Resolve pending commit links

Update the framework to 150819a, follow the new milestone workflow template, review RC-027
Merge pull request 'MIL-010: gym accounts from a credentials file' (#59 ) from mil-010-credentials-file into main
2026-10-29 00:11:16 +08:00 · 2026-10-29 00:11:15 +08:00 · 2026-10-28 15:26:25 +01:00 · 2026-10-28 22:26:18 +08:00 · 2026-10-28 22:26:03 +08:00 · 2026-10-28 22:22:00 +08:00
107 changed files with 3075 additions and 83 deletions
@@ -0,0 +1,85 @@
+# Publish accepted milestones to the git host.
+#
+# On a push to main that changes docs/milestones/ or docs/sqa/reviews/, this
+# workflow runs `sync-project.sh --accepted-only --apply`: every milestone whose
+# latest Version History row is Accepted and whose latest review record is a Go
+# is created or updated as a Milestone with its tasks as Issues. A Proposed
+# milestone, or one without a Go review, is skipped. Nothing is closed or
+# deleted. Each run updates the issues of every accepted milestone (the sync is
+# idempotent, but the host records the edit).
+#
+# INSTALLING THIS FILE IS THE STANDING AUTHORIZATION TO PUBLISH. A project opts
+# in by copying it; an agent does not install it on its own.
+#
+# Copy it to .gitea/workflows/sync-milestones.yml (Gitea Actions) or
+# .github/workflows/sync-milestones.yml (GitHub Actions). Then:
+#   1. Create a CI secret and put its name in the two `secrets.` lines below
+#      (TOKEN_FOR_GITEA_ISSUES is the name the
+#      framework uses). Gitea does not accept secret names that
+#      start with GITEA_ or GITHUB_.
+#   2. Give the token the smallest scope that works: read and write on issues
+#      and milestones. If the framework submodule is a private repository, the
+#      token must also be able to read it (it is used to check it out).
+#   3. Run it once by hand from the Actions page with dry_run left on: it prints
+#      the plan and creates nothing. Turn dry_run off only for a real run.
+# A runner must be registered for the repository (or its organisation).
+#
+# The steps are one command each: the logic is framework/scripts/ci-publish-milestones.sh,
+# a file that is linted and tested (QC-WF-001 criterion 20).
+
+name: Publish accepted milestones
+
+on:
+  push:
+    branches: [main]
+    paths:
+      - "docs/milestones/**"
+      - "docs/sqa/reviews/**"
+  workflow_dispatch:
+    inputs:
+      dry_run:
+        description: "Preview only: print the plan and create nothing"
+        type: boolean
+        default: true
+
+# Two merges in a row queue instead of racing.
+concurrency:
+  group: publish-accepted-milestones
+  cancel-in-progress: false
+
+jobs:
+  publish:
+    permissions:
+      contents: read
+    runs-on: ubuntu-latest
+    timeout-minutes: 15
+    steps:
+      - name: Check out the project
+        uses: actions/checkout@v4.2.2
+        with:
+          token: ${{ secrets.TOKEN_FOR_GITEA_ISSUES }}
+
+      # The framework submodule is registered with an SSH address, which the runner
+      # cannot use. The address is rewritten to HTTPS for this step only, and the token
+      # goes in through the environment, never on a command line. .gitmodules stays.
+      - name: Fetch the framework submodule over HTTPS
+        env:
+          GIT_CONFIG_COUNT: "2"
+          GIT_CONFIG_KEY_0: url.https://git.tirsystem.com/.insteadOf
+          GIT_CONFIG_VALUE_0: ssh://git@git.tirsystem.com:10022/
+          GIT_CONFIG_KEY_1: http.https://git.tirsystem.com/.extraheader
+          GIT_CONFIG_VALUE_1: "Authorization: token ${{ secrets.TOKEN_FOR_GITEA_ISSUES }}"
+        run: git submodule update --init --recursive
+
+      - name: Make sure curl and jq are available
+        run: bash framework/scripts/ci-publish-milestones.sh ensure-tools
+
+      - name: Publish accepted milestones
+        env:
+          # Gitea reads TOKEN_FOR_GITEA_ISSUES, GitHub's gh reads GH_TOKEN; both
+          # come from the one secret named above.
+          TOKEN_FOR_GITEA_ISSUES: ${{ secrets.TOKEN_FOR_GITEA_ISSUES }}
+          GH_TOKEN: ${{ secrets.TOKEN_FOR_GITEA_ISSUES }}
+          # Empty on a push (a real run); "true" or "false" on a manual run.
