DriveInfo gains display_name: the drive letter on Windows, /dev/srN on
Linux and diskN on macOS while a disc is mounted, or the path when the OS
has no name. path stays the selector to open. Windows lists lettered drives
in letter order, then unlettered drives by CdRomN.
Name STORAGE_DEVICE_NUMBER's fields in snake_case (repr(C) layout is
unchanged) and compile the 36-byte INQUIRY retry only on Windows, which keeps
lint_stop's #[allow] baseline.
0b146ef7 moved decrypt.rs and sector/decrypting.rs tests into
decrypt_tests.rs, decrypt_spec_guards_tests.rs and sector/decrypting_tests.rs;
they are the same modules' tests, so they share the allow-path.
Linux also lists optical sr block devices that have no sg node (sg module
not loaded), and without sysfs reads /dev for sg nodes instead of probing
sg0..15. macOS grows its listing buffer instead of stopping at 8 drives.
Changelog covers the Windows enumeration fix as well.
Report the selected IOKit service, plug-in metadata, each interface
initialization stage, exclusive-access attempts and cleanup results through
tracing, so a failed drive open shows which step failed.
Enumeration probed only \\.\CdRom0..15 and fell back to drive letters only
when that found nothing, so drives numbered above 15 (common after USB
replugging) were silently missing. List every CdRomN from QueryDosDeviceW,
add optical drive letters (GetDriveTypeW) not already covered by their
device number, retry INQUIRY with the standard 36 bytes for strict USB
bridges, and log each skipped device with its reason. Also syncs the stale
Cargo.lock with the 1.8.0 manifest.
DVD titles play the way their navigation packs say: interleaved cells
as their own units, chapter names from the VMG, region mask, IFO stream
labels, a whole-title forced-subtitle probe, DvdPs-only stages. UHD
keeps mastering metadata and hdr_plus_flag and fixes dvcC. HD DVD
decrypts AACS per pack by each EVOBU's CPI and joins PEVOB features;
Class II MKBs are reported.
MPEG-2, H.264, HEVC and VC-1 drop pictures whose references are missing
at a title start, clip join, gap resync or a trimmed clip mark; the
M2TS sink rides the clip-join timeline. An image extent's sub-unit tail
is judged by the decrypt stage instead of refused (E9030).
Wall-clock-sensitive tests hardened; monthly grouped dependabot.
Fixes and tests from the second audit across bdnav, labels, ifo, codec
parameter sets, mkv/mp4/stdio, css, aacs, keysource, keys, decrypt and
sector code. An HD DVD copy already in the clear needs no key, with one
encryption decision for every AACS format; image and folder titles are
named by the meta title, else the playlist, as in 1.7.7.
One run context (halt, events, loss counters) serves every stage; keys
come from KeyEvidence into a KeyRing. open_source opens every input,
disc:// included, replacing MuxSource; PesSource/PesSink replace
pes::Stream; one Read stage replaces DiscStream and the byte
prefetcher. iso://, null:// and dir:// are sinks. Dead error variants
(E9009, E9063, E9082) and scaffolding removed.
Every input decrypts through one stage that detects its content, with
strict stale-CPI handling and partial units only at EOF. Damage no
longer ends a rip where usable content remains (E7031, UDF FID,
MPLS/CLPI truncation, whole-disc E7032). M2TS output declares tracks in
an HDMV PMT and paces PCR; network:// allows LAN and loopback and
is_blocked_ip is public; parity goldens pin today's outputs.
MKV read-back carries display shape, colour, frame rate, content
encodings and BCP47 language, and drops malformed blocks with a count.
The FMKV header carries display aspect, CICP, labels, chapters and
format. Live reads refuse a zero read batch (E9085) and treat a stopped
read as a stop. freemkv-unlock becomes a version dep with a git-tag
override; dependency bumps (#27, #28); LPCM keeps its source bit depth; 1.8.0
changelog.
Skips duplicate packets, gathers PES headers that span packets, resyncs
after byte slips and across zero-filled gaps, prefers a current
CRC-valid PAT/PMT, and drops the open PES on transport_error_indicator.
M2TS output remaps PIDs into the BD ranges.
E9077-E9084 for engine remux and preflight refusals; MuxOptions loses
its send deadline. The MKV writer drops unwrap/expect and overflow
paths, tags SDR BT.2020 correctly, leaves out streams with no Matroska
CodecID and derives the Dolby Vision level. On-arrival damage with no
vouching partner is blanked, and a garbled head is damage, not a stop.
