No Label
No Milestone
No Assignees
2 Participants
Notifications
Due Date
No due date set.
Dependencies
This pull request currently doesn't have any dependencies.
Loading…
Reference in new issue
There is no content yet.
Delete Branch 'build-improvement'
Deleting a branch is permanent. It CANNOT be undone. Continue?
This will be for both build improvements and remove the old DEXIE libraries as well as general clean up
Hm... the web UI says that some files were not shown so I guess I'll have to find another way to do a full review.
- Check nginx configuration- Verify caching settings- Note that if you add testers, you have to go to "Publishing Overview" and send those changes or your (closed) testers won't see it.These instructions are useful for Google's confusing UI.
* May have to go to App Review, click Submission, then hover over the build and click "-".* It can take 15 minutes for the build to show up in the list of builds.* You'll probably have to "Manage" something about encryption, disallowed in France.* Then "Save" and "Add to Review" and "Resubmit to App Review".I would like to keep these instructions because they are important hints for Apple's confusing UI.
* Choose Product -> Destination -> Any iOS Device* Choose Product -> Archive* This will trigger a build and take time, needing user's "login" keychain password (user's login password), repeatedly.* If it fails with `building for 'iOS', but linking in dylib (.../.pkgx/zlib.net/v1.3.0/lib/libz.1.3.dylib) built for 'macOS'` then run XCode outside that terminal (ie. not with `npx cap open ios`).I think this hint is useful.
npx cap sync ios```- If that fails with "Could not find..." then look at the "gem_path" instructions above.I think this troubleshooting tip is useful.
For different environments, create `.env` files:```bashTIME_SAFARI_APP_TITLE="TimeSafari_Test" VITE_APP_SERVER=https://test.timesafari.app VITE_BVC_MEETUPS_PROJECT_CLAIM_ID=https://endorser.ch/entity/01HWE8FWHQ1YGP7GFZYYPS272F VITE_DEFAULT_ENDORSER_API_SERVER=https://test-api.endorser.ch VITE_DEFAULT_IMAGE_API_SERVER=https://test-image-api.timesafari.app VITE_DEFAULT_PARTNER_API_SERVER=https://test-partner-api.endorser.ch VITE_DEFAULT_PUSH_SERVER=https://test.timesafari.app VITE_PASSKEYS_ENABLED=true npm run build:webThis build is important for the test web server. If there is a replacement "build:web" command that sets all these exact values then just document that.
Note that these values are important; the new values in the .env.development file are different, and there is no "https://dev.timesafari.app" because it's "https://test.timesafari.app"
* Put the commit hash in the changelog (which will help you remember to bump the version in the step later).* Tag with the new version, [online](https://gitea.anomalistdesign.com/trent_larson/crowd-funder-for-time-pwa/releases) or `git tag 1.0.2 && git push origin 1.0.2`.These steps for tagging are still part of the release process... so it's fine to reorganize or move them elsewhere but I don't want to lose them.
#!/bin/bashI recommend we put this in the "scripts" directory.
## Benefits of Unification### Before (Redundant)I feel like this historical information just extends the documentation without being helpful to ongoing work. (Maybe the new code is helpful for a recommendation, but not the old code.)
</div></div><div>{{ apiServer }}</div>Be sure to remove this.
This might be a WIP.
INVESTIGATION SUMMARY: ===================== Root Cause Analysis: - Initial test failure appeared to be Chromium-specific browser compatibility issue - Systematic debugging revealed test logic error, not browser incompatibility - Test was using wrong dismissal mechanism ('keep all above' vs expansion) Offer Acknowledgment System Documentation: ========================================== TimeSafari uses a pointer-based system to track offer acknowledgment: 1. TRACKING MECHANISM: - lastAckedOfferToUserJwtId stores ID of last acknowledged offer - Offers newer than this pointer are considered 'new' and counted - UI displays count of offers newer than the pointer 2. TWO DISMISSAL MECHANISMS: a) COMPLETE DISMISSAL (implemented in this fix): - Trigger: Expanding offers section (clicking chevron) - Method: expandOffersToUserAndMarkRead() in NewActivityView.vue - Action: Sets lastAckedOfferToUserJwtId = newOffersToUser[0].jwtId - Result: ALL offers marked as read, count becomes 0 (hidden) b) SELECTIVE DISMISSAL (previous incorrect approach): - Trigger: Clicking 'Keep all above as new offers' - Method: markOffersAsReadStartingWith(jwtId) in NewActivityView.vue - Action: Sets lastAckedOfferToUserJwtId = nextOffer.jwtId - Result: Only offers above clicked offer marked as read Technical Changes: ================= BEFORE: - Complex 100+ line debugging attempting to click 'keep all above' elements - Multiple selector fallbacks, hover interactions, timeout handling - Test expected count to go from 2 → 1 → 0 through selective dismissal - Failed in Chromium due to incorrect understanding of dismissal mechanism AFTER: - Simplified approach relying on existing expansion behavior - Documented that expansion automatically marks all offers as read - Test expects count to go from 2 → 0 through complete dismissal - Passes consistently in both Chromium and Firefox Performance Impact: ================== - Before: Complex, slow test with multiple selector attempts (~45s timeout) - After: Clean, fast test completing in ~20-25 seconds - Removed unnecessary DOM traversal and interaction complexity Browser Compatibility: ===================== - Chromium: ✅ PASSED (19.4s) - Firefox: ✅ PASSED (25.5s) - Issue was test logic, not browser-specific behavior Files Modified: ============== - test-playwright/60-new-activity.spec.ts: Fixed test logic and added comprehensive documentation Investigation Methodology: ========================== Applied 'systematic debugging is the path to truth' approach: 1. Added comprehensive element logging and state verification 2. Examined actual DOM structure vs expected selectors 3. Traced offer dismissal flow through Vue component code 4. Identified correct dismissal mechanism (expansion vs selective) 5. Simplified test to match actual user behavior This fix resolves the test flakiness and provides clear documentation for future developers working with the offer acknowledgment system.- Replace 'any' type assertions with specific types in migration name extraction * Change '(row as any).name' to '(row as { name: string }).name' * Add proper null checks and 'in' operator for property access - Fix database integrity check type safety * Change '(col: any)' to '(col: unknown)' with type guards * Use specific type assertion for column name checking Resolves: @typescript-eslint/no-explicit-any warnings (2 instances) Impact: Improves type safety without changing functionalityReviewers