gnu_parallel-20211122-bp154.1.21

- update to 20211122:
  * Bug fixes and man page updates.
  * Don't postpone output until $jobslots jobs have been started.
  * Bug fixes and man page updates.

- Update to release 20210922
  * --ctag/--ctagstring colors the tag in different colors for
    each job.
  * You can use unit prefixes (k, m, g, etc) with -n -N -L.

- update to 20210722:
  * --results no longer prints the result to standard output
  * parset supports associative arrays in bash, ksh, zsh.
  * Online HTML is now generated by Sphinx.
  * Bug fixes and man page updates.

- Update to release 20210522
  * --plus includes {%%regexp} and {##regexp}.

- Update to release 20210422
  * Bug fixes and man page updates.

- Update to release 20210322
  * parallel: Partial Reuse licensing support.
  * parallel: Bugfix {0%}/{0#}. Added warning_once().

- Update to release 20210222
  * --filter only run jobs where the filter is true. The filter
    can contain replacement strings and Perl code such as: '{1} <
    {2}+1'.
  * --template takes a text file as a template with replacement
    strings. Then it replaces the replacement strings and saves
    it under a new filename.
  * --plus implements {0%} {0#} replacement string for zero
    padded slot and sequence.

- Update to release 20210122
  * --memsuspend suspends jobs when there is little memory free.
    This way you can run jobs in parallel that add up to more RAM
    than is available.
  * $PARALLEL_ARGHOSTGROUPS and the replacement string {agrp}
    will give the hostgroup given on the argument when using
  - -hostgroup.

- Update to release 20201222
  * --pipe engine changed making --pipe alpha quality.
  * --results -.json outputs results as JSON objects on stdout
    (standard output).
  * --delay 123auto will auto-adjust --delay. If jobs fail due to
    being spawned too quickly, --delay will exponentially
    increase.
  * --plus sets {hgrp} that gives the intersection of the
    hostgroups of the job and the sshlogin that the job is run
    on.

- update to 20201122:
  * Bug fixes and man page updates.

- Update to release 20201022
  * --termseq now works for jobs running remotely.

- Update to release 20200922
  * New CPU detection for GNU/Linux

- Update to release 20200822
  * Bug fixes and man page updates.

- Update to release 20200722
  * Bug fixes and man page updates.

- Update to release 20200522
  * While running a job $PARALLEL_JOBSLOT is the jobslot of the
    job. It is equal to {%} unless the job is being retried. See
    {%} for details.

- Update to release 20200422
  * parsort makes GNU sort run faster for files with more
    than 1M lines.

- Update to release 20200322
  * Bug fixes and man page updates.

- Update to release 20200222
  * --blocktimeout dur - Time out for reading block when using
  - -pipe. If it takes longer than dur to read a full block, use
    the partial block read so far.

- Update to release 20191222
  * No new functionality was introduced.

- Remove /usr/bin/env rpmlint errors:
  * fix shebang in %{_bindir}/parallel
  * add rpmlint filters for env_parallel.* and parset, as these are not intended
    to be executed
- Change URLs to use https instead of http

- update to 20180422:
  * Parse input sources as CSV with --csv. With --pipe only consider
    full CSV-records
  * Time in --bar is printed as 1d02h03m04s
  * Optimization of --tee: It spawns a process less per value

- update to 20180322:
  * niceload -p can now take multiple pids separated by comma
  * --timeout gives a warning when killing processes
  * --embed now uses the same code for all supported shells
  * --delay can now take arguments like 1h12m07s

- Update to new upstream release 20180222
  * --embed makes it possible to embed GNU parallel
    in a shell script.

- update to 20171222:
  * env_parset for ash, dash, ksh, sh, zsh
  * Automatically create hostgroups if argument ends in @sshlogin

- update to 20171022:
  * Initial support for XDG Base Directory $XDG_* environment
    variables

- update to 20170922:
  * Use '--limit myprog' to make a dynamic job limit. Just return 0
    to spawn another job, 1 to not spawn another job, and 2 to kill
    the youngest job.
  * PARALLEL_RSYNC_OPTS and --rsync-opts sets the options for rsync
    (Default: -rlDzR).
  * Download all of humble bundle books in parallel
    https://til.hashrocket.com/posts/ggt1jaes4y-download-all-of-humble-bundle-books-in-parallel

- update to 20170822:
  * --line-buffer no longer uses tempfiles. This is faster and
    makes it possible for a single process to output more data than
    there is free disk space.
  * ship documentation license texts

- update to 20170722:
  * Bug fixes and man page updates

- update to 20170622:
  * The command line can now contain \257 (Macron) or UTF
    characters containing \257.
- includes changes from 20170522:
  * --timeout can now use s=second,m=minute,h=hour,d=day.
  * --dr is now an alias for --dry-run.
- remove donation nag screen, add parallel-remove-nag-screen.patch
  See "man 1 parallel" or "parallel --citation" for the upstream
  request for citation or project donation.

