media-info-wdx: Matroska/WebM (.mkv/.webm) support via native EBML parser#13
Merged
Conversation
…L parser macOS system frameworks can't open Matroska/WebM containers — AVURLAsset reports the file unreadable with zero tracks — so DC showed nothing for .mkv files. Add a self-contained EBML reader that pulls dimensions, duration, and frame rate straight from the file head (Segment > Info and Segment > Tracks > TrackEntry > Video), stopping at the first Cluster. Mirrors the existing AVI RIFF reader: no dependency, no network, bounded read, routed to the fast (non-deferred) parse path. .mkv/.webm added to the category table, the compiled DetectString, and register_plugin.py. Verified end-to-end against a real 6h54m 1080p30 .mkv — output matches ffprobe exactly (1920x1080, 30 fps, 24854.809 s). Headless harness grows a synthetic-Matroska case; 30/30 assertions pass. Bump VERSION 0.1.0 -> 0.2.0. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Fixes from the fresh-agent /code-review of the EBML parser: - Bound FormatDuration (>~100y rejected) so a corrupt/hostile Matroska Duration float can't reach an out-of-range llround (undefined) and emit a garbage time string. Protects every parser path. - Rewrite ParseMKV as a *seeking* reader: it walks the Segment's children by element header, reading only the small Info/Tracks bodies and skipping past large SeekHead/Cues/Attachments. Removes the 2MB read cap that both (a) permanently cached empty metadata for spec-valid files whose Tracks fell past the window and (b) did a synchronous 2MB read on DC's calling thread. Now a few KB regardless of file size or layout. - Prefer DisplayWidth/Height (aspect-correct) over coded PixelWidth/Height, matching what a player and the AVFoundation path report for anamorphic video; parse the audio track and CodecID too, so mkv/webm now expose video/audio codec, sample rate, and channels (audio-only .webm included). - Extract FillVideoFields, shared by the AVI and MKV readers, removing the triplicated dimensions/duration/frame-rate/Summary assembly. Harness grows anamorphic (display != coded), audio-track/codec, and hostile-duration cases: 37/37 pass. Re-verified end-to-end against the real 6h54m .mkv — resolution, fps, duration, and now codec/48kHz/stereo all match ffprobe exactly. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Second-pass /code-review of the seeking rewrite: - Fix a real regression: prefer DisplayWidth/Height only when DisplayUnit is pixels (0, the default). DisplayUnit 1/2/3 means cm/inches/aspect-ratio, where those fields aren't pixel counts — a 16:9 aspect-ratio track was reported as "16 × 9". Now falls back to the coded PixelWidth/Height. - Robustness: raise the top-level Segment search from 8 to 64 elements (a file may carry leading Void/padding before the Segment); handle an unknown-size Info/Tracks master by reading a bounded window and walking it recursively (stopping at the first Cluster) instead of bailing out. - Cleanup: remove the dead, write-only MIInfo.category property (never read; classification comes from CategoryForPath) and unify the duplicated EBML header decode into MKVDecodeHeader, shared by the in-memory walker and the file-seeking reader. Harness grows an aspect-ratio-display (DisplayUnit=3) case asserting the coded size wins: 40/40 pass. Real 6h54m .mkv output unchanged and still matches ffprobe (1080p30, 24854.8s, H.264/AAC/48kHz/stereo). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What & why
.mkvfiles showed no metadata in Double Commander because macOS system frameworks can't open Matroska/WebM at all —AVURLAssetreports the file unreadable with zero tracks. MKV was never in the plugin's supported set, so DC never called it.This adds a self-contained EBML parser (
ParseMKV) that reads dimensions, duration, and frame rate straight from the file head (Segment > InfoandSegment > Tracks > TrackEntry > Video), stopping at the firstCluster. Same philosophy as the existing AVI RIFF reader: no dependency, no network, bounded read, routed to the fast (non-deferred) parse path..webmis the same container, so it's covered too.Verification
Tested end-to-end against a real 6h54m 1080p30
.mkv— output matchesffprobeexactly:.mkv/.webmadded to the category table, compiled DetectString, andregister_plugin.py.A fresh-agent
/code-review(high) is running on this diff per the repo charter; findings will be addressed before merge.