alias $_i | perl -pe '1..1 and s/^alias //'`"
  48         done
  49     }
  50     _names_of_maybe_FUNCTIONS() {
  51         set | perl -ne '/^([A-Z_0-9]+)\s*\(\)\s*\{?$/i and print "$1\n"'
  52     }
  53     _names_of_FUNCTIONS() {
  54         # myfunc is a function
  55         LANG=C type `_names_of_maybe_FUNCTIONS` |
  56             perl -ne '/^(\S+) is a function$/ and not $seen{$1}++ and print "$1\n"'
  57     }
  58     _bodies_of_FUNCTIONS() {
  59         LANG=C type "$@" | perl -ne '/^(\S+) is a function$/ or print'
  60     }
  61     _names_of_VARIABLES() {
  62         # This may screw up if variables contain \n and =
  63         set | perl -ne 's/^(\S+?)=.*/$1/ and print;'
  64     }
  65     _bodies_of_VARIABLES() {
  66         # Crappy typeset -p
  67         for _i in "$@"
  68         do
  69             perl -e 'print @ARGV' "$_i="
  70             eval echo \"\$$_i\" | perl -e '$/=undef; $a=<>; chop($a); print $a' |
  71                 perl -pe 's/[\002-\011\013-\032\\\#\?\`\(\)\{\}\[\]\^\*\<\=\>\~\|\; \"\!\$\&\202-\377]/\\$&/go;'"s/'/\\\'/g; s/[\n]/'\\n'/go;";
  72             echo
  73         done
  74     }
  75     _ignore_HARDCODED() {
  76         # These names cannot be detected
  77         echo '(_|TIMEOUT)'
  78     }
  79     _ignore_READONLY() {
  80         readonly | perl -e '@r = map {
  81                 chomp;
  82                 # sh on UnixWare: readonly TIMEOUT
  83                 # ash: readonly var='val'
  84                 # ksh: var='val'
  85                 s/^(readonly )?([^= ]*)(=.*|)$/$2/ or
  86                 # bash: declare -ar BASH_VERSINFO=([0]="4" [1]="4")
  87                 # zsh: typeset -r var='val'
  88                   s/^\S+\s+\S+\s+(\S[^=]*)(=.*|$)/$1/;
  89                 $_ } <>;
  90             $vars = join "|",map { quotemeta $_ } @r;
  91             print $vars ? "($vars)" : "(,,nO,,VaRs,,)";
  92             '
  93     }
  94     _remove_bad_NAMES() {
  95         # Do not transfer vars and funcs from env_parallel
  96         _ignore_RO="`_ignore_READONLY`"
  97         _ignore_HARD="`_ignore_HARDCODED`"
  98         # Macos-grep does not like long patterns
  99         # Old Solaris grep does not support -E
 100         # Perl Version of:
 101         # grep -Ev '^(...)$' |
 102         perl -ne '/^(
 103                      PARALLEL_ENV|
 104                      PARALLEL_TMP|
 105                      _alias_NAMES|
 106                      _bodies_of_ALIASES|
 107                      _bodies_of_FUNCTIONS|
 108                      _bodies_of_VARIABLES|
 109                      _error_PAR|
 110                      _function_NAMES|
 111                      _get_ignored_VARS|
 112                      _grep_REGEXP|
 113                      _ignore_HARD|
 114                      _ignore_HARDCODED|
 115                      _ignore_READONLY|
 116                      _ignore_RO|
 117                      _ignore_UNDERSCORE|
 118                      _list_alias_BODIES|
 119                      _list_function_BODIES|
 120                      _list_variable_VALUES|
 121                      _make_grep_REGEXP|
 122                      _names_of_ALIASES|
 123                      _names_of_FUNCTIONS|
 124                      _names_of_VARIABLES|
 125                      _names_of_maybe_FUNCTIONS|
 126                      _parallel_exit_CODE|
 127                      _prefix_PARALLEL_ENV|
 128                      _prefix_PARALLEL_ENV|
 129                      _remove_bad_NAMES|
 130                      _remove_readonly|
 131                      _variable_NAMES|
 132                      _warning_PAR|
 133                      _which_PAR)$/x and next;
 134             # Filter names matching --env
 135             /^'"$_grep_REGEXP"'$/ or next;