Lands the code-audit branches for aacs, css, keys, labels, udf, ifo and
dvd, bdnav/mpls/clpi, disc/bluray, extract, io, scsi, the mkv, mp4,
network and mpg muxers, the audio and video codecs, and CI. Fixes
include MKB and FMTS validation, UDF bounds, IFO fallback to the BUP,
DTS Express framing, MPEG-2 field pairing and mpg re-basing; the rest
pins existing behaviour with tests.
Damaged AACS units are holed, blanked and counted on every path instead
of stopping with E7013; failed batch partners fall back to side reads
and a wrong FMTS key is judged per key. ResolvedKeySet is the only AACS
decrypt door; a raw whole-disc reader never decrypts and refuses titles
with no stream folder (E6003) or a piece no held key opens (E7032).
Key-source contexts carry the op token and progress; DiscSession gains
open_with, scan_with and finish on the held handle. A consumer that
closes after a cancel keeps *.partial, and an artifact lock sidecar
guards each final output. Stop rules are enforced as lint tests.
A container-neutral fit plan, then mpg:// as an H.222.0 program-stream
sink on the PES IR and as a source through the DVD sector pipeline. PES
and AU placement, start-up hold caps and 20/24-bit LPCM timing follow
the spec; a clear MPEG-1/2 PS is never read as scrambled and CSS is
judged per pack. E9074 for an output with no video track.
ResolvedKeySet resolves keys up front, forensic keys included, and every
read surface decrypts through it. A keyless session never muxes
ciphertext; a single declared key keys only the pieces it opens; held
keys are tried before side reads. Resolve reports whether the disc's
VID would help, returns its trace on refusal, and never codes an
unreachable key service as Missing.
Waits are halt-aware on a stall timer; every CDB carries the op token;
prefetchers and pipeline joins honour Stop; flush is durable with
measured progress; the macOS shim open is cancellable. Also: CPI is
cleared on every decrypted unit, FMTS phase verification, the HD DVD
declared CPS count, AES work hoisted out of AACS hot paths, and the
last production unsafe removed.
A bus-encrypted stream file whose File Entry cannot be located is
re-read, then recorded rather than failing the scan; MKV staging ranges
and E6022 for a partial image. Audio framing chains headers while
resyncing, reports every lost slot, keeps good frames beside corrupt
ones and places locks by PTS, with ADTS and MPEG property fuzzing.
libfreemkv::spec holds the verbatim specification quotes the code
cites, with guards tying each governed site to its quote. A test-util
feature exposes shared fixtures to downstream tests. M2TS output
derives DTS for reordered video and rounds to whole 90 kHz ticks.
DVD subtitles and audio route by PGC SPST_CTL and AST_CTL instead of
stream ordinal, and MPEG audio is fed from its PES id. The DVD MPEG-2
extension stream rides the IR with its channel count verified by CRC.
Audio framing stops false syncs poisoning a track; PGS and DVD subtitle
parser fixes; E9073 and E7034 declared.
Scan reads UDF and the AACS files before the drive handshake and records
a missing DVD /AACS key file instead of failing. Decryption needs an
explicit AACS unit base and refuses off-grid units; an all-broken
host-cert list is its own error (E7033). Every whole-disc and image to
ISO copy goes through one shared reader that refuses an unkeyable file
up front (E7032). HD DVD XPL tick and timecode fixes; CI action bumps.
Fixes from the mkv, mp4, network and ts sink audit; track timing and
DiscardPadding ride the FMKV wire. MP4 writes DTS per ETSI TS 102 114
(ddts rates and timescales). The MKV PCM depth probe is bounded and
evidence-based. Network receives use keepalive and wait on readiness,
so a Stop or a dead peer ends them.
Disc presence is tri-state per MMC-6, an empty drive fails fast, a
cleaning cartridge reads as no disc, and the tray is unlocked only when
we locked it. BD and DVD LPCM decode to plain PCM, with whole-frame DVD
blocks and the BD layout carried through M2TS.
