feat(internalization-room): the room writes down what was heard - #232
Merged
Merged
Conversation
The tablet already refuses 'terminei' until the team's own recording has played to its end; nothing recorded that fact. The finish route now accepts an optional report of the played ranges and the clip's duration, keeps it on the telling-back (and on the superseded attempt when the team retakes), and a pure rule says whether the reported playback covered the whole clip within 750 ms of tolerance. No report at all is a legacy client and changes nothing — this is evidence for the Refine artifact, not a new wall for old tablets. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This was referenced Aug 22, 2026
joaocarvoli
marked this pull request as ready for review
August 22, 2026 16:37
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Substitui o #202, que não podia mais ser mergeado.
Por que um PR novo
Os 44 PRs abaixo deste foram squash-merged, então o
maincarrega o conteúdo deles sob commits novos, sem história em comum. A branch do #202 continuava carregando a cópia antiga de tudo isso — 42 commits atrás domain— e mergeá-la teria somado esta feature e revertido em silêncio ~500 linhas do trabalho mais recente domain: as correções de revisão feitas durante cada squash, junto com os testes de regressão que as protegem.A CI não teria pego: a branch traz as suas próprias cópias antigas desses testes. Um check verde teria certificado a regressão.
O que este PR é
Exatamente o commit próprio do #202 (
adbb792, "the room writes down what was heard") aplicado sobre omainatual, e nada mais:Zero deleções.
played_rangeseclip_duration_msnofinishda retrotradução,played_ranges_cover_clipcom tolerância de 750 ms, e o teste que os cobre.Uma diferença deliberada em relação ao commit original: o parâmetro
response: Responsefoi deixado de fora dofinish, porque não é usado em lugar nenhum do corpo da função e obrigaria um import só para ficar parado ali.Gates locais:
ruff,ruff format,mypylimpos; suíte completa 1531 passando. A única falha étest_sound_necklace/test_working_time.py, que vive nomain, não é tocada por este PR e passa 3 de 3 isolada — flake pré-existente sob carga da suíte inteira.🤖 Generated with Claude Code
https://claude.ai/code/session_015r75P6pEvJJNofz5159G9A