 136             /^'"$_ignore_UNDERSCORE"'$/ and next;
 137             # Remove readonly variables
 138             /^'"$_ignore_RO"'$/ and next;
 139             /^'"$_ignore_HARD"'$/ and next;
 140             print;'
 141     }
 142     _get_ignored_VARS() {
 143         perl -e '
 144             for(@ARGV){
 145                 $next_is_env and push @envvar, split/,/, $_;
 146                 $next_is_env=/^--env$/;
 147             }
 148             if(grep { /^_$/ } @envvar) {
 149                 if(not open(IN, "<", "$ENV{HOME}/.parallel/ignored_vars")) {
 150                     print STDERR "parallel: Error: ",
 151                     "Run \"parallel --record-env\" in a clean environment first.\n";
 152                 } else {
 153                     chomp(@ignored_vars = <IN>);
 154                 }
 155             }
 156             if($ENV{PARALLEL_IGNORED_NAMES}) {
 157                 push @ignored_vars, split/\s+/, $ENV{PARALLEL_IGNORED_NAMES};
 158                 chomp @ignored_vars;
 159             }
 160             $vars = join "|",map { quotemeta $_ } @ignored_vars;
 161             print $vars ? "($vars)" : "(,,nO,,VaRs,,)";
 162             ' -- "$@"
 163     }
 164
 165     # Get the --env variables if set
 166     # --env _ should be ignored
 167     # and convert  a b c  to (a|b|c)
 168     # If --env not set: Match everything (.*)
 169     _make_grep_REGEXP() {
 170         perl -e '
 171             for(@ARGV){
 172                 /^_$/ and $next_is_env = 0;
 173                 $next_is_env and push @envvar, split/,/, $_;
 174                 $next_is_env = /^--env$/;
 175             }
 176             $vars = join "|",map { quotemeta $_ } @envvar;
 177             print $vars ? "($vars)" : "(.*)";
 178             ' -- "$@"
 179     }
 180     _which_PAR() {
 181         # type returns:
 182         #   ll is an alias for ls -l (in ash)
 183         #   bash is a tracked alias for /bin/bash
 184         #   true is a shell builtin (in bash)
 185         #   myfunc is a function (in bash)
 186         #   myfunc is a shell function (in zsh)
 187         #   which is /usr/bin/which (in sh, bash)
 188         #   which is hashed (/usr/bin/which)
 189         #   gi is aliased to `grep -i' (in bash)
 190         #   aliased to `alias | /usr/bin/which --tty-only --read-alias --show-dot --show-tilde'
 191         # Return 0 if found, 1 otherwise
 192         LANG=C type "$@" |
 193             perl -pe '$exit += (s/ is an alias for .*// ||
 194                                 s/ is aliased to .*// ||
 195                                 s/ is a function// ||
 196                                 s/ is a shell function// ||
 197                                 s/ is a shell builtin// ||
 198                                 s/.* is hashed .(\S+).$/$1/ ||
 199                                 s/.* is (a tracked alias for )?//);
 200                       END { exit not $exit }'
 201     }
 202     _warning_PAR() {
 203         echo "env_parallel: Warning: $@" >&2
 204     }
 205     _error_PAR() {
 206         echo "env_parallel: Error: $@" >&2
 207     }
 208
 209     if _which_PAR parallel >/dev/null; then
 210         true parallel found in path
 211     else
 212         _error_PAR 'parallel must be in $PATH.'
 213         return 255
 214     fi
 215
 216     # Grep regexp for vars given by --env
 217     _grep_REGEXP="`_make_grep_REGEXP \"$@\"`"
 218     unset _make_grep_REGEXP
 219
 220     # Deal with --env _
 221     _ignore_UNDERSCORE="`_get_ignored_VARS \"$@\"`"
 222     unset _get_ignored_VARS
 223
 224     # --record-env
 225     if perl -e 'exit grep { /^--record-env$/ } @ARGV' -- "$@"; then
 226         true skip
 227     else
 228         (_names_of_ALIASES;
 229          _names_of_FUNCTIONS;
 230          _names_of_VARIABLES) |
 231             cat > $HOME/.parallel/ignored_vars
 232         return 0
 233     fi
 234
 235     # --session
 236     if perl -e 'exit grep { /^--session$/ } @ARGV' -- "$@"; then
 237         true skip
 238     else
 239         # Insert ::: between each level of session
 240         # so you can pop off the last ::: at --end-session
 241         PARALLEL_IGNORED_NAMES="`echo \"$PARALLEL_IGNORED_NAMES\";
 242           echo :::;
 243           (_names_of_ALIASES;
 244            _names_of_FUNCTIONS;
 245            _names_of_VARIABLES) | perl -ne '
 246             BEGIN{
 247               map { $ignored_vars{$_}++ }
 248                 split/\s+/, $ENV{PARALLEL_IGNORED_NAMES};
 249             }
 250             chomp;
 251             for(split/\s+/) {
 252               if(not $ignored_vars{$_}) {
 253                 print $_,\"\\n\";
 254               }
 255             }
 256             '`"
 257         export PARALLEL_IGNORED_NAMES
 258         return 0
 259     fi
 260     if perl -e 'exit grep { /^--end.?session$/ } @ARGV' -- "$@"; then
 261         true skip
 262     else
 263         # Pop off last ::: from PARALLEL_IGNORED_NAMES
 264         PARALLEL_IGNORED_NAMES="`perl -e '
 265           $ENV{PARALLEL_IGNORED_NAMES} =~ s/(.*):::.*?$/$1/s;
 266           print $ENV{PARALLEL_IGNORED_NAMES}
 267         '`"
 268         return 0
 269     fi
 270     # Grep alias names
 271     _alias_NAMES="`_names_of_ALIASES | _remove_bad_NAMES | xargs echo`"
 272     _list_alias_BODIES="_bodies_of_ALIASES $_alias_NAMES"
 273     if [ "$_alias_NAMES" = "" ] ; then
 274         # no aliases selected
 275         _list_alias_BODIES="true"
 276     fi
 277     unset _alias_NAMES
 278
 279     # Grep function names
 280     _function_NAMES="`_names_of_FUNCTIONS | _remove_bad_NAMES | xargs echo`"
 281     _list_function_BODIES="_bodies_of_FUNCTIONS $_function_NAMES"
 282     if [ "$_function_NAMES" = "" ] ; then
 283         # no functions selected
 284         _list_function_BODIES="true"
 285     fi
 286     unset _function_NAMES
 287
 288     # Grep variable names
 289     _variable_NAMES="`_names_of_VARIABLES | _remove_bad_NAMES | xargs echo`"
 290     _list_variable_VALUES="_bodies_of_VARIABLES $_variable_NAMES"
 291     if [ "$_variable_NAMES" = "" ] ; then
 292         # no variables selected
 293         _list_variable_VALUES="true"
 294     fi
 295     unset _variable_NAMES
 296
 297     PARALLEL_ENV="`
 298         $_list_alias_BODIES;
 299         $_list_function_BODIES;
 300         $_list_variable_VALUES;
 301     `"
 302     export PARALLEL_ENV
 303     unset _list_alias_BODIES _list_variable_VALUES _list_function_BODIES
 304     unset _bodies_of_ALIASES _bodies_of_VARIABLES _bodies_of_FUNCTIONS
 305     unset _names_of_ALIASES _names_of_VARIABLES _names_of_FUNCTIONS
 306     unset _ignore_HARDCODED _ignore_READONLY _ignore_UNDERSCORE
 307     unset _remove_bad_NAMES _grep_REGEXP
 308     unset _prefix_PARALLEL_ENV
 309     # Test if environment is too big
 310     if `_which_PAR true` >/dev/null 2>/dev/null ; then
 311         parallel "$@"
 312         _parallel_exit_CODE=$?
 313         # Clean up variables/functions
 314         unset PARALLEL_ENV
 315         unset _which_PAR _which_TRUE
 316         unset _warning_PAR _error_PAR
 317         # Unset _parallel_exit_CODE before return
 318         eval "unset _parallel_exit_CODE; return $_parallel_exit_CODE"
 319     else
 320         unset PARALLEL_ENV;
 321         _error_PAR "Your environment is too big."
 322         _error_PAR "You can try 3 different approaches:"
 323         _error_PAR "1. Run 'env_parallel --session' before you set"
 324         _error_PAR "   variables or define functions."
 325         _error_PAR "2. Use --env and only mention the names to copy."