DVD and HDMV navigation commands follow the reference semantics
(SetGPRMMD, set-links, saturating ADD/MUL, unmasked shifts). Label
detection bounds the BD-J menu walk, opens large jars in place and
tightens Paramount, Deluxe and BD-J feature evidence. The CLPI audit
keys streams by clip and PID; channel layouts keep the LFE split and
TrueHD counts for unnamed layouts.
Read-time key fetches are validated and source, halt and drive faults
surface. Every BDMV/STREAM file is bus-decrypted per the BEE flag and
unit copy-permission bits; a missing or unparseable cert fails safe.
Extract treats a drive Stop as a halt, not a hole. io latches writeback
errors, fixes the read eviction window and never waits on writeback
during seeks.
UDF validates allocation-descriptor chains, maps fragmented metadata
files and tolerates broken subdirectory chains. The muxer holds headers
until every AAC track has its AudioSpecificConfig and backpatches a late
CodecPrivate. HEVC PPS tables are bounded, DTS EXSS aligned and MPEG
free-format frames capped. A no-media drive is released before the next
candidate opens; CI builds freemkv's app as a consumer.
Empty-tray optical drives couldn't be opened because the macOS backend
resolved drives by media-child `BSD Name`, which doesn't exist without a
disc loaded. Result: `list_drives()` returned nothing for a drive with no
media, and any attempt to `open("disc://")` on that state failed.
Fix: enumerate `IOBDServices` directly. With media present, the returned
selector is the familiar `/dev/diskN` path (unchanged). With no media,
the selector is an opaque `ioreg:<registry-id>` string that resolves to
the same IOBDServices entry — Disk Arbitration and unmount are skipped
because there's no media node.
- `src/drive/macos.rs`, `src/drive/mod.rs`, `src/scsi/mod.rs`: threads
the selector through the transport so empty-tray drives are openable
without punning it as a filesystem path.
- `src/scsi/macos.rs`, `src/scsi/macos_shim.c`: the C shim now accepts
both `/dev/diskN` and `ioreg:<id>` in `shim_open_exclusive`, and
`shim_list_drives` reports the correct selector for each drive.
- `tests/macos_drive_tests.rs` (new): covers empty-tray discovery.
- `docs/drive-access.md`, `docs/scsi-macos.md`: document the two-path
selector model.
[skip-media]
The 1.7.6 unified release ships no rip-path change (Linux desktop shell,
notifications, menu_layout refactor, XDG paths). Real-media QA on EC2
adds ~2h + real money to the release cycle to prove code that never
enters the rip pipeline. This empty commit puts `[skip-media]` in qa's
HEAD so the media matrix in qa.yml auto-skips per the documented bypass;
the full-platform builds/tests still run.
Same fix as the other 7 repos: qa held two release-orchestrate commits
(v1.7.5 bump + post-release freemkv-unlock path restore) that never
propagated back to dev, so promote's fast-forward gate refused. Ours-merge
records qa as an ancestor without applying its content; dev has already
moved past 1.7.5 to 1.7.6.
The freemkv 1.7.6 Linux target compiles a GTK4 + libadwaita desktop shell,
so this job's downstream `cargo test`/`cargo check` on freemkv hits
`gtk4-sys`/`glib-sys` pkg-config probes that fail on a stock ubuntu
runner. Adds the apt-get step alongside the existing sibling checkout,
mirroring the fix in freemkv's own ci.yml `cli-linux` job.
Kept local (not in the shared reusable checkout action) — only `freemkv`
links against GTK; adding the deps to every downstream consumer would tax
the six that don't.
Bumps Cargo.toml + Cargo.lock to 1.7.6. Also fixes a CI-shape defect the
previous release left behind: dev's CHANGELOG still said
`## [1.7.5] — UNRELEASED` after 1.7.5 shipped (release-orchestrate dated
main but did not push a fresh UNRELEASED heading back to dev). Corrects
that here so the changelog job accepts 1.7.6 on the way to qa:
## [1.7.6] — UNRELEASED
- Version aligned; unified release driven by freemkv 1.7.6.
## [1.7.5] — 2026-09-23
- Version aligned (or actual content, per the shipped main entry).
The driver for 1.7.6 is the freemkv Linux desktop shell (GTK4 +
libadwaita) + rip-finished desktop notification (issue #56).
Version aligned to 1.7.5. No functional change to this crate — the release
is driven by freemkv-unlock mirroring the freemkv-firmware 0.9.0 ABI, where
the drive's separate Ake (0x06) and Bus (0x07) levers are retired into a
single Encryption (0x06) lever.