+          DRY_RUN: ${{ inputs.dry_run }}
+        run: bash framework/scripts/ci-publish-milestones.sh publish
@@ -43,3 +43,4 @@ Thumbs.db

 # Selenium Chrome profile created by the bot (holds the practice site data)
 /chrome_profile/
+.env.gym
@@ -8,7 +8,7 @@ ENV PYTHONDONTWRITEBYTECODE=1 \
    SNACK_AND_LIFT_E2E=1

 RUN apt-get update \
-    && apt-get install -y --no-install-recommends chromium chromium-driver fonts-liberation fonts-noto-color-emoji \
+    && apt-get install -y --no-install-recommends chromium chromium-driver fonts-liberation fonts-dejavu-core fonts-noto-color-emoji \
    && rm -rf /var/lib/apt/lists/*

 WORKDIR /app
@@ -8,6 +8,7 @@ A Selenium bot that books classes at the fictional gym "Snack and Lift" for you:
 - [Requirements](#-requirements)
 - [Setup](#-setup)
 - [Run](#-run)
+- [Lecture steps and screenshots](#-lecture-steps-and-screenshots)
 - [Tests](#-tests)
 - [License](#-license)
 - [Links](#-links)
@@ -91,6 +92,8 @@ $env:SNACK_AND_LIFT_EMAIL = "yourname@test.com"
 $env:SNACK_AND_LIFT_PASSWORD = "your password"
 ```

+Instead of setting variables by hand, you can put the same lines in a file named `.env.gym` in the project folder (`SNACK_AND_LIFT_EMAIL=...` and `SNACK_AND_LIFT_PASSWORD=...`, one per line). The bot reads it at start; `--env-file PATH` names another file, and a real environment variable wins over the file. The file `.env.gym` is ignored by git: never commit it. Two optional lines, `SNACK_AND_LIFT_ADMIN_EMAIL` and `SNACK_AND_LIFT_ADMIN_PASSWORD`, hold the admin account of the practice site; the bot loads them but does not use them yet.
+
 ### Start the bot

 With the environment active:
@@ -112,6 +115,22 @@ docker build -t snack-and-lift .
 docker run --rm snack-and-lift
 ```

+## 📸 Lecture steps and screenshots
+
+Each step has a wiki page that explains what it teaches, in [English](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/wiki/Home), [Dansk](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/wiki/Home-da) and [Bahasa Melayu](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/wiki/Home-ms). All nine images are screenshots of a real run against the live site. Step 8 ends in SUCCESS; the My Bookings page also lists bookings from earlier runs, which the check prints as "Also booked (not checked)".
+
+| Step | Wiki page | Screenshot |
+| --- | --- | --- |
+| 1 | [Setup and navigation](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/wiki/Step-1) | ![Step 1](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-1.png) |
+| 2 | [Automated login](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/wiki/Step-2) | ![Step 2](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-2.png) |
+| 3 | [Book the next Tuesday class](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/wiki/Step-3) | ![Step 3](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-3.png) |
+| 4 | [Check if a class is booked](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/wiki/Step-4) | ![Step 4](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-4.png) |
+| 5 | [Counters and summary](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/wiki/Step-5) | ![Step 5](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-5.png) |
+| 6 | [Every Tuesday and Thursday class](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/wiki/Step-6) | ![Step 6](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-6.png) |
+| 7 | [Verify on My Bookings](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/wiki/Step-7) | ![Step 7](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-7.png) |
+| 8 | [Time travel QA](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/wiki/Step-8) | ![Step 8](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-8.png) |
+| 9 | [Resilience (retry)](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/wiki/Step-9) | ![Step 9](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-9.png) |
+
 ## 🧪 Tests

 ```bash
@@ -14,9 +14,9 @@ document of a type. `Primary File` may contain a glob (e.g.
 | BC | Business Case | docs/business-case.md | 002 |
 | SA | Stakeholder Analysis | docs/stakeholder-analysis.md | 002 |
 | PP | Project Plan | docs/project-plan.md | 002 |
-| MIL | Milestone / Gateway | docs/milestones/*.md | 009 |
+| MIL | Milestone / Gateway | docs/milestones/*.md | 011 |
 | ADR | Architecture Decision Record | docs/adr/adr-NNNN-*.md | 0004 |
-| RC | SQA Review Record | docs/sqa/reviews/rc-*.md | 021 |
+| RC | SQA Review Record | docs/sqa/reviews/rc-*.md | 028 |
 | TM | Traceability Matrix | docs/sqa/traceability-matrix.md | 002 |
 | DICT | Domain Dictionary | docs/dictionary.md | 002 |

@@ -0,0 +1,74 @@
+# MIL-009 Verification fix: only compare the bookings of the run
+
+## Metadata
+| Key | Value |
+| --- | --- |
+| ID | MIL-009 |
+| CrossReference | [BC-001], [MIL-006] |
+| Language | en |
+| Domain | it |
+
+## Version History
+| Date | Status | Author | Reviewer | Change | Commit |
+| --- | --- | --- | --- | --- | --- |
+| 2026-10-28 | Accepted | Jens Tirsvad Nielsen | AI as stand-in for S02 | Initial version, added after a real run (issue #50) | [3b9e381] |
+
+---
+
+## Purpose
+
+Decide whether the My Bookings check ignores bookings that lie outside the window the bot worked on, so that a real account with bookings from earlier runs ends in SUCCESS, and whether the report never prints a negative count.
+
+## Deliverable
+
+A corrected comparison in `domain/verification.py` and its use in `application/verification.py`, unit tests and an end-to-end fixture with extra future bookings, new screenshots for lecture steps 7 to 9 from a real run, and updated README and wiki notes.
