Two color-handling gaps vs GNU: the 'mt' capability in GREP_COLORS (which
sets both the selected- and context-match colors) was ignored, and the
deprecated GREP_COLOR variable produced no warning. Handle 'mt', and emit
GNU's 'GREP_COLOR=... is deprecated; use GREP_COLORS=mt=...' warning when
color output is actually produced. Fixes the GNU testsuite 'color-colors'
test.
A back-reference to a non-existent group, e.g. (.)\2, makes oniguruma
fail with 'invalid backref number/name'. GNU words this per engine:
'reference to non-existent subpattern' under -P (PCRE2) and 'Invalid
back reference' for basic/extended (gnulib regex). Translate
ONIGERR_INVALID_BACKREF accordingly (gnu_error_message now takes the
regex mode). Fixes the GNU testsuite 'pcre-wx-backref' test.
Also drop pipe_in() from the compile-error tests added in the previous
commits: those patterns are rejected before stdin is read, so feeding
input raced with the child exiting and intermittently panicked the test
harness with a broken pipe under parallel execution.
A pathological pattern such as -P '((a+)*)+$' makes oniguruma exceed its
match retry limit. The onig crate's search_with_encoding/match_with_encoding
panic on that error, so uu_grep aborted with a Rust panic. Switch to the
fallible *_with_param variants and propagate the error: match_line/is_match
now return io::Result, the failure is mapped to GNU's wording ('exceeded
PCRE's backtracking limit') and surfaces as a normal exit-code-2 diagnostic.
Fixes the GNU testsuite 'pcre-abort' test.
A reversed range like [b-a] makes oniguruma fail with 'empty range in
char class', which uu_grep wrapped as 'invalid pattern "[b-a]": ...'.
GNU grep instead prints the bare POSIX diagnostic 'Invalid range end'
and exits 2. Translate the oniguruma error code
(ONIGERR_EMPTY_RANGE_IN_CHAR_CLASS) to GNU's wording, leaving other
compile errors to fall back to oniguruma's text. Fixes the GNU testsuite
'reversed-range-endpoints' test.
GNU grep flags a bracket expression of the form [:name:] (an almost
certain misspelling of [[:name:]]) with a dedicated diagnostic and exit
code 2, whereas oniguruma silently treats it as the character set
{':','n','a','m','e'}. Port GNU's colon_warning_state logic from
parse_bracket_exp (gnulib dfa.c) so basic/extended patterns produce the
same error. Fixes the GNU testsuite 'warn-char-classes' test.
GNU grep tolerates options given more than once (boolean flags are
idempotent, value options take the last occurrence), but clap rejected
them with "cannot be used multiple times" (exit 2). Set
args_override_self(true) so clap replaces rather than errors; args with
ArgAction::Append (-e/-f/--include/--exclude) still accumulate.
Found by the differential fuzzer (e.g. `grep -e .* -E -n -o -n -x`).
Add a regression test covering repeated booleans, repeated value options
(last wins), and that -e still accumulates patterns.
Sylvestre Ledru
Leonard Hecker