hotfix: migration fix for pro v0.1.9 compatibility (#295)

[sujithatzackriya]

• Model selection bug fix: commands updated

• Model selection bug fix: Model selection commands added

• Fix UTF-8 string slicing panic in API response logging

Fixed panic that occurred when logging API responses containing multi-byte UTF-8 characters (e.g., Punjabi/Gurmukhi script). The issue was caused by direct byte-index slicing which could split multi-byte characters.

Root Cause:

• Line 298: &response_text[..std::cmp::min(200, response_text.len())] could slice in the middle of multi-byte UTF-8 characters
• Line 211: Similar issue with auth token truncation

Solution:

• Replaced byte-index slicing with character-aware truncation using .chars().take(N).collect::<String>()
• This safely respects UTF-8 character boundaries

Error Before Fix:

thread panicked at src/api/api.rs:298:50:
byte index 200 is not a char boundary; it is inside 'ਮ' (bytes 198..201)

• resolved conflicts and some fixes related to conflict resolving

• onboarding added for graceful database migration

• warning cleaning

• ui edits settings button changes and small tweaks

• dev and build scripts update

• Fixed the loading issue after the recording is done. monitor idle state along with error state

• Initial system audio recording check disabled

• Core audio disabled

• Dynamic routing and permission checks added

• ffmpeg related cmd popup in the middle fix

• Permission buttons

• DB and Model directory commands and logic added

• Updated settings navigation

• Added toaster

• Fixed warning during build for mac specific command

• Added the pipeline change to wait for 1.5 seconds of silence for better speech processing - more chunks during normal speech

• Added the pipeline change to wait for 1.5 seconds of silence for better speech processing - more chunks during normal speech

• Fixed the transcript lose due to skip chunk issue

• Added sidebar behaviour; Added bluetooth device fix

• Add commands for device events

• Document bluetooth playback issues

• Handle BT devices; resampling

• Handle BT devices; resampling; show warning

• Handle BT devices; resampling; show warning

• Notification settings update

• ollama endpoint addition and some ux fixes

• settings tab creation for summary models

• Added FFMPEG bug fix - find ffmpeg path ratherthan using hardcoded path

• fix initial ollama model fetch even for non ollama providers

• Not fetching issue fixed

• mplemented comprehensive analytics tracking across the Meetily application to monitor user behavior, feature adoption, and application health metrics. All analytics are logged to PostgreSQL via the existing PostHog infrastructure with minimal local state management in analytics.json.

• Bugfix : Analytics tauri store version update

• Enhanced analytics with ability to capture major details such as device selection, language selection, preference changes, recording settings

• Package missmatch error was shown during build; fixed it

• config data sync fix

• Build error fixes are applied

• ffmpeg popup issue fix

• Added windows/linux build fix

• Implement Homebrew database auto-detection feature

Added a new component, HomebrewDatabaseDetector, to check for the existence of a Homebrew database and retrieve its size. Integrated this component into the LegacyDatabaseImport dialog. Updated Tauri commands to support database detection functionality.

• Enhance Homebrew database detection to support multiple paths

Updated the HomebrewDatabaseDetector component to check for the existence of the Homebrew database in both Apple Silicon and Intel Mac paths. The detected path is now displayed dynamically, improving user experience and compatibility.

• Refactor AnalyticsConsentSwitch and AnalyticsProvider to remove default analyticsOptedIn setting

Updated both components to eliminate the default value for analyticsOptedIn in the load function, streamlining the analytics preference handling process.

• Fix model selection logic to respect saved configuration, enable selection in settings page, and optimize modal loading

• Fix model download progress bug

• Fix model loading bug

• Removed unwanted logs

• Updated with home page and settings page bug fix for model selection

• Update version to 0.0.6 and set default analyticsOptedIn in multiple components

Bumped the application version to 0.0.6 in package.json, About component, Sidebar, and Cargo.toml. Updated AnalyticsConsentSwitch and AnalyticsProvider to include a default value for analyticsOptedIn in the load function, enhancing analytics preference management.

• Refactor HomebrewDatabaseDetector to enhance import functionality and user interaction

Updated the HomebrewDatabaseDetector component to replace the onDatabaseFound callback with onImportSuccess and onDecline. Added user prompts for importing existing databases with success and error handling, improving the overall user experience. Integrated the updated component into LegacyDatabaseImport. Bumped version to 0.0.6 in tauri.conf.json.

