# Mutants that NO test can kill, with the reason each one is unreachable by any
# test that could be written.
#
# THE BAR IS IMPOSSIBILITY, NOT COST. A mutant that a test could kill, where
# nobody has written that test, does not belong here however unappealing the
# test looks. It stays a visible survivor with a GoingsOn task against it. The
# two classes are easy to conflate and the difference is the whole value of the
# number: an exclusion list that also absorbs "not worth it" reports zero while
# real coverage gaps sit underneath it.
#
# The evidence for that bar is in the sibling files. Of docengine's nine
# survivors, four were coverage gaps; of custom-pages' six, five looked
# equivalent at a glance and were not. So: try to kill it first, and add an
# entry only with the reason written out.
#
# The patterns are regexes matched against the WHOLE listing line, `file:line:col`
# included, so `+`, `*` and `|` need escaping. A bare `||` is an empty
# alternation and silently excludes every mutant in the crate.
#
# An entry carries a line number when the description alone would also match a
# mutant that is being CAUGHT -- cargo-mutants names a mutant by its operator and
# its enclosing function and nothing else, so one function with two similar
# operators needs the position to tell them apart. A stale anchor stops matching
# and the mutant returns as a visible survivor, which is the safe direction.
# Re-derive by diffing `cargo mutants --list` with and without this file.

exclude_re = [
    # The two oracle wrappers in `mod oracle`. An oracle asserts what must hold
    # independently of the code under test, so it only speaks when that code is
    # wrong; replacing its body with `()` is invisible to every test that passes.
    # Killing it would mean shipping a broken planner for a test to catch, which
    # is not a test that can exist alongside a correct one.
    #
    # The evidence these two work is the other half of the same run rather than
    # a mutant: all 34 mutants in `MultipartPlan::{new,auto,part_len,part_range}`
    # were caught, and they were caught AT THESE ASSERTIONS.
    "replace oracle::check_plan with \\(\\)",
    "replace oracle::check_auto with \\(\\)",

    # `while bytes_read < part_size` -> `<=` in the part-filling loop. At
    # equality the extra iteration reads into `&mut buf[part_size..]`, an empty
    # slice, and a read into an empty buffer returns `Ok(0)` without touching
    # the file -- which is the loop's own break arm. One wasted call to `read`
    # and an identical buffer, so no observation distinguishes it.
    "src/lib.rs:1072:30: replace < with <= in S3Client::run_multipart_upload",
]