Repository navigation
[FEATURE] Say what the app is doing in the system log - #131
Merged
Merged
Conversation
There was not one Logger call in the project. When dictation stopped hearing anything on Bluetooth, AVAudioEngine had stopped itself and said so to nobody, and finding it took four throwaway binaries reproducing the audio stack outside the app. Four categories under one subsystem: session, audio, speech, delivery. The lines are the ones that were silent by construction — the route change and the rebuild that answers it, a session beginning and how it ended, a model download, and every path where shaping quietly returns the words as spoken. Nothing that carries what the user said goes through it. Every value marked public is an error description, a locale identifier, an enum case name, a count or a duration. The word count of a session is logged; the words are not. Closes #129 Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_012Abfbei4sXuCngAWEH8Z2k
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
There was not one Logger call anywhere in the project. When dictation stopped hearing anything on Bluetooth, AVAudioEngine had stopped itself and said so to nobody: log show returned nothing from this process, and finding it took four throwaway binaries reproducing the audio stack outside the app.
Four categories under one subsystem, com.tgomareli.Talkify: session, audio, speech and delivery. The lines added are the ones that were silent by construction. The audio route changing and whether the rebuild that answers it worked. A session beginning, with its trigger slot and locale, and how it ended, with the outcome and the word count. A speech model downloading. And every path where shaping quietly hands back the words as spoken, which is correct for the user and unhelpful when someone asks why their prompt did nothing.
The constraint that shaped it: nothing carrying what the user said goes through it. Logger redacts interpolations unless they are marked public, so the rule on top is that only error descriptions, locale identifiers, enum case names, counts and durations are ever public. A session's word count is logged; the words are not. Every call site is listed in the type's doc comment and I audited each public interpolation against that rule.
Verified that the subsystem, the categories and the privacy annotations behave, including that a public value renders rather than reading as private. I could not drive a real dictation session through the status menu to capture an end-to-end trace, so the call sites themselves are verified by reading rather than by running.
Closes #129