src/env_parallel.ksh

   1 #!/usr/bin/env ksh
   2
   3 # This file must be sourced in ksh:
   4 #
   5 #   source `which env_parallel.ksh`
   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
  31 env_parallel() {
  32     # env_parallel.ksh
  33
  34     _names_of_ALIASES() {
  35         alias | perl -pe 's/=.*//'
  36     }
  37     _bodies_of_ALIASES() {
  38         alias "$@" | perl -pe 's/^/alias /;
  39                      sub warning { print STDERR "env_parallel: Warning: @_\n"; }
  40                      if(/^alias (\S+)=\$.*\\n/) {
  41                          warning("Alias \"$1\" contains newline.");
  42                          warning("Make sure the command has at least one newline after \"$1\".");
  43                          warning("See BUGS in \"man env_parallel\".");
  44                      }'
  45
  46     }
  47     _names_of_maybe_FUNCTIONS() {
  48         true not used
  49     }
  50     _names_of_FUNCTIONS() {
  51         typeset +p -f | perl -pe 's/\(\).*//'
  52     }
  53     _bodies_of_FUNCTIONS() {
  54         functions "$@"
  55     }
  56     _names_of_VARIABLES() {
  57         typeset +p | perl -pe 's/^typeset .. //'
  58     }
  59     _bodies_of_VARIABLES() {
  60         typeset -p "$@"
  61     }
  62     _ignore_HARDCODED() {
  63         # These names cannot be detected
  64         echo '(_|TIMEOUT|IFS)'
  65     }
  66     _ignore_READONLY() {
  67         readonly | perl -e '@r = map {
  68                 chomp;
  69                 # sh on UnixWare: readonly TIMEOUT
  70                 # ash: readonly var='val'
  71                 # ksh: var='val'
  72                 # mksh: PIPESTATUS[0]
  73                 s/^(readonly )?([^=\[ ]*?)(\[\d+\])?(=.*|)$/$2/ or
  74                 # bash: declare -ar BASH_VERSINFO=([0]="4" [1]="4")
  75                 # zsh: typeset -r var='val'
  76                   s/^\S+\s+\S+\s+(\S[^=]*)(=.*|$)/$1/;
  77                 $_ } <>;
  78             $vars = join "|",map { quotemeta $_ } @r;
  79             print $vars ? "($vars)" : "(,,nO,,VaRs,,)";
  80             '
  81     }
  82     _remove_bad_NAMES() {
  83         # Do not transfer vars and funcs from env_parallel
  84         _ignore_RO="`_ignore_READONLY`"
  85         _ignore_HARD="`_ignore_HARDCODED`"
  86         # Macos-grep does not like long patterns
  87         # Old Solaris grep does not support -E
  88         # Perl Version of:
  89         # grep -Ev '^(...)$' |
  90         perl -ne '/^(
  91                      PARALLEL_ENV|
  92                      PARALLEL_TMP|
  93                      _alias_NAMES|
  94                      _bodies_of_ALIASES|
  95                      _bodies_of_FUNCTIONS|
  96                      _bodies_of_VARIABLES|
  97                      _error_PAR|
  98                      _function_NAMES|
  99                      _get_ignored_VARS|
 100                      _grep_REGEXP|
 101                      _ignore_HARD|
 102                      _ignore_HARDCODED|
 103                      _ignore_READONLY|
 104                      _ignore_RO|
 105                      _ignore_UNDERSCORE|
 106                      _list_alias_BODIES|
 107                      _list_function_BODIES|
 108                      _list_variable_VALUES|
 109                      _make_grep_REGEXP|
 110                      _names_of_ALIASES|
 111                      _names_of_FUNCTIONS|
 112                      _names_of_VARIABLES|
 113                      _names_of_maybe_FUNCTIONS|
 114                      _parallel_exit_CODE|
 115                      _prefix_PARALLEL_ENV|
 116                      _prefix_PARALLEL_ENV|