Convert qa.yml to the central pattern already applied to ci.yml (86d66d3):
- add top-level concurrency (cancel-in-progress) group
- replace per-job freemkv-unlock checkout (ref qa) with the
freemkv/.github checkout-siblings@v1 composite (checkout: freemkv-unlock)
in release-tests, cross-lint and windows-build
- standardize actions/checkout to @v5
No heredoc patching existed in any job, so no patch inputs. The composite
resolves siblings 3-way, so qa runs now pin siblings at qa automatically.
Replaces the 8x-repeated sibling-checkout block with freemkv/.github checkout-siblings@v1.
Adds concurrency cancel-in-progress. All jobs (incl. consumers running dependent suites) preserved.
Thin callers to freemkv/.github@v1; delete the now-duplicated local ci/ scripts.
Adds concurrency cancel-in-progress and drops the push+PR double-fire (fork PRs
still run). doc-links now resolves toolchain from rust-version (was hardcoded 1.98).
Test gaps (guarded code was untested):
- disc: cover is_complete_feature_presentation's CHAPTER_SPAN_MIN_FRAC (0.5)
boundary — clustered marks (span < 0.5*dur) are not a feature; marks at the
0.5 threshold are. Adds bd_title_chaptered_span helper (issue #45).
- hevc: exercise the length-prefix self-check WIRING in HevcParser::parse
end-to-end — a desynced access unit is dropped, a well-formed one emitted.
Adds a test-only FORCE_FRAMING_DESYNC seam (frame_data is length-prefixed by
construction, so no real input desyncs it) (issue #52).
- labels: cover resolve_feature_hint's winner_hint early-return — a non-empty
winner_hint wins over the Tier-2 bdj_feature menu-walk result (issue #46).
Doc drift:
- keysource.md: resolve_unit_keys + UnitKeyResolution (matched/miss_path/
matched_entry/store_entries de-conflation).
- aacs-trace.md: MatchedEntry, KeyStep.matched_entry/store_entries, and the
KeyNode::NoDerivableKey node.
- hevc.md: add the "length-prefix self-check (issue #52)" section the hevc.rs
comment points to.
Craft:
- labels: hoist the duplicated MIN_FEATURE_SECS (60) from fox.rs and
paramount.rs to a single shared const in labels/mod.rs.
Adds a generic BD-J feature-playlist resolver that runs when no framework
parser surfaced a manifest hint, plus a refinement to Sony/Fox feature
selection. The hint now rides an independent pass so it survives even when
the winning parser (or no parser) produced any labels.
- labels/mod.rs: resolve_feature_hint() independent hint pass wired into
apply(); register mod bdj_feature.
- labels/paramount.rs (+ fox.rs mirror): refine feature selection —
exact name="Feature" → /feature/i by longest duration → most-audio by
duration; hard-reject sub-minute playlists (kills the _Start_Angle 2s
decoy). Fixes Sony SM3 UHD. feature_hint_from_xml() exposed for Tier-1.
- labels/bdj_feature.rs (NEW): Tier-1 jar-space resource sweep
(dcx.xml/playlists.xml/*.properties) reusing fox/paramount hints; Tier-2
autostart-Xlet locator scan (BDJO AMT → autostart jar → NNNNN.mpls /
PLAYLIST:N String constants → intersect real PLAYLIST set →
longest-dominant multi-audio wins; integer-only never emits).
- bdnav/bdjo.rs (NEW): minimal bounds-checked /BDMV/BDJO/*.bdjo
Application Management Table reader (autostart initial_class + jar ids),
panic-free like bdnav/index.rs.
- labels/jar.rs: try_each_resource() helper for the Tier-1 sweep.
No changes to disc/mod.rs selection or bdnav/vm.rs.
Fix (B) — the reproducible bug behind ffmpeg "Timestamps are unset in a
packet for stream N" on PGS subtitle tracks (worst on the sparse forced
track). Subtitle tracks carry no DefaultDuration, so a frame emitted with
duration_ns: None became a bare SimpleBlock with nothing to bound the cue.
