A full verification of the application against docs/BRIEF.md, CLAUDE.md, the UI kit, the tests and the Docker and client setups, repeated until it converged. Every finding was confirmed by an independent skeptic before it was fixed, and every fix was reviewed independently before merging.
Pass
Confirmed
Areas
1
71
Server hardening and deploy, ingest/upload, imaging and metadata, proxies/render nodes/sync, Settings pages, shell and notifications, app fixes, tests and CI (new installer job), setup and .titano project bundles
2
37
Proxies and decode tiers, project positions crossing copies through sync, server scans and maintenance, desktop settings and polling, a portable copy inside OneDrive, setup and deployment guidance
3
16
Editors open while photos arrive, partial writers of the index row, image passes after another writer's import, error-text and documentation gaps
4
7
Lock order of index writes (PostgreSQL deadlocks), analyses and scans racing another writer
5
4
Tag lock order on PostgreSQL, the orphans fix's follow-up analyses, deep reports saved over newer ones, frames versions shared by two indexes
6
2
An unreadable photo counted as a change, a deep report made stale by proxies
All are fixed with tests; choices are recorded in docs/DECISIONS.md (rows dated 2026-10-12/13), user-facing changes in docs/CHANGELOG.md, and docs/STATUS.md summarises the verification. New strings are in all 38 cultures (36 machine translations, each independently reviewed).
Known open item: the screenshot pixel diff (BRIEF §20.2) is wired into CI but compares nothing until the first baselines are accepted from ci-screenshots (STATUS › Next steps).
Test plan
dotnet build Titano.slnx -c Release: 0 warnings, 0 errors (Windows projects included through EnableWindowsTargeting).
Linux suites with PostgreSQL 16: Core 1189, Data 489, Server 148, Tools 25 passed, 0 failed; Data and Server repeated for flakiness.
Docker: the current image built and run with PostgreSQL 17, a fresh root-owned archive, remapped ports: create, upload, download (originals identical), unauthenticated large chunk answered 401, scan/maintenance/rebuild keep a sequence waiting for photos, hand-copied photos get thumbnails and proxies, data kept across a restart.
Windows CI (build/test matrix, App and Core.Windows suites, screenshots, installer job): not run — every job is stopped within seconds by the account's GitHub Actions billing/spending limit (see the comment below).
## Summary
A full verification of the application against `docs/BRIEF.md`, `CLAUDE.md`, the UI kit, the tests and the Docker and client setups, repeated until it converged. Every finding was confirmed by an independent skeptic before it was fixed, and every fix was reviewed independently before merging.
| Pass | Confirmed | Areas |
|---|---|---|
| 1 | 71 | Server hardening and deploy, ingest/upload, imaging and metadata, proxies/render nodes/sync, Settings pages, shell and notifications, app fixes, tests and CI (new **installer** job), setup and `.titano` project bundles |
| 2 | 37 | Proxies and decode tiers, project positions crossing copies through sync, server scans and maintenance, desktop settings and polling, a portable copy inside OneDrive, setup and deployment guidance |
| 3 | 16 | Editors open while photos arrive, partial writers of the index row, image passes after another writer's import, error-text and documentation gaps |
| 4 | 7 | Lock order of index writes (PostgreSQL deadlocks), analyses and scans racing another writer |
| 5 | 4 | Tag lock order on PostgreSQL, the orphans fix's follow-up analyses, deep reports saved over newer ones, frames versions shared by two indexes |
| 6 | 2 | An unreadable photo counted as a change, a deep report made stale by proxies |
All are fixed with tests; choices are recorded in `docs/DECISIONS.md` (rows dated 2026-10-12/13), user-facing changes in `docs/CHANGELOG.md`, and `docs/STATUS.md` summarises the verification. New strings are in all 38 cultures (36 machine translations, each independently reviewed).
Known open item: the screenshot pixel diff (BRIEF §20.2) is wired into CI but compares nothing until the first baselines are accepted from `ci-screenshots` (STATUS › Next steps).
## Test plan
- [x] `dotnet build Titano.slnx -c Release`: 0 warnings, 0 errors (Windows projects included through EnableWindowsTargeting).
- [x] Linux suites with PostgreSQL 16: Core 1189, Data 489, Server 148, Tools 25 passed, 0 failed; Data and Server repeated for flakiness.
- [x] `i18n check` (2892 keys, 38 cultures, 0 errors, 0 warnings), `xaml-lint`, `package-guard`, `dotnet format whitespace`.
- [x] Docker: the current image built and run with PostgreSQL 17, a fresh root-owned archive, remapped ports: create, upload, download (originals identical), unauthenticated large chunk answered 401, scan/maintenance/rebuild keep a sequence waiting for photos, hand-copied photos get thumbnails and proxies, data kept across a restart.
- [ ] Windows CI (build/test matrix, App and Core.Windows suites, screenshots, installer job): not run — every job is stopped within seconds by the account's GitHub Actions billing/spending limit (see the comment below).
🤖 Generated with [Claude Code](https://claude.com/claude-code)
https://claude.ai/code/session_012iLMBNwBE2GnphAsu4cjaC
CI on 5960516 (run 38247624738): every job (Build and test on Windows and Ubuntu, PostgreSQL 16/17, Docker image, Installer, Screenshots branch) failed within 2 seconds without being assigned a runner, so there are no steps and no logs. The Screenshots branch job of the previous run, 38245684628, had already failed the same way at 12:09 UTC, while the other jobs of that run had just been running normally.
This does not come from the code in this PR: no job started. The likely cause is the account's GitHub Actions allowance. Artifact uploads already report "Artifact storage quota has been hit", and storage and minutes come under the same billing limits. The last completed runs passed the Installer, Docker image, PostgreSQL 16/17 and Ubuntu jobs. Their only Windows test failures were two tests, both now fixed on this branch (SettingsScreenSmokeTests in 162197b, the single-instance race in 5960516).
Needed to resume: in the account's Settings › Billing, check the Actions minutes and storage usage (raise the spending limit, or wait for the monthly reset), and delete old workflow artifacts. Then push again or re-run the workflow. I'm keeping the PR watched.
CI on 5960516 (run 38247624738): every job (Build and test on Windows and Ubuntu, PostgreSQL 16/17, Docker image, Installer, Screenshots branch) failed within 2 seconds without being assigned a runner, so there are no steps and no logs. The Screenshots branch job of the previous run, 38245684628, had already failed the same way at 12:09 UTC, while the other jobs of that run had just been running normally.
This does not come from the code in this PR: no job started. The likely cause is the account's GitHub Actions allowance. Artifact uploads already report "Artifact storage quota has been hit", and storage and minutes come under the same billing limits. The last completed runs passed the Installer, Docker image, PostgreSQL 16/17 and Ubuntu jobs. Their only Windows test failures were two tests, both now fixed on this branch (`SettingsScreenSmokeTests` in 162197b, the single-instance race in 5960516).
Needed to resume: in the account's Settings › Billing, check the Actions minutes and storage usage (raise the spending limit, or wait for the monthly reset), and delete old workflow artifacts. Then push again or re-run the workflow. I'm keeping the PR watched.
---
_Generated by [Claude Code](https://claude.ai/code)_
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Summary
A full verification of the application against
docs/BRIEF.md,CLAUDE.md, the UI kit, the tests and the Docker and client setups, repeated until it converged. Every finding was confirmed by an independent skeptic before it was fixed, and every fix was reviewed independently before merging..titanoproject bundlesAll are fixed with tests; choices are recorded in
docs/DECISIONS.md(rows dated 2026-10-12/13), user-facing changes indocs/CHANGELOG.md, anddocs/STATUS.mdsummarises the verification. New strings are in all 38 cultures (36 machine translations, each independently reviewed).Known open item: the screenshot pixel diff (BRIEF §20.2) is wired into CI but compares nothing until the first baselines are accepted from
ci-screenshots(STATUS › Next steps).Test plan
dotnet build Titano.slnx -c Release: 0 warnings, 0 errors (Windows projects included through EnableWindowsTargeting).i18n check(2892 keys, 38 cultures, 0 errors, 0 warnings),xaml-lint,package-guard,dotnet format whitespace.🤖 Generated with Claude Code
https://claude.ai/code/session_012iLMBNwBE2GnphAsu4cjaC
CI on
5960516(run 38247624738): every job (Build and test on Windows and Ubuntu, PostgreSQL 16/17, Docker image, Installer, Screenshots branch) failed within 2 seconds without being assigned a runner, so there are no steps and no logs. The Screenshots branch job of the previous run, 38245684628, had already failed the same way at 12:09 UTC, while the other jobs of that run had just been running normally.This does not come from the code in this PR: no job started. The likely cause is the account's GitHub Actions allowance. Artifact uploads already report "Artifact storage quota has been hit", and storage and minutes come under the same billing limits. The last completed runs passed the Installer, Docker image, PostgreSQL 16/17 and Ubuntu jobs. Their only Windows test failures were two tests, both now fixed on this branch (
SettingsScreenSmokeTestsin162197b, the single-instance race in5960516).Needed to resume: in the account's Settings › Billing, check the Actions minutes and storage usage (raise the spending limit, or wait for the monthly reset), and delete old workflow artifacts. Then push again or re-run the workflow. I'm keeping the PR watched.
Generated by Claude Code