 117                      _remove_bad_NAMES|
 118                      _remove_readonly|
 119                      _variable_NAMES|
 120                      _warning_PAR|
 121                      _which_PAR)$/x and next;
 122             # Filter names matching --env
 123             /^'"$_grep_REGEXP"'$/ or next;
 124             /^'"$_ignore_UNDERSCORE"'$/ and next;
 125             # Remove readonly variables
 126             /^'"$_ignore_RO"'$/ and next;
 127             /^'"$_ignore_HARD"'$/ and next;
 128             print;'
 129     }
 130     _get_ignored_VARS() {
 131         perl -e '
 132             for(@ARGV){
 133                 $next_is_env and push @envvar, split/,/, $_;
 134                 $next_is_env=/^--env$/;
 135             }
 136             if(grep { /^_$/ } @envvar) {
 137                 if(not open(IN, "<", "$ENV{HOME}/.parallel/ignored_vars")) {
 138                     print STDERR "parallel: Error: ",
 139                     "Run \"parallel --record-env\" in a clean environment first.\n";
 140                 } else {
 141                     chomp(@ignored_vars = <IN>);
 142                 }
 143             }
 144             if($ENV{PARALLEL_IGNORED_NAMES}) {
 145                 push @ignored_vars, split/\s+/, $ENV{PARALLEL_IGNORED_NAMES};
 146                 chomp @ignored_vars;
 147             }
 148             $vars = join "|",map { quotemeta $_ } @ignored_vars;
 149             print $vars ? "($vars)" : "(,,nO,,VaRs,,)";
 150             ' -- "$@"
 151     }
 152
 153     # Get the --env variables if set
 154     # --env _ should be ignored
 155     # and convert  a b c  to (a|b|c)
 156     # If --env not set: Match everything (.*)
 157     _make_grep_REGEXP() {
 158         perl -e '
 159             for(@ARGV){
 160                 /^_$/ and $next_is_env = 0;
 161                 $next_is_env and push @envvar, split/,/, $_;
 162                 $next_is_env = /^--env$/;
 163             }
 164             $vars = join "|",map { quotemeta $_ } @envvar;
 165             print $vars ? "($vars)" : "(.*)";
 166             ' -- "$@"
 167     }
 168     _which_PAR() {
 169         # type returns:
 170         #   ll is an alias for ls -l (in ash)
 171         #   bash is a tracked alias for /bin/bash
 172         #   true is a shell builtin (in bash)
 173         #   myfunc is a function (in bash)
 174         #   myfunc is a shell function (in zsh)
 175         #   which is /usr/bin/which (in sh, bash)
 176         #   which is hashed (/usr/bin/which)
 177         #   gi is aliased to `grep -i' (in bash)
 178         #   aliased to `alias | /usr/bin/which --tty-only --read-alias --show-dot --show-tilde'
 179         # Return 0 if found, 1 otherwise
 180         LANG=C type "$@" |
 181             perl -pe '$exit += (s/ is an alias for .*// ||
 182                                 s/ is aliased to .*// ||
 183                                 s/ is a function// ||
 184                                 s/ is a shell function// ||
 185                                 s/ is a shell builtin// ||
 186                                 s/.* is hashed .(\S+).$/$1/ ||
 187                                 s/.* is (a tracked alias for )?//);
 188                       END { exit not $exit }'
 189     }
 190     _warning_PAR() {
 191         echo "env_parallel: Warning: $@" >&2
 192     }
 193     _error_PAR() {
 194         echo "env_parallel: Error: $@" >&2
 195     }
 196
 197     if _which_PAR parallel >/dev/null; then
 198         true parallel found in path
 199     else
 200         _error_PAR 'parallel must be in $PATH.'