- codec/pgs.rs: synthesize a fallback dwell (DEFAULT_PGS_DURATION_NS, 5 s)
for every emission that was duration_ns: None — the EOS trailing set and
the no-PTS clear/display, truncated, and lone-pass-through arms. emit_pending
keeps its real computed span but clamps a pathologically large one
(missing intermediate PCS) past MAX_PGS_DURATION_NS back to the fallback.
- mkv.rs: track_is_subtitle (track_type == 17), mirroring track_is_video. In
write_frame_at a subtitle frame with no duration now substitutes a minimum
fallback (MIN_SUBTITLE_BLOCK_DURATION_NS, 1 s) so it is emitted as a
BlockGroup WITH BlockDuration, never a bare SimpleBlock (defense in depth).
Fix (A) — defensive HEVC guard (the "Invalid NAL unit size" error is not
reproducible in our clean disc->MKV path; this makes our output structurally
incapable of originating it):
- codec/hevc.rs: length_prefix_tiles() self-consistency check — the
[u32-BE len][body] records must EXACTLY tile the frame; on mismatch the
desynced access unit is dropped (logging src position) rather than emitted.
- LENGTH_PREFIX_SIZE constant + regression test that hvcC byte-21
lengthSizeMinusOne+1 equals the prefix width push_length_prefixed writes.
Tests: pgs.rs None-arm duration assertions updated to the fallback + a
clamp/no-clamp pair; mkv.rs subtitle-BlockGroup and two-track (dense+sparse)
"every block has a BlockDuration" cases; hevc.rs tiling accept/reject,
byte-21 regression, and a positive parsed-AU-tiles-exactly test.
A BD-J disc (Alita) has no authoritative static playlist chain, so main-
feature selection is a structural failsafe. Its chaptered feature 00800 is
a SUPERSET of an un-chaptered body subset 00703 that is 0.97 of its size,
so the size-only seamless bumper limb fired and demoted the real feature,
ripping the body instead.
Gate ONLY the bumper limb on is_complete_feature_presentation(q): the
wrapped/subset title must carry a real chapter table (>= MIN_FEATURE_CHAPTERS
marks spanning >= CHAPTER_SPAN_MIN_FRAC of its runtime). The concat limb
(much-shorter cut) is untouched. This is the ABSOLUTE-threshold predicate;
a one-line swap to the RELATIVE form (p.chapters.len() <= q.chapters.len())
is documented at the call site and in the predicate doc, pending a hoard
sweep of SM3's real chapter counts.
Also add lowest playlist_id as the final canonical_title_order tie-break so
equal-size seamless siblings (00800-00808) pick 00800 deterministically.
Fixtures re-based on the real chapter discriminator (not a size coincidence)
and new tests cover both topologies (Alita superset-feature, SM3 subset-
feature) plus the un-chaptered-subset keep-superset case.
The keydb lookup returned an empty Vec for two very different outcomes:
a true hash miss, and a MATCHED disc that derived no key (e.g. an MK-only
entry with no VID). Both rendered as `keydb > no entry`, misleading
reporters whose disc IS present in the keydb.
Add KeySource::resolve_unit_keys -> UnitKeyResolution { keys, matched,
miss_path, matched_entry, store_entries }; the trait default delegates to
get_unit_keys (matched=false) so every existing source is unchanged.
resolve_and_apply_traced now renders MatchedDisc + NoVid/NoDerivableKey
for a matched-but-no-key disc and reserves NoEntry for a genuine miss.
New KeyNode::NoDerivableKey, a MatchedEntry shape (booleans + lengths, no
key material), and KeyStep.{matched_entry, store_entries} carry the
diagnostics to the app-layer renderer. Tests cover both verdict paths.
Bus content-range de-bus:
- F1 disc/mod.rs: always install content ranges (extracted wire_bus_removal);
empty ranges => Some([]) => de-bus NOTHING (fail-safe), not None => de-bus
everything — closes silent clear-sector corruption on the AACS cert route.
- F4 aacs/content.rs, F5 decrypt.rs: content gate is now a partition_point
binary search (O(log n)/sector) over the sorted, merged-disjoint ranges,
replacing the per-sector O(ranges) linear scan on the de-bus hot path.
- F2 scsi/linux.rs: RECOVERY_THREADS cap uses atomic reserve-then-check
(fetch_add first, fall back on overshoot) — removes the TOCTOU that let
concurrent multi-drive rips exceed MAX_RECOVERY_THREADS.