+
+## Go / No-Go Criteria
+
+| # | Criterion (objectively checkable) | Go | No-Go |
+| --- | --- | --- | --- |
+| 1 | Unit tests show that entries outside the window of the run are listed as a note and do not cause MISMATCH | 0 failures; a test uses entries after the last handled class | A test fails or is missing |
+| 2 | Unit tests show that the missing count is never negative and that a class the bot booked but the page lacks is still a MISMATCH | 0 failures | A test fails or is missing |
+| 3 | Unit and container tests of the existing steps still pass with at least 80% coverage of the logic modules | Both true | Either false |
+| 4 | The code review against `qc-programming-python.md` is `Go` | `RC-*` is `Go` | `No-Go` |
+| 5 | `docs/images/step-8.png` from a real run shows SUCCESS | S01 reran the bot with their own account and the image is committed | The image shows MISMATCH or is not replaced |
+
+## Dependencies
+
+| Depends on | Reason |
+| --- | --- |
+| MIL-006 | Changes the verification that MIL-006 built |
+
+## Traceability
+
+| Business Case objective / KPI / user story | Reference |
+| --- | --- |
+| O5, success criterion 3 | [BC-001] |
+| CN-001, CN-003 | [SA-001] |
+
+## Ownership
+
+| Role | Stakeholder ID (SA) |
+| --- | --- |
+| Owner | S01 |
+| Approving reviewer | S02 (AI as stand-in, autonomous mode, not independent) |
+
+## Target Date
+
+2026-11-06 — inside the planned window 2026-10-28 to 2026-11-06 of [PP-001]; the Business Case sets no deadline.
+
+## Tasks
+
+| # | Task | Summary | Needs its own Use Case/User Story? | Reference |
+| --- | --- | --- | --- | --- |
+| 1 | Limit the comparison to the window of the run | In `build_verification` count only page entries dated from the first to the last class the bot handled; list the other entries as "also booked, not checked". The missing count never goes below zero. | No | O5 |
+| 2 | Test the corrected comparison | Unit tests for entries outside the window, a booked class that is missing on the page, and the count; an end-to-end fixture with extra future bookings. | No | O5 |
+| 3 | Refresh the step 7 to 9 screenshots and the notes | S01 reruns the bot with their own account and `--screenshot-dir`; the AI commits the images and removes the MISMATCH note from the README and the step 8 wiki pages in three languages. | No | O5, O8 |
+
+---
+
+[BC-001]: ../business-case.md
+[MIL-006]: ./mil-006-verify-and-time-travel.md
+[PP-001]: ../project-plan.md
+[SA-001]: ../stakeholder-analysis.md
+[3b9e381]: https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/commit/3b9e381cd42c56cc4543869c5b80c8e0014b4948
@@ -0,0 +1,74 @@
+# MIL-010 Gym accounts from a credentials file
+
+## Metadata
+| Key | Value |
+| --- | --- |
+| ID | MIL-010 |
+| CrossReference | [BC-001], [MIL-003] |
+| Language | en |
+| Domain | it |
+
+## Version History
+| Date | Status | Author | Reviewer | Change | Commit |
+| --- | --- | --- | --- | --- | --- |
+| 2026-10-28 | Accepted | Jens Tirsvad Nielsen | AI as stand-in for S02 | Initial version, requested by S01 in chat | [434bdaf] |
+
+---
+
+## Purpose
+
+Decide whether the bot can read the student and the admin account of the practice site from the personal file `.env.gym`, so that S01 does not set environment variables by hand before each run, while the file stays out of the repository and the tests.
+
+## Deliverable
+
+`credentials.py` that loads the student account and, load-only, the admin account from the environment, plus a small reader of the `.env.gym` file in the infrastructure layer that the command line merges under the real environment, unit tests with dummy values only, and README and wiki instructions.
+
+## Go / No-Go Criteria
+
+| # | Criterion (objectively checkable) | Go | No-Go |
+| --- | --- | --- | --- |
+| 1 | Unit tests show that the student and the admin account load from file values, and that a real environment variable wins over the file | 0 failures | A test fails or is missing |
+| 2 | Unit tests show that a missing or empty variable raises an error that names the variable and never prints a value, and that the admin account is optional | 0 failures | A test fails or is missing |
+| 3 | No test reads `.env` or `.env.gym`; `.env.gym` is in `.gitignore`; no password appears in the repository | `git grep` finds no real credential; tests use `tmp_path` | A real file is read or a credential is committed |
+| 4 | All earlier tests pass with at least 80% coverage of the logic modules | Both true | Either false |
+| 5 | The code review against `qc-programming-python.md` is `Go` | `RC-*` is `Go` | `No-Go` |
+
+## Dependencies
+
+| Depends on | Reason |
+| --- | --- |
+| MIL-003 | Changes the credentials handling that MIL-003 built |
+
+## Traceability
+
+| Business Case objective / KPI / user story | Reference |
+| --- | --- |
+| O2, O7 | [BC-001] |
+| CN-004, CN-005, CN-009 | [SA-001] |
+
+## Ownership
+
+| Role | Stakeholder ID (SA) |
+| --- | --- |
+| Owner | S01 |
+| Approving reviewer | S02 (AI as stand-in, autonomous mode, not independent) |
+
+## Target Date
+
+2026-11-06 — inside the planned window 2026-10-28 to 2026-11-06 of [PP-001]; the Business Case sets no deadline.