 201         return 255
 202     fi
 203
 204     # Grep regexp for vars given by --env
 205     _grep_REGEXP="`_make_grep_REGEXP \"$@\"`"
 206     unset _make_grep_REGEXP
 207
 208     # Deal with --env _
 209     _ignore_UNDERSCORE="`_get_ignored_VARS \"$@\"`"
 210     unset _get_ignored_VARS
 211
 212     # --record-env
 213     if perl -e 'exit grep { /^--record-env$/ } @ARGV' -- "$@"; then
 214         true skip
 215     else
 216         (_names_of_ALIASES;
 217          _names_of_FUNCTIONS;
 218          _names_of_VARIABLES) |
 219             cat > "$HOME"/.parallel/ignored_vars
 220         return 0
 221     fi
 222
 223     # --session
 224     if perl -e 'exit grep { /^--session$/ } @ARGV' -- "$@"; then
 225         true skip
 226     else
 227         # Insert ::: between each level of session
 228         # so you can pop off the last ::: at --end-session
 229         PARALLEL_IGNORED_NAMES="`echo \"$PARALLEL_IGNORED_NAMES\";
 230           echo :::;
 231           (_names_of_ALIASES;
 232            _names_of_FUNCTIONS;
 233            _names_of_VARIABLES) | perl -ne '
 234             BEGIN{
 235               map { $ignored_vars{$_}++ }
 236                 split/\s+/, $ENV{PARALLEL_IGNORED_NAMES};
 237             }
 238             chomp;
 239             for(split/\s+/) {
 240               if(not $ignored_vars{$_}) {
 241                 print $_,\"\\n\";
 242               }
 243             }
 244             '`"
 245         export PARALLEL_IGNORED_NAMES
 246         return 0
 247     fi
 248     if perl -e 'exit grep { /^--end.?session$/ } @ARGV' -- "$@"; then
 249         true skip
 250     else
 251         # Pop off last ::: from PARALLEL_IGNORED_NAMES
 252         PARALLEL_IGNORED_NAMES="`perl -e '
 253           $ENV{PARALLEL_IGNORED_NAMES} =~ s/(.*):::.*?$/$1/s;
 254           print $ENV{PARALLEL_IGNORED_NAMES}
 255         '`"
 256         return 0
 257     fi
 258     # Grep alias names
 259     _alias_NAMES="`_names_of_ALIASES | _remove_bad_NAMES | xargs echo`"
 260     _list_alias_BODIES="_bodies_of_ALIASES $_alias_NAMES"
 261     if [ "$_alias_NAMES" = "" ] ; then
 262         # no aliases selected
 263         _list_alias_BODIES="true"
 264     fi
 265     unset _alias_NAMES
 266
 267     # Grep function names
 268     _function_NAMES="`_names_of_FUNCTIONS | _remove_bad_NAMES | xargs echo`"
 269     _list_function_BODIES="_bodies_of_FUNCTIONS $_function_NAMES"
 270     if [ "$_function_NAMES" = "" ] ; then
 271         # no functions selected
 272         _list_function_BODIES="true"
 273     fi
 274     unset _function_NAMES
 275
 276     # Grep variable names
 277     _variable_NAMES="`_names_of_VARIABLES | _remove_bad_NAMES | xargs echo`"
 278     _list_variable_VALUES="_bodies_of_VARIABLES $_variable_NAMES"
 279     if [ "$_variable_NAMES" = "" ] ; then
 280         # no variables selected
 281         _list_variable_VALUES="true"
 282     fi
 283     unset _variable_NAMES
 284
 285     PARALLEL_ENV="`
 286         $_list_alias_BODIES;
 287         $_list_function_BODIES;
 288         $_list_variable_VALUES;
 289     `"
 290     export PARALLEL_ENV
 291     unset _list_alias_BODIES _list_variable_VALUES _list_function_BODIES
 292     unset _bodies_of_ALIASES _bodies_of_VARIABLES _bodies_of_FUNCTIONS
 293     unset _names_of_ALIASES _names_of_VARIABLES _names_of_FUNCTIONS
 294     unset _ignore_HARDCODED _ignore_READONLY _ignore_UNDERSCORE
 295     unset _remove_bad_NAMES _grep_REGEXP
 296     unset _prefix_PARALLEL_ENV
 297     # Test if environment is too big
 298     if `_which_PAR true` >/dev/null 2>/dev/null ; then
 299         parallel "$@"
 300         _parallel_exit_CODE=$?
