fix(mediorum): keep audio analysis from completing uploads - #570
Merged
Merged
Conversation
raymondjacobson
deleted the
codex/fix-audio-analysis-upload-completion
branch
September 8, 2026 18:37
rickyrombo
added a commit
that referenced
this pull request
Sep 15, 2026
Unit tests pin each piece of content auth; nothing ran the whole chain. This uploads audio over tus with an asserted user to one node, waits for the attestation to commit before the upload reads done, and validates track writes at a different node's mempool: the owner's create passes, another user naming the same cid is rejected, an unattested cid is rejected, and edits that resend the cids (as the web client does) keep passing. Also pins that an audio upload naming no user is refused at tus create, and it caught the pre-#570 done-without-cid race on a stale devnet image. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
rickyrombo
added a commit
that referenced
this pull request
Sep 16, 2026
…et (#574) Unit tests pin each piece of content auth; nothing ran the whole chain. This uploads audio over tus with an asserted user to one node, waits for the attestation to commit before the upload reads done, and validates track writes at a different node's mempool: the owner's create passes, another user naming the same cid is rejected, an unattested cid is rejected, and edits that resend the cids (as the web client does) keep passing. Also pins that an audio upload naming no user is refused at tus create, and it caught the pre-#570 done-without-cid race on a stale devnet image. Co-authored-by: Claude Fable 5.1 <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.
Audio analysis can publish an audio upload as
donebefore its CID is persisted. During inline analysis,transcodeFullAudiohas populated only the in-memory results. Successful analysis reloads the older database row and writesdonewith no CID; transcode writes the complete results afterward. Clients polling between these operations can publish a CIDless track. Core can expose the intermediate operation across block boundaries.Keep the overall upload status owned by transcode completion, which already persists results after required attestation. Both analysis success and failure now refresh the persisted row and update only analysis fields, preserving upload status, transcode results, and mirrors. Failed analysis remains non-blocking for transcode completion.
Validation:
doneoperation without a320CID.gofmtandgit diff --checkpass. The Docker-based Mediorum suite must run in CI; Docker is unavailable locally.This fixes a source-level mechanism, not an independently verified attribution of the production incidents. The premature write dates to #120; #364 retained peer broadcasts with Core writes disabled by default, and #366 later removed peer transport. July onset needs deployment/flag history. Existing CIDless track metadata and stuck transcodes require separate repair/investigation.
Related: AudiusProject/apps#14585