+
+## Tasks
+
+| # | Task | Summary | Needs its own Use Case/User Story? | Reference |
+| --- | --- | --- | --- | --- |
+| 1 | Load the student and the admin account | In `credentials.py` read `SNACK_AND_LIFT_EMAIL` and `SNACK_AND_LIFT_PASSWORD` (required) and `SNACK_AND_LIFT_ADMIN_EMAIL` and `SNACK_AND_LIFT_ADMIN_PASSWORD` (optional, load-only; nothing uses the admin yet). Errors name the variables, never a value. | No | O2 |
+| 2 | Read the `.env.gym` file | Add a small reader (no new dependency) in `infrastructure/` for `KEY=VALUE` lines; the command line merges the file under the real environment, and a missing file is not an error. | No | O2, O7 |
+| 3 | Test and document | Unit tests with dummy values in `tmp_path`; update the README and the Getting Started wiki pages (three languages) and keep `.env.gym` in `.gitignore`. | No | O7, O8 |
+
+---
+
+[BC-001]: ../business-case.md
+[MIL-003]: ./mil-003-login.md
+[PP-001]: ../project-plan.md
+[SA-001]: ../stakeholder-analysis.md
+[434bdaf]: https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/commit/434bdafefdf0a95eed175dc89f27fbd91b19c91f
@@ -4,20 +4,21 @@
 | Key | Value |
 | --- | --- |
 | ID | PP-001 |
-| CrossReference | [BC-001], [SA-001], [MIL-001], [MIL-002], [MIL-003], [MIL-004], [MIL-005], [MIL-006], [MIL-007], [MIL-008] |
+| CrossReference | [BC-001], [SA-001], [MIL-001], [MIL-002], [MIL-003], [MIL-004], [MIL-005], [MIL-006], [MIL-007], [MIL-008], [MIL-009], [MIL-010] |
 | Language | en |
 | Domain | it |

 ## Version History
 | Date | Status | Author | Reviewer | Change | Commit |
 | --- | --- | --- | --- | --- | --- |
-| 2026-10-28 | Accepted | Jens Tirsvad Nielsen | S01 | Initial version: eight gateways | [50e8331] |
+| 2026-10-28 | Accepted | Jens Tirsvad Nielsen | S01 | Add MIL-009, a verification fix found in a real run (S01 accepted the plan in chat) | [3b9e381] |
+| 2026-10-28 | Accepted | Jens Tirsvad Nielsen | S01 | Add MIL-010, gym accounts from a credentials file (S01 asked for it in chat) | [434bdaf] |

 ---

 ## Purpose

-This plan schedules eight gateways (`MIL-001` to `MIL-008`) that take the Snack and Lift booking bot from an accepted baseline to a documented, tested and published repository. The Business Case sets no deadline, so the window 2026-10-28 to 2026-11-06 is a planning assumption of this plan. Each gateway ends in a decision and a deliverable and is delivered by one pull request on its own branch.
+This plan schedules ten gateways (`MIL-001` to `MIL-010`) that take the Snack and Lift booking bot from an accepted baseline to a documented, tested and published repository. The Business Case sets no deadline, so the window 2026-10-28 to 2026-11-06 is a planning assumption of this plan. Each gateway ends in a decision and a deliverable and is delivered by one pull request on its own branch.

 ## Planning Assumptions

@@ -39,6 +40,8 @@ This plan schedules eight gateways (`MIL-001` to `MIL-008`) that take the Snack
 | Lecture steps 7 and 8 | [MIL-006] | 2026-11-01 to 2026-11-02 | 2026-11-02 | S01 | none (technical tasks) | My Bookings verification, time travel, screenshots 7 and 8 | |
 | Lecture step 9 | [MIL-007] | 2026-11-02 to 2026-11-04 | 2026-11-04 | S01 | none (technical tasks) | `retry`, resilient `login`, `book_class`, `get_my_bookings`, screenshot 9 | |
 | Documentation, wiki and release | [MIL-008] | 2026-11-04 to 2026-11-06 | 2026-11-06 | S01 | none (documentation) | README, Doxygen, wiki in en, da and ms, final checks | |
+| Verification fix | [MIL-009] | 2026-11-04 to 2026-11-06 | 2026-11-06 | S01 | none (technical tasks) | Verification compares only the window of the run; new screenshots 7 to 9 | |
+| Gym accounts from a file | [MIL-010] | 2026-11-04 to 2026-11-06 | 2026-11-06 | S01 | none (technical tasks) | `.env.gym` credentials for the student and the admin account | |

 ```plantuml
@startgantt
@@ -77,7 +80,7 @@ Project starts 2026-10-28
 ## Dependencies

 ```
-MIL-001 → MIL-002 → MIL-003 → MIL-004 → MIL-005 → MIL-006 → MIL-007 → MIL-008
+MIL-001 → MIL-002 → MIL-003 → MIL-004 → MIL-005 → MIL-006 → MIL-007 → MIL-008 → MIL-009 → MIL-010
 ```

 A No-Go on a gateway stops all later gateways until it is fixed and re-reviewed. The end date then moves by the days spent on the rework; S01 decides a move of the end date.
@@ -109,4 +112,8 @@ A No-Go on a gateway stops all later gateways until it is fixed and re-reviewed.