 301         # Clean up variables/functions
 302         unset PARALLEL_ENV
 303         unset _which_PAR _which_TRUE
 304         unset _warning_PAR _error_PAR
 305         # Unset _parallel_exit_CODE before return
 306         eval "unset _parallel_exit_CODE; return $_parallel_exit_CODE"
 307     else
 308         unset PARALLEL_ENV;
 309         _error_PAR "Your environment is too big."
 310         _error_PAR "You can try 3 different approaches:"
 311         _error_PAR "1. Run 'env_parallel --session' before you set"
 312         _error_PAR "   variables or define functions."
 313         _error_PAR "2. Use --env and only mention the names to copy."
 314         _error_PAR "3. Try running this in a clean environment once:"
 315         _error_PAR "     env_parallel --record-env"
 316         _error_PAR "   And then use '--env _'"
 317         _error_PAR "For details see: man env_parallel"
 318         return 255
 319     fi
 320 }
 321
 322 parset() {
 323     _parset_PARALLEL_PRG=parallel
 324     _parset_main "$@"
 325 }
 326
 327 env_parset() {
 328     _parset_PARALLEL_PRG=env_parallel
 329     _parset_main "$@"
 330 }
 331
 332 _parset_main() {
 333     # If $1 contains ',' or space:
 334     #   Split on , to get the destination variable names
 335     # If $1 is a single destination variable name:
 336     #   Treat it as the name of an array
 337     #
 338     #   # Create array named myvar
 339     #   parset myvar echo ::: {1..10}
 340     #   echo ${myvar[5]}
 341     #
 342     #   # Put output into $var_a $var_b $var_c
 343     #   varnames=(var_a var_b var_c)
 344     #   parset "${varnames[*]}" echo ::: {1..3}
 345     #   echo $var_c
 346     #
 347     #   # Put output into $var_a4 $var_b4 $var_c4
 348     #   parset "var_a4 var_b4 var_c4" echo ::: {1..3}
 349     #   echo $var_c4
 350
 351     _parset_NAME="$1"
 352     if [ "$_parset_NAME" = "" ] ; then
 353         echo parset: Error: No destination variable given. >&2
 354         echo parset: Error: Try: >&2
 355         echo parset: Error: ' ' parset myarray echo ::: foo bar >&2
 356         return 255
 357     fi
 358     if [ "$_parset_NAME" = "--help" ] ; then
 359         echo parset: Error: Usage: >&2
 360         echo parset: Error: ' ' parset varname GNU Parallel options and command >&2
 361         echo
 362         parallel --help
 363         return 255
 364     fi
 365     if [ "$_parset_NAME" = "--version" ] ; then
 366         echo "parset 20210922 (GNU parallel `parallel --minversion 1`)"
 367         echo "Copyright (C) 2007-2021 Ole Tange, http://ole.tange.dk and Free Software"
 368         echo "Foundation, Inc."
 369         echo "License GPLv3+: GNU GPL version 3 or later <https://gnu.org/licenses/gpl.html>"
 370         echo "This is free software: you are free to change and redistribute it."
 371         echo "GNU parallel comes with no warranty."
 372         echo
 373         echo "Web site: https://www.gnu.org/software/parallel"
 374         echo
 375         echo "When using programs that use GNU Parallel to process data for publication"
 376         echo "please cite as described in 'parallel --citation'."
 377         echo
 378         return 255
 379     fi
 380     shift
 381
 382     # Bash: declare -A myassoc=( )
 383     # Zsh: typeset -A myassoc=( )
 384     # Ksh: typeset -A myassoc=( )
 385     if (typeset -p "$_parset_NAME" 2>/dev/null; echo) |
 386             perl -ne 'exit not (/^declare[^=]+-A|^typeset[^=]+-A/)' ; then
 387         # This is an associative array
 388         eval "`$_parset_PARALLEL_PRG -k --parset assoc,"$_parset_NAME" "$@"`"
 389         # The eval returns the function!
 390     else
 391         # This is a normal array or a list of variable names
 392         eval "`$_parset_PARALLEL_PRG -k --parset var,"$_parset_NAME" "$@"`"
 393         # The eval returns the function!
 394     fi
 395 }