 326         _error_PAR "3. Try running this in a clean environment once:"
 327         _error_PAR "     env_parallel --record-env"
 328         _error_PAR "   And then use '--env _'"
 329         _error_PAR "For details see: man env_parallel"
 330         return 255
 331     fi
 332 }
 333
 334 parset() {
 335     _parset_PARALLEL_PRG=parallel
 336     _parset_main "$@"
 337 }
 338
 339 env_parset() {
 340     _parset_PARALLEL_PRG=env_parallel
 341     _parset_main "$@"
 342 }
 343
 344 _parset_main() {
 345     # If $1 contains ',' or space:
 346     #   Split on , to get the destination variable names
 347     # If $1 is a single destination variable name:
 348     #   Treat it as the name of an array
 349     #
 350     #   # Create array named myvar
 351     #   parset myvar echo ::: {1..10}
 352     #   echo ${myvar[5]}
 353     #
 354     #   # Put output into $var_a $var_b $var_c
 355     #   varnames=(var_a var_b var_c)
 356     #   parset "${varnames[*]}" echo ::: {1..3}
 357     #   echo $var_c
 358     #
 359     #   # Put output into $var_a4 $var_b4 $var_c4
 360     #   parset "var_a4 var_b4 var_c4" echo ::: {1..3}
 361     #   echo $var_c4
 362
 363     _make_TEMP() {
 364         # mktemp does not exist on some OS
 365         perl -e 'use File::Temp qw(tempfile);
 366                  $ENV{"TMPDIR"} ||= "/tmp";
 367                  print((tempfile(DIR=>$ENV{"TMPDIR"}, TEMPLATE => "parXXXXX"))[1])'
 368     }
 369
 370     _parset_NAME="$1"
 371     if [ "$_parset_NAME" = "" ] ; then
 372         echo parset: Error: No destination variable given. >&2
 373         echo parset: Error: Try: >&2
 374         echo parset: Error: ' ' parset myarray echo ::: foo bar >&2
 375         return 255
 376     fi
 377     if [ "$_parset_NAME" = "--help" ] ; then
 378         echo parset: Error: Usage: >&2
 379         echo parset: Error: ' ' parset varname GNU Parallel options and command >&2
 380         echo
 381         parallel --help
 382         return 255
 383     fi
 384     if [ "$_parset_NAME" = "--version" ] ; then
 385         echo "parset 20201022 (GNU parallel `parallel --minversion 1`)"
 386         echo "Copyright (C) 2007-2020 Ole Tange, http://ole.tange.dk and Free Software"
 387         echo "Foundation, Inc."
 388         echo "License GPLv3+: GNU GPL version 3 or later <https://gnu.org/licenses/gpl.html>"
 389         echo "This is free software: you are free to change and redistribute it."
 390         echo "GNU parallel comes with no warranty."
 391         echo
 392         echo "Web site: https://www.gnu.org/software/parallel"
 393         echo
 394         echo "When using programs that use GNU Parallel to process data for publication"
 395         echo "please cite as described in 'parallel --citation'."
 396         echo
 397         return 255
 398     fi
 399     shift
 400     echo "$_parset_NAME" |
 401         perl -ne 'chomp;for (split /[, ]/) {
 402             # Allow: var_32 var[3]
 403             if(not /^[a-zA-Z_][a-zA-Z_0-9]*(\[\d+\])?$/) {
 404                 print STDERR "parset: Error: $_ is an invalid variable name.\n";
 405                 print STDERR "parset: Error: Variable names must be letter followed by letters or digits.\n";
 406                 print STDERR "parset: Error: Usage:\n";
 407                 print STDERR "parset: Error:   parset varname GNU Parallel options and command\n";
 408                 $exitval = 255;
 409             }
 410         }
 411         exit $exitval;
 412         ' || return 255
 413     _exit_FILE=`_make_TEMP`
 414     if perl -e 'exit not grep /,| /, @ARGV' "$_parset_NAME" ; then
 415         # $_parset_NAME contains , or space
 416         # Split on , or space to get the names
 417         eval "`
 418             # Compute results into files
 419             ($_parset_PARALLEL_PRG --files -k "$@"; echo $? > "$_exit_FILE") |
 420                 # var1= cat tmpfile1; rm tmpfile1
 421                 # var2= cat tmpfile2; rm tmpfile2
 422                 parallel -q echo {2}='\`cat {1}; rm {1}\`' :::: - :::+ \`
 423                     echo "$_parset_NAME" |
 424                         perl -pe 's/,/ /g'
 425                          \`
 426             `"
 427     else
 428         # $_parset_NAME does not contain , or space
 429         # => $_parset_NAME is the name of the array to put data into
 430         # Supported in: bash zsh ksh mksh
 431         # Arrays do not work in: sh ash dash
 432         eval "$_parset_NAME=( $(
 433             # Compute results into files. Save exit value
 434             ($_parset_PARALLEL_PRG --files -k "$@"; echo $? > "$_exit_FILE") |
 435                 perl -pe 'chop;$_="\"\`cat $_; rm $_\`\" "'
 436             ) )"
 437     fi
 438     unset _parset_NAME _parset_PARALLEL_PRG _parallel_exit_CODE
 439     # Unset _exit_FILE before return
 440     eval "unset _exit_FILE; return \`cat $_exit_FILE; rm $_exit_FILE\`"
 441 }