 [MIL-006]: ./milestones/mil-006-verify-and-time-travel.md
 [MIL-007]: ./milestones/mil-007-resilience.md
 [MIL-008]: ./milestones/mil-008-documentation.md
+[MIL-009]: ./milestones/mil-009-verification-fix.md
+[MIL-010]: ./milestones/mil-010-credentials-file.md
 [50e8331]: https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/commit/50e8331bd21910a72ae56b74c9ff6a3677a709b3
+[3b9e381]: https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/commit/3b9e381cd42c56cc4543869c5b80c8e0014b4948
+[434bdaf]: https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/commit/434bdafefdf0a95eed175dc89f27fbd91b19c91f
@@ -0,0 +1,62 @@
+# SQA Review Record
+
+## Metadata
+| Key | Value |
+| --- | --- |
+| ID | RC-021 |
+| CrossReference | [MIL-007], [QC-PY-001] |
+
+## Version History
+| Date | Status | Author | Reviewer | Change | Commit |
+| --- | --- | --- | --- | --- | --- |
+| 2026-10-28 | Accepted | Jens Tirsvad Nielsen | AI as stand-in for S02 | Review record, full review of the source and tests of MIL-007 | [5d28e68] |
+
+---
+
+## Artifact Under Review
+
+- Instance reviewed: [MIL-007]
+- Checklist used: [QC-PY-001] (`QC-PY-001`)
+- Scope: full review of the source and tests of MIL-007
+- Language and domain: n/a
+- Language reviewer: none
+- Mode: autonomous, not independent (the reviewer is `AI as stand-in for S02`)
+
+## Checklist Results
+
+| # | Criterion | Status | Evidence/Notes |
+| --- | --- | --- | --- |
+| 1 | Packages, modules, functions, variables, classes and constants follow PEP 8 casing (`snake_case`, `PascalCase`, `UPPER_SNAKE`) | Pass | Modules, functions and variables are `snake_case`, classes `PascalCase`, constants `UPPER_SNAKE`; `ruff check` with the N rules passes. |
+| 2 | Names state purpose in the domain's language; no unexplained abbreviations, no single-letter names outside tiny scopes | Pass | Names state purpose (for example `retry`, `login`, `book_class`, `get_my_bookings`); no single-letter names outside comprehensions. |
+| 3 | Code is produced by the project's formatter and passes its linter with no unexplained suppressions | Pass | `ruff format` leaves 0 files changed and `ruff check` reports 0 findings; there is no suppression. |
+| 4 | Every function and method signature is type-annotated, including `-> None` | Pass | Every function and method has annotations (ruff ANN rules pass), including `-> None`. |
+| 5 | No bare `except:`, no swallowed exceptions; specific exceptions are raised and the cause is kept (`raise ... from`) | Pass | No bare `except`; specific exceptions are caught, and re-raised with `raise ... from` where the cause matters. |
+| 6 | No mutable default arguments and no shadowed builtins | Pass | No mutable default arguments; no builtin is shadowed (ruff B and A-style checks pass). |
+| 7 | Files, locks and connections are managed with context managers | Pass | Files and the browser are managed with `with`/`try-finally`; no unmanaged file handle exists. |
+| 8 | Public modules, classes and functions have docstrings that say what, not how | Pass | Public modules, classes and functions have Doxygen docstrings (`@brief`, `@param`, `@return`). |
+| 9 | Logging uses `logging`, not `print`; no secrets or personal data in log output | Pass | No print calls; messages go through `logging`. |
+| 10 | Classes and operations trace to the Design Class Diagram they implement; deviations are recorded | N-A | No Design Class Diagram exists; ADR-0001 records that design artifacts are skipped. Classes follow ADR-0002. |
+| 11 | Tests exist for new behaviour, are named for the behaviour, and do not depend on order or the network | Pass | 237 unit tests (retry, resilient login, booking and bookings read, a seeded chaos test at 50% failure: 97 of 100 seeds complete, seeds 0 to 17 must all pass) need no browser or network; 20 container tests run the adapter against failing local fixtures; coverage of the non-infrastructure modules is 100%. |
+| 12 | Type checker runs in strict mode without errors; `Any` is justified in a comment | N-A | No strict type checker is configured; annotations are complete and checked by ruff. |
+| 13 | Dependencies are declared and pinned in the project's dependency file, none unused | Pass | `pyproject.toml` declares Selenium (runtime) and pytest, pytest-cov, ruff (dev) with lower bounds; none is unused. |
+| 14 | Each piece of knowledge (a business rule, constant, format, validation or query) is defined in one place; code is merged only where it expresses the same knowledge, not where it merely looks alike | Pass | Each rule, constant and message exists once (constants in `constants.py`, rules in `domain/`). |
+| 15 | Imports point inward: business-rule modules import no framework, database, UI or delivery-mechanism code, and there are no import cycles | Pass | `domain` and `application` import no Selenium; only `infrastructure` imports it; no import cycle exists. |
+| 16 | SOLID holds: a class or module has one reason to change; new behaviour is added by extension, not by editing a type test in several places; subclasses honour the contract of their base; protocols are narrow; high-level code depends on protocols or abstract types and receives its implementations | Pass | One reason to change per module; the adapter hides Selenium behind a port; fakes honour the port contract. |
+| 17 | Source is laid out in the layers and folders of the clean architecture folder structure (`domain`, `application`, `infrastructure`, `presentation` packages under `src/`, and `unit`, `integration` and `e2e` under `tests/`), in `snake_case`, and imports between the layers point inward | Pass | Layout follows `src/snack_and_lift/{domain,application,infrastructure}` and `tests/{unit,integration,e2e}`; there is no presentation layer for a command-line bot (ADR-0002). |
+| 18 | File and folder names are English, contain no space and no character outside letters, digits, `-`, `_` and `.`, and follow the casing of the language for files | Pass | All file and folder names are English lowercase with `_`, `-` and `.` only. |
+
+## Overall Verdict
+
+Go-with-conditions - the code of MIL-007 meets every Mandatory criterion of QC-PY-001 and there is no blocking finding, but part of the gate evidence cannot be produced by the stand-in. Gate criteria 1 (retry unit tests) and 2 (a booking only counts when the page confirms it) are met in tests. Criterion 3 (three container runs on the live site at 50% simulated failure) is not verified, and criterion 4 is not met: `step-9.png` comes from a run with an account. The framework forbids the stand-in to enter passwords or register on a third-party site, and the network simulation is an admin control behind a login. Known behaviour: a wrong password or an empty bookings page is also retried seven times (about 3 seconds) before exit code 1. The review is not independent: the AI wrote and reviewed it as stand-in for S02, and S01 can challenge it in the end-of-run summary.
