pr: accept -e at the end of a short flag cluster

[m0g3r]

Fixes #13895.

pr -tre failed with a value is required for '--expand-tabs <[CHAR][WIDTH]>' but none was supplied, while GNU prints oi.

-e takes an optional attached argument, which clap cannot express, so recreate_arguments fills in the default before parsing. Its check was ^-e, which only matches an argument that starts with -e, so -e at the end of a cluster like -tre was never given the default and reached clap bare.

Now it matches a cluster of value-less short flags ending in e and appends the default instead of replacing the whole argument, so -tre becomes -tre\t8 and a bare -e still becomes -e\t8. Options that take a value are excluded from the cluster so -se keeps meaning -s e.

I used recreate_arguments rather than num_args(0..=1) as #13900 does for -n: with num_args(0..=1) a bare -e swallows the following operand, so pr -t -e file reports the file name as an invalid -e argument. The second test below guards that.

Test

test_expand_tab_at_end_of_short_flag_cluster fails on main with the error from the issue and passes with this change. The full pr suite passes, 246 tests, and cargo clippy -p uu_pr --all-targets -- -D warnings is clean.

---

Written with AI assistance. I have reviewed the diff and checked the behaviour against GNU pr myself.

[oech3]

Please contribute to clap instead of working around at here.

[github-actions]

Binary size comparison:

Individual binary size comparison VS main (threshold: >=5% AND >=4 KB).

Total size of compared binaries: 149.02 MB (+92 KB, +0.06%)

Significant per-binary changes:
  [           1.10 MB ->    1.18 MB  (+76 KB, +6.74%)
  mkdir       1.10 MB ->    1.17 MB  (+76 KB, +6.76%)
  mknod       1.09 MB ->    1.17 MB  (+76 KB, +6.79%)
  test        1.10 MB ->    1.18 MB  (+76 KB, +6.74%)
  install     1.25 MB ->    1.32 MB  (+72 KB, +5.62%)
  mkfifo      1.09 MB ->    1.16 MB  (+72 KB, +6.45%)
  chmod       1.15 MB ->    1.22 MB  (+68 KB, +5.76%)