- F3 pes.rs: PesFrame::deserialize reads each chunk straight into data's tail;
drops the scratch Vec + double-copy, keeps the incremental-growth DoS guard.
- F7 disc/mod.rs: six inline rank_titles thresholds -> named consts.
Tests: bus fail-safe + scan wiring (both routes), per-sector gate straddle,
read_directory type-3 continuation/AED-skip/cycle-guard, strengthened
without_bus_removal no-op guard. F6 (pub decrypt_bus) left by-design.
Relocate AACS bus decryption to one spot at the bottom of the sector chain:
the Drive's SectorSource impl applies decrypt_bus_in_content (gated to
encrypted-content ranges), with the stage decided once from the unlock/
handshake result (BusStage::from_read_data_key). Remove read_data_key from
DecryptKeys::Aacs and AacsState and the two downstream decrypt_bus call sites;
every reader (key sampler, mux, unit-key validation) now sees already-de-bussed
content and never reasons about bus encryption. Add BusRemovalSectorSource
stream + tests; keep the AacsBusKeyUnavailable fail-loud gate.
The PlayItem STN (stream-number) table offset was hard-coded to 32, which
is only correct for a single-angle PlayItem. A multi-angle PlayItem inserts
number_of_angles(1) + flags(1) + (number_of_angles-1) angle references (10
bytes each) right after still_time, so its STN begins at 32 + 2 +
(number_of_angles-1)*10. The fixed offset lands inside that angle block and
misreads the stream counts as garbage (e.g. n_video=51 on Spider-Man 3 UHD
00245/00246.mpls, number_of_angles=2 => real offset 44), dropping every
stream including the primary video. That makes DiscTitle::has_video() return
false, which drops a multi-angle/seamless-branch feature from the disc's nav
candidate set and loses it in main-title selection.
Compute the STN offset from is_multi_angle (bit 4 of item[10]) + number_of
_angles. Add a regression test building a 2-angle PlayItem and asserting the
video stream is found past the angle block.
Validated on a 92-disc real UHD/BD corpus (offline scan): 0 main-feature
pick changes vs baseline, 3 multi-angle titles (Spider-Man 3 244/245/246)
recover has_video. Adds examples/corpus_titledump.rs as the regression tool.
The fallback composite gate in rank_titles flagged title P as a "play-all
wrapper" whenever some other title Q's clip set was a proper subset of P's,
Q had video, and Q.size >= 0.5 * P.size. main_feature_order then ranks
standalone-before-composite, demoting P below the decoy Q.
On a seamless-branch feature this inverts the pick: the real 00800 (2h03,
4 clips = body + intro/credits/seamless segments) is a SUPERSET of the
single-clip body decoy 00700 (1h50), and 700 >= 50% of 800's size, so 800
was mis-flagged composite and lost to the shorter 700 — the wrong title
gets ripped.
Add a duration-first wrapper-shape gate: P is only a composite when the
subset Q is EITHER a much shorter cut (Q.duration < 0.85 * P.duration, a
concat/play-all padding Q's runtime) OR nearly size-equal (Q.size >= 0.90 *
P.size, a bumper/outro wrapper). A near-full-length but markedly smaller
body subset (Q runs ~0.89 of P's duration yet holds far fewer bytes,
because P's branch segments add bytes not runtime) is a real seamless
branch, so P is kept. Uses per-title duration_secs/size_bytes, stays
capacity-independent and feature-gate-safe.
Preserves the real play-all avoidance: the 00245 bumper-wrapper decoy
(Q ~0.96 of both duration and size) stays demoted via the size limb, and
a concat play-all (subset < half the runtime) via the duration limb.
Add regression tests: seamless_branch_feature_not_demoted_issue_45 and
true_play_all_concat_wrapper_still_demoted. All existing selection tests
pass.
Bring every flagged comment block within its cap (inline // <=3 lines,
/// item-doc prose <=8): tighten wording, or convert an over-long inline
block that documents the item below it into a /// doc, preserving all
technical facts. No code, string literals, or logic changed.
Probe dir_is_case_insensitive against the real target volume by creating
the dir first. Previously the test probed a not-yet-created dir, which
returns a conservative true; that disagreed with extract_tree's own
post-mkdir probe on a case-sensitive host (Linux CI), panicking the test.
Both branches (collision on insensitive, clean coexist on sensitive) still
verified.
Matthew Jackson