+
+## Action Items
+
+| Action | Owner | Due |
+| --- | --- | --- |
+| Enable the network simulation at 50% as admin, run the bot three times with your own account and keep `docs/images/step-9.png` | S01 | when S01 has time |
+
+---
+
+[MIL-007]: ../../milestones/mil-007-resilience.md
+[QC-PY-001]: ../../../framework/qc/code/qc-programming-python.md
+[5d28e68]: https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/commit/5d28e682b41e1a03b0cac1a6b2ea4977750d45b8
@@ -0,0 +1,69 @@
+# SQA Review Record
+
+## Metadata
+| Key | Value |
+| --- | --- |
+| ID | RC-022 |
+| CrossReference | [MIL-008], [QC-WF-001] |
+
+## Version History
+| Date | Status | Author | Reviewer | Change | Commit |
+| --- | --- | --- | --- | --- | --- |
+| 2026-10-28 | Accepted | Jens Tirsvad Nielsen | AI as stand-in for S02 | Review record, the workflows publish-wiki.yml and sync-milestones.yml of MIL-008 task 5 | [0301fd2] |
+
+---
+
+## Artifact Under Review
+
+- Instance reviewed: [MIL-008]
+- Checklist used: [QC-WF-001] (`QC-WF-001`)
+- Scope: the workflows publish-wiki.yml and sync-milestones.yml of MIL-008 task 5
+- Language and domain: n/a
+- Language reviewer: none
+- Mode: autonomous, not independent (the reviewer is `AI as stand-in for S02`)
+
+## Checklist Results
+
+| # | Criterion | Status | Evidence/Notes |
+| --- | --- | --- | --- |
+| 1 | The file meets the YAML checklist ([QC-YML-001]) | Pending | `yamllint` is not installed on the stand-in's machine, so the YAML checklist QC-YML-001 was not run; the files parse (actionlint reads them) and use 2-space indentation. |
+| 2 | The file is in `.gitea/workflows/` or `.github/workflows/` with the extension `.yml` or `.yaml`, and the same workflow does not exist in both directories | Pass | Both files are in `.gitea/workflows/` with the extension `.yml`; neither exists in `.github/workflows/`. |
+| 3 | The file declares `name`, `on` and `jobs`; `on` lists explicit events and, where it matters, `branches` or `paths` | Pass | Both files declare `name`, `on` and `jobs`; `on` lists `push` on `main` with `paths`, and `workflow_dispatch`. |
+| 4 | A header comment states the purpose, the triggers, the inputs, the secrets it needs, what it changes outside the repository, and the decision or milestone it implements | Pass | Both headers state the purpose, triggers, secrets, what changes outside the repository and the installation steps (copied from the framework template); the decision is MIL-008 task 5. |
+| 5 | Every job and step has a descriptive `name`, unique within its job | Pass | `sync-milestones.yml` names every step. `publish-wiki.yml` has one job that calls a reusable workflow and has no steps of its own. |
+| 6 | The file passes `actionlint` (or the project's workflow validation test) and the tool version is stated | Pass | `actionlint` 1.7.12 reports no finding for both files (exit code 0). |
+| 7 | Every job declares `permissions` explicitly with the least access that works (`contents: read` or none by default); write access only on the job that needs it | Pass | `sync-milestones.yml` sets `contents: read` on its job; `publish-wiki.yml` sets `contents: read` for the whole workflow (added to the framework template). The wiki write uses the token secret. |
+| 8 | Every action used is pinned to a full commit SHA or an exact release tag, never to a branch; a moving major tag (`@v1`) is accepted only for an action published by the same organisation under a stated release policy | Pass | `actions/checkout@v4.2.2` is an exact release tag. `publish-wiki.yml` uses `TirSystem/git-workflows/.gitea/workflows/publish-wiki-pages.yml@v1`, a moving major tag of a reusable workflow of the same organisation, which the framework template accepts. |
+| 9 | No secret is written in the file; a secret is read as `secrets.NAME`, passed to a step through `env:`, never placed on a command line or echoed | Pass | Both files read `secrets.SYNC_TOKEN` and `secrets.TOKEN_FOR_GITEA_WIKI` only; no secret value is in the repository. |
+| 10 | Untrusted text (`inputs.*`, event fields such as a title, branch name or label, step outputs derived from them) is never interpolated into a `run:` script; it is passed through `env:` and used as a quoted variable | Pass | No `run:` block interpolates `inputs.*` or event text; `DRY_RUN` is passed through `env:`. |