• Downgraded whisper-rs and changed beam size for windows for achieving more speed

• Initial parkeet test

• Big bad refactoring - Optimizing the transcription

• summary section removed in home(recording page)

• transcript metadata location change and transcript max-w set

• streaming effect added in the transcriptview

• Added the UX fixes to new model nanagement and transcription flow

• auto-scrolling base

• auto scroll done

• simplified model selection with simple names and accordion

• Fixed issue with transcription history management issue that occured after merge

• UX improvements - Model management, UI button position fixes

• Default model selection - changed to parkeet

• Removed the onSave

• Transcript filter update

• Updated audio processing with better control

• UX improvements

• UX improvements

• Removed repeated letters and words being shown - bad UX. cleanup transcript

• Remove vulkan as default for linux

• Recording pause timer update

• Better transcription accuracy after changing the redumption time.

• Removed normalizer

• readme updated

• Porting to core audio from screen capure kit

• Fixed windows powershell pop up issue

• Advanced mic noise removal

• Recording audio fixes; Recording status indicator ux fix

• Windows specific redumption time

• Windows specific redumption time

• Audio quality issue fix

• Removed the debug log

• Added EBU R128 Normalization to system audio

• Removed EBU R128 Normalization to system audio

• Removed BT device warning

• Better audio mixing and device detection logic for better processing

• Before FFMPEG mixer addition

• Removed the echo in recorded system audio

• Bluetooth device fix: Added advanced resampling

• Removed the unwanted notification library

• Justify transcript text

• UI changes; version update

• Name update

• Fix macOS system audio echo caused by duplicate capture paths

• Fix : Analytics focused fixed

• Enhancement : Analytics

• Model management fix - parakeet

• Model management fix - whisper

• UX enhancements - added toasters

• UI enhancements : Recording notifications; Navigation on sidebar

• Fix: Download model issue - duplication

• Sync recording status with sidebar

• Transcript processing logic bug fix

• Fix: Recording metadata, transcript saving, and meeting details integration

• Added basic templates

• Added template files

• Added template based selection for processing

• Added more templates

• Details page cleanup

• Details page cleanup: added hooks

• Details page cleanup: moved to components

• Changed the UI behaviour : blovknotes summary

• Added the ollama download/fetch/update options

• Added the ollama download/fetch/update options

• Added new components from shadcn

• Added tauri config update for template fetching

• Auto generate meeting notes

• Updated model management

• Updated model management; enhanced UX

• Updated model management: Added the download status chcking for ollama models

• UI enhancement - button positioning, UI woking enhancement

• Build error fixes; Ollama custom download

• Build error fixes; Ollama custom download

• Dynamic context length selection for ollama models

• Added the meeting recordings button and functionality

• Added button group styling

• Fixes for Smooth UX

• Fixes for Smooth UX

• Fixed auto scroll missmatch

• Fixed auto scroll missmatch

• UI fixes

• Fix: Default model issue fixed

• installation content in readme updated

• acknowledgement added

• Added new images and gif

• Meetily ux edits (Meetily ux edits #209)

• parakeet defaulted, recording page buttons changed, sidebar icon, etc

• sidebar search input updated

• backend defaulting of parakeet model

• only backend activetime duration syncing

• transcription models selection scrolling issue

• PArakeet model selection fix

• Open the model downloader if no model exist

• Ollama frontend download stuff fixed

---

• removed unused code(function, states, variable)

• SettingsModal extracted, added place holders in page

• TranscriptPanel extracted

• all component refactored from page

• Contexts Extracted and implemented in recording page(app/page.tsx)

• configs,recording,storage and transcription service extracted and applied to page and 3 Contexts

• Hooks extracted and refactored components and page again for using the new context and hooks

• mistake fix

• auto summary toggle and stop generation in provider level to the UI

• summary loss on generation fail or cancellation fixed

• emoji remove

• Bugfix/sidebar start and modal fix (Bugfix/sidebar start and modal fix #242)

• sidebar start fixed along with tray stopping issue fix

• language modal fixed

• settings navigation from tray fix

• auto save toggling persistance fix

• language button visibility fix according to provider

• redundant recording button removed

• incremental saver removed for auto saved turned off recording

• removed permission check for linux, as it is not relevent

• added timestamp to the summary generation prompt

• Updated workflow

• Added permission checker and dialogue box

• Revert "Added permission checker and dialogue box"

This reverts commit 8c4881a.

• Revert "Updated workflow"

This reverts commit fe1aac9.

• base version done

• dev and build script update and ts error fix

• forcing vulkan build for windows platform

• appImage bundling issue with llama helper executable not found - Fixed

• seperated logic of llam-helper bundling from tauri auto

• linux build ensured

• windows build script updated for sidecar building and sidecar spawning with cmd fixed

• model unloading on provider added

• some minor optimization fro performance

• timeout and stop builtin fixed, priority reduced for llama-helper, dev script update

• initial onboarding

• simplified steps, and added mac specific permission check, UI enhancement and twaks

• Added initial fix with the changes made to improve model download performance and add detailed progress tracking (MB/speed) to the UI for both

• Fixed permission check bug

• modele size updated

• removed mistral 7b

• data injection script added

• claude models list updated

• model list while downloading fixed, cancel download fixed

• added stats to llama helper, increased timeout for long script handling

• onboarding completion order fix, token estimation fix, orphaned process on app exit fix

• download remove call twice fix, parakeet onboarding state update fix, cleanup of partial download error, removing duplicate completed progress event on builtin model

• custom-openai endpoint added

• Built fix

• onboarding simplified with 4 steps with background download support of models, gemini backend setup

• download error handling in UI while downloading and not available scenario

• custom endpoint model settings data loading fix, test endpoint fix

• migrations changed and removed gemini for compatibility

• build fix and model recommendation update on onboarding

• toast and model download related error handling

• ota update added

• updated gitignore for env.example

• onboarding and toast styliing and tweaks

• simple edit

• transcript recovery added

• fix: ffmpeg audio bug in release build

• Add GitHub Actions workflows with manual triggers

• Workflow file updates with version management and the release build edits

• Fix signing command issues

• Signing logic fix

• Signing logic fix for windows

• Fix vulkan issues

• Avoid the llama-helper error with addition of vulkan command

• fix: in onboarding already downloaded parakeet model causing error

• restricting windows system audio device change

• use audio devices according to user preference, removed the restriction of system audio device change of windows devices

• fix: release microphone when recording stops

  • Add stream.pause() before dropping CPAL stream to stop callbacks immediately
  • Clear microphone_device and system_device Arc references in stop_recording()
  • Call state.cleanup() explicitly in stop_streams_and_force_flush()
  • Store transcript listener ID and unlisten on stop
  • Add brief sleep after aborting Core Audio task to allow cleanup

  The microphone was staying active because:

  1. CPAL stream callbacks held Arc references
  2. AudioCapture held Arc which held Arc
  3. Device references were never cleared in stop_recording()

• fix: Windows shutdown hang and mic array audio issues

  • Stop device monitor FIRST in stop_streams_and_force_flush() to prevent continuous WASAPI polling during shutdown (fixes 90+ second hang)
  • Fix audio_to_mono() for Windows microphone arrays (>2 channels): Only use first 2 channels to avoid destructive interference from beam-forming/noise cancellation auxiliary channels
  • Reduce WASAPI device enumeration log verbosity (info -> debug)

• added user id display in settings

• recording stop from tray fixed, recovery dialog for on going meeting fixed, added timeouts for reliability

• pausing from tray synchronization with recording controls via context

• fix; meerging stop operation from tray and UI, elimination of desyn an duplicate recovery dialogs

• cleanup of sessionstorage usage

• model downloading error handling and progress toast updation

• cleanup up db connection and checkpoints and graceful closing of db pool

• removing right click context menu for production build

• feat: add transcript pagination and virtualization for performance

• Add backend pagination with SQL LIMIT/OFFSET (100 transcripts per page)
• Implement virtual scrolling with @tanstack/react-virtual (10+ segment threshold)
• Create usePaginatedTranscripts, useAutoScroll, and useTranscriptStreaming hooks
• Add VirtualizedTranscriptView with conditional rendering and infinite scroll
• Integrate pagination into meeting details page (replaces api_get_meeting)
• Fix streaming text effect, status bar overlap, and auto-scroll bugs

Performance: 80-90% faster initial load, 60-80% RAM reduction for large meetings

• perf: optimize settings general tab with lazy loading

Move preference settings to ConfigContext with ref-based caching to prevent redundant IPC calls on every tab visit.

• ui-enh: shadcn tabs implemented in settings

• fix: copy all transcripts and disable auto-scroll on details page

• Copy operation now fetches all transcripts from DB instead of using paginated state
• Auto-scroll disabled on meeting details to allow manual navigation
• Removed auto scrolling if the user scrolled up the transcript

• migrations added

• chore: version bump to v0.2.0

• chore: README update

• change: added requested changes

• Updated the workflows to remove missmatches

• readme pro link updates

• Updated README and added pro screenshot

• Updated README and added pro screenshot

• hotfix: migration fix for pro v0.1.9 compatibility

---

Description

[Provide a detailed description of your changes]

Related Issue

[Link to the issue this PR addresses (e.g., "Fixes #123")]

Type of Change

• Bug fix
• New feature
• Documentation update
• Performance improvement
• Code refactoring
• Other (please describe)

Testing

• Unit tests added/updated
• Manual testing performed
• All tests pass

Documentation

• Documentation updated
• No documentation needed

Checklist

• Code follows project style
• Self-reviewed the code
• Added comments for complex code
• Updated README if needed
• Branch is up to date with devtest
• No merge conflicts

Screenshots (if applicable)

[Add screenshots here if your changes affect the UI]

Additional Notes

[Add any additional information that might be helpful for reviewers]