- update to 20170422:
  * --halt supports done in addition to success and fail.
  * parset will set variables in parallel in Bash.
  * Bug fixes and man page updates.
- includes changes from 20170322:
  * --rpl can now take arguments by adding '(regexp)' in the
    replacement string.
  * Bug fixes and man page updates.
- includes changes from 20170222:
  * env_parallel now supports /bin/sh, ash, dash.
  * --tee will pipe all data to all jobs. Used with --pipe/
  - -pipepart and :::. Small example: seq 3333 | parallel --pipe
  - -tee --tag 'grep {} | wc -l' ::: {0..9}
  * If --results {replstr} ends in /: save in dir. Otherwise save
    as file. Example: parallel --results {} echo {} ::: a b c
  * Bug fixes and man page updates.
- includes changes from 20170206alpha:
  * --tee introduced.

- Update to new upstream release 20170122
  * sql now uses a temporary credentials file for MySQL to avoid
    warnings.
  * --sqlmaster and --sqlworker have been reworked.
  * --sqlworker only reads from the DBURL: any command and any
    arguments are ignored.
  * --sqlmaster +DBURL will append commands to the DBURL and not
    erase the table.
  * If --results contain a replacement string, no tree structure
    with args will be generated.
  * --fg --tmux or --fg --tmuxpane will connect to the tmux.

- Update to new upstream release 20161222
  * --results foo.csv will save output as a CSV-file. Can be
    directly imported into R or LibreCalc.
  * When using --pipepart a negative --block-size is not
    interpreted as a block size but as the number of blocks each
    jobslot should have.
  * --sqlmaster/--sqlandworker will append jobs to the DBURL if the
    DBURL is prepended with +.

- Update to new upstream release 20161122
  * --record-env can now be used with env_parallel for bash, ksh,
    pdksh, and zsh.

- Update to version 20161022:
  * First command no longer gets stdin as it causes unpredictable
    issues.
  * Bug fixes and man page updates.
- Changes for version 20160922
  * parcat - cat files or fifos in parallel

- Update to new upstream release 20160822
  * --tmuxpane opens jobs in panes in tmux. Useful if you want to
  monitor progress of less than 100 simultaneous jobs.
  * --linebuffer now treats \r as line ending, too.
  * Perl changes forces use of floats to be given with leading
  zero, so 0.1 and not .1
  * --xapply renamed to --link.

- Update to new upstream release 20160722
  * env_parallel is now ready for wider testing. It is still beta
  quality.
  * env_parallel is heavily modified for all shells and testing has
  been increased.
  * Selectively choosing what to export using --env now works for
  env_parallel (bash, csh, fish, ksh, pdksh, tcsh, zsh).
  * --round-robin now gives more work to a job that processes
  faster instead of same amount to all jobs.
  * --pipepart works on block devices on GNU/Linux.

- Update to new upstream release 20160622
  * $PATH can now be exported using --env PATH. Useful if GNU
  Parallel is not in your path on remote machines.
  * If --block is left out, --pipepart will use a block size that
  will result in 10 jobs per jobslot.

- Update to new upstream release 20160522
  * niceload --net pauses the program if the internet connection
  is overloaded.

- Update to new upstream release 20160422
  * :::+ and ::::+ work like ::: and :::: but links this input source
  to the previous input source in a --xapply fashion. Contrary to
  - -xapply values do not wrap: The shortest input source determines
  the length.
  * --line-buffer --keep-order now outputs continuously from the
  oldest job still running. This is more what you would expect than
  the earlier behaviour where --keep-order had no effect with
  - -line-buffer.
  * env_parallel supports tcsh, csh, pdksh. In fish it now supports
  arrays. In csh/tcsh it now supports variables, aliases, and
  arrays with no special chars. In pdksh it supports aliases,
  functions, variables, and arrays.

- Update to new upstream release 20160322
  * env_parallel is a function that exports the environment
  (functions, aliases, variables, and arrays) to GNU Parallel.
  * niceload --prg now searches for substrings if no process with the
  name is found.