+| 11 | A trigger that gives a run secrets or a write token on code from a fork or an untrusted branch (`pull_request_target`, `workflow_run`) does not check out or execute that code | Pass | Neither workflow uses `pull_request_target` or `workflow_run`. |
+| 12 | A workflow that changes state outside the repository (release, tag, publish, sync, deploy) is started by hand or by a protected branch, and a manual run defaults to a dry run with an explicit input to apply | Pass | Both run on a push to `main` or by hand; the manual run of `sync-milestones.yml` defaults to a dry run. `publish-wiki.yml` has no dry run (it is a call of the reusable workflow). |
+| 13 | The token the workflow uses has only the scope the job needs, and where the host's built-in token is enough it is used instead of a personal token | Pending | The scope of both tokens is set by S01 when the secrets are created. |
+| 14 | Every job sets `runs-on` to an explicit label and `timeout-minutes` | Pass | `sync-milestones.yml` sets `runs-on: ubuntu-latest` and `timeout-minutes: 15`. `publish-wiki.yml` calls a reusable workflow, where the called workflow sets them (a calling job cannot). |
+| 15 | `concurrency` is set where two runs would conflict: a check cancels the older run (`cancel-in-progress: true`); a publishing job queues (`cancel-in-progress: false`) | Pass | `sync-milestones.yml` sets `concurrency` with `cancel-in-progress: false`; `publish-wiki.yml` leaves it to the reusable workflow. |
+| 16 | Triggers are filtered (`branches`, `paths`, `types`) so the workflow runs only for the changes it concerns | Pass | Both are filtered by `branches: [main]` and `paths`. |
+| 17 | Tools and dependencies are installed at pinned versions; a version is a workflow input with a default or a value in one place | Pass | `sync-milestones.yml` installs nothing itself; `curl` and `jq` are checked by the framework script. The wiki workflow installs nothing. |
+| 18 | A dependency cache is used where the install is slow enough to matter | N-A | No dependency is installed that a cache would speed up. |
+| 19 | A failing step names the tool, the file and the reason in its log (annotation or `::error::` line) | Pending | Not verified: the log messages come from the framework script and the reusable workflow, which the stand-in did not run. |
+| 20 | A `run:` block is within the limit on embedded program text of criterion 7 of [QC-YML-001]; a longer script is a file (`tools/`, or a shared repository fetched at a pinned ref) that the step calls with one command, and the workflow passes its settings through `env:` | Pass | Each `run:` block is one command; the logic is in `framework/scripts/ci-publish-milestones.sh`. |
+| 21 | Logic used by more than one workflow or repository is a reusable workflow (`workflow_call`) or a composite action; each input has a `description`, a `type` and a `default` where one is sensible | Pass | `publish-wiki.yml` calls the reusable workflow; `sync-milestones.yml` calls one framework script. The inputs of `workflow_dispatch` have a description, type and default. |
+| 22 | Every workflow feature used is verified on the instance that runs it, and the host and version are recorded; a feature the target hosts differ on is noted in the header | Pending | Neither workflow has run on the Gitea instance yet: the secrets do not exist and a merge to `main` is what triggers the first run. |
+| 23 | A reusable workflow states, in its header, how a caller pins it and the script it fetches, so the workflow and its script are the same release | N-A | Neither file is a reusable workflow; both are callers. |
+| 24 | A static test of the repository's workflows exists and runs on every pull request: valid YAML, required keys, explicit permissions and runner, no inline script above the limit | N-A | No static test of the workflows exists in this project; the checks are `actionlint` on request. |
+
+## Overall Verdict
+
+Go-with-conditions - the two workflows are the framework's templates, copied into `.gitea/workflows/`, and pass `actionlint`. Mandatory criteria 1 (the YAML checklist: `yamllint` was not run) and 22 (feature checked on the instance) are Pending, because the stand-in cannot create the secrets `TOKEN_FOR_GITEA_WIKI` and `SYNC_TOKEN` or watch a run on the host. The workflows will fail or skip until S01 creates them. The review is not independent: the AI wrote and reviewed it as stand-in for S02, and S01 can challenge it in the end-of-run summary.
+
+## Action Items
+
+| Action | Owner | Due |
+| --- | --- | --- |
+| Create the secrets `TOKEN_FOR_GITEA_WIKI` (wiki write) and `SYNC_TOKEN` (issues and milestones) in the organisation, enable the wiki, then run both workflows once by hand | S01 | before the next merge to main |
+| Run `yamllint` on `.gitea/workflows/` | S01 | when convenient |
+
+---
+
+[MIL-008]: ../../milestones/mil-008-documentation.md
+[QC-WF-001]: ../../../framework/qc/code/qc-workflow.md
+[0301fd2]: https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/commit/0301fd2f0bc22a26ee8d71ec41dab2bce0d38dc3
@@ -0,0 +1,65 @@
+# SQA Review Record
+
+## Metadata
+| Key | Value |
+| --- | --- |
+| ID | RC-023 |
+| CrossReference | [MIL-009], [QC-MIL-001] |
+
+## Version History
+| Date | Status | Author | Reviewer | Change | Commit |
+| --- | --- | --- | --- | --- | --- |
+| 2026-10-28 | Accepted | Jens Tirsvad Nielsen | AI as stand-in for S02 | Review record, full review | [3b9e381] |
+
+---
+
+## Artifact Under Review
+
+- Instance reviewed: [MIL-009]
+- Checklist used: [QC-MIL-001] (`QC-MIL-001`)
+- Scope: full review
+- Language and domain: en / it
+- Language reviewer: none
+- Mode: autonomous, not independent (the reviewer is `AI as stand-in for S02`)
+
+## Checklist Results
+
+| # | Criterion | Status | Evidence/Notes |
+| --- | --- | --- | --- |
+| 1 | A concrete deliverable is defined for every gate | Pass | Deliverable section names concrete files, runs or documents. |
+| 2 | Explicit Go/No-Go criteria are stated for each gate | Pass | Go/No-Go table has objectively checkable criteria with Go and No-Go columns. |
+| 3 | Dependencies on other milestones are explicitly mapped | Pass | Dependencies table names the preceding milestone and the reason. |
+| 4 | Each milestone is traceable to a Business Case objective or KPI | Pass | Traceability cites O5 of BC-001 and the concerns of SA-001. |
+| 5 | Milestone owner and approving reviewer are identified | Pass | Ownership names S01 as owner and S02 (AI as stand-in) as approving reviewer. |
+| 6 | Milestone has a defined target date consistent with project constraints | Pass | Target Date lies inside the window 2026-10-28 to 2026-11-06 of PP-001. |
+
+## Language and Domain Results
+
+| # | Criterion | Status | Evidence/Notes |
+| --- | --- | --- | --- |
+| 1 | The Metadata table has a `Language` row and a `Domain` row, and neither is a placeholder | Pass | Metadata has `Language` = en and `Domain` = it. |
+| 2 | `Language` is a BCP 47 code and `Domain` is a value from the registry's domain list | Pass | `en` is BCP 47; `it` is in the registry domain list. |
+| 3 | The content (prose and table cells) is written in the stated language | Pass | All prose and table cells of MIL-009 are English (stand-in assessment, see the verdict). |
+| 4 | The register matches the one the registry gives for the artifact type | Pass | Register of the type is given in the registry (Languages); the text keeps to it. |
+| 5 | Domain terms are the PO terms of the domain's dictionary, with no synonyms | Pass | Terms follow DICT-001: bot, booking, verification (stand-in assessment, see the verdict). |
+| 6 | Metadata keys, section headings, IDs and statuses are in English | Pass | Keys, headings, IDs and statuses are English. |
+| 7 | No translated twin (`<name>.<language>.md`) exists beside the document | Pass | No translated twin exists in `docs/`. |
+| 8 | A change of language or domain since the previous accepted version has a Version History row and was reviewed again | N-A | First version; no language change. |
+| 9 | A reviewer competent in the domain, and in the language, has confirmed that the domain terms are used correctly | Pass | The stand-in reviewer reads English and checked the IT terms in context (stand-in assessment, see the verdict). |
+| 10 | Abbreviations are spelled out on first use, in the stated language | Pass | Abbreviations are spelled out or defined where first used, or are the framework's own IDs. |
+
+## Overall Verdict
+
+Go — MIL-009 meets every Mandatory criterion. No blocking finding. The review is not independent: the AI authored and reviewed it as stand-in for S02, and S01 can challenge it in the end-of-run summary. Criteria 3, 5 and 9 of QC-LANG-001 were assessed by the stand-in only; the framework would hold them `Pending` for an independent language reviewer. S01 is asked to confirm them in the end-of-run summary.
+
+## Action Items
+
+| Action | Owner | Due |
+| --- | --- | --- |
+| None | S01 | - |
+
+---
+
+[MIL-009]: ../../milestones/mil-009-verification-fix.md
+[QC-MIL-001]: ../../../framework/qc/artifacts/qc-milestones-gateways.md
+[3b9e381]: https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/commit/3b9e381cd42c56cc4543869c5b80c8e0014b4948
@@ -0,0 +1,62 @@
+# SQA Review Record
+
+## Metadata
+| Key | Value |
+| --- | --- |
+| ID | RC-024 |
+| CrossReference | [MIL-009], [QC-PY-001] |
+
+## Version History
+| Date | Status | Author | Reviewer | Change | Commit |
+| --- | --- | --- | --- | --- | --- |
+| 2026-10-28 | Accepted | Jens Tirsvad Nielsen | AI as stand-in for S02 | Review record, full review of the source and tests of MIL-009 | [52427ca] |
+
+---
+
+## Artifact Under Review
+
+- Instance reviewed: [MIL-009]
+- Checklist used: [QC-PY-001] (`QC-PY-001`)
+- Scope: full review of the source and tests of MIL-009
+- Language and domain: n/a
+- Language reviewer: none
+- Mode: autonomous, not independent (the reviewer is `AI as stand-in for S02`)
+
+## Checklist Results
+
+| # | Criterion | Status | Evidence/Notes |
+| --- | --- | --- | --- |
+| 1 | Packages, modules, functions, variables, classes and constants follow PEP 8 casing (`snake_case`, `PascalCase`, `UPPER_SNAKE`) | Pass | