shuf: add --repetition to support repetition in output

main(): Process new option.  Replace input_numbers_option_used()
with a local variable.  Re-organize argument processing.
usage(): Describe the new option.
(write_random_numbers): A new function to generate a
permutation of the specified input range with repetition.
(write_random_lines): Likewise for stdin and --echo.
(write_permuted_numbers):  New function refactored from
write_permuted_output().
(write_permuted_lines): Likewise.
* tests/misc/shuf.sh: Add tests for --repetitions option.
* doc/coreutils.texi: Mention --repetitions, add examples.
* TODO: Mention an optimization to avoid needing to
read all of the input into memory with --repetitions.
* NEWS: Mention new shuf option.

1 files changed, 62 insertions, 0 deletions

diff --git a/tests/misc/shuf.sh b/tests/misc/shuf.sh
index 492fd4186..a25a6f8ad 100755
--- a/tests/misc/shuf.sh
+++ b/tests/misc/shuf.sh
@@ -94,4 +94,66 @@ shuf -i0-9 -o A -o B &&
 shuf -i0-9 --random-source A --random-source B &&
   { fail=1; echo "shuf did not detect multiple --random-source usage.">&2 ; }
 
+# Test --repetition option
+
+# --repetition without count should return one line
+shuf --rep -i0-10 > exp || framework_failure_
+c=$(wc -l < exp) || framework_failure_
+test "$c" -eq 1 || { fail=1; echo "--repetition default count is not 1">&2 ; }
+
+# --repetition can output more values than the input range
+shuf --rep -i0-9 -n1000 > exp || framework_failure_
+c=$(wc -l < exp) || framework_failure_
+test "$c" -eq 1000 || { fail=1; echo "--repetition with --count failed">&2 ; }
+
+# Check output values (this is not bullet-proof, but drawing 1000 values
+# between 0 and 9 should produce all values, unless there's a bug in shuf
+# or a very poor random source, or extremely bad luck)
+c=$(sort -nu exp | paste -s -d ' ') || framework_failure_
+test "$c" = "0 1 2 3 4 5 6 7 8 9" ||
+  { fail=1; echo "--repetition produced bad output">&2 ; }
+
+# check --repetition with non-zero low value
+shuf --rep -i222-233 -n2000 > exp || framework_failure_
+c=$(cat exp | sort -nu | paste -s -d ' ') || framework_failure_
+test "$c" = "222 223 224 225 226 227 228 229 230 231 232 233" ||
+ { fail=1; echo "--repetition produced bad output with non-zero low">&2 ; }
+
+# --repetition,-i,count=0 should not fail and produce no output
+shuf --rep -i0-9 -n0 > exp || framework_failure_
+# file size should be zero (no output from shuf)
+test \! -s exp ||
+  { fail=1; echo "--repetition,-i0-9,-n0 produced bad output">&2 ; }
+
+# --repetition with -e, without count, should return one line
+shuf --rep -e A B C D > exp || framework_failure_
+c=$(cat exp | wc -l) || framework_failure_
+test "$c" -eq 1 ||
+  { fail=1; echo "--repetition,-e default count is not 1">&2 ; }
+
+# --repetition with STDIN, without count, should return one line
+printf "A\nB\nC\nD\nE\n" | shuf --rep > exp || framework_failure_
+c=$(wc -l < exp) || framework_failure_
+test "$c" -eq 1 ||
+  { fail=1; echo "--repetition,STDIN default count is not 1">&2 ; }
+
+# --repetition with STDIN,count - can return move values than input lines
+printf "A\nB\nC\nD\nE\n" | shuf --rep -n2000 > exp || framework_failure_
+c=$(wc -l < exp) || framework_failure_
+test "$c" -eq 2000 ||
+  { fail=1; echo "--repetition,STDIN,count failed">&2 ; }
+
+# Check output values (this is not bullet-proof, but drawing 2000 values
+# between A and E should produce all values, unless there's a bug in shuf
+# or a very poor random source, or extremely bad luck)
+c=$(sort -u exp | paste -s -d ' ') || framework_failure_
+test "$c" = "A B C D E" ||
+  { fail=1; echo "--repetition,STDIN,count produced bad output">&2 ; }
+
+# --repetition,stdin,count=0 should not fail and produce no output
+printf "A\nB\nC\nD\nE\n" | shuf --rep -n0 > exp || framework_failure_
+# file size should be zero (no output from shuf)
+test \! -s exp ||
+  { fail=1; echo "--repetition,STDIN,-n0 produced bad output">&2 ; }
+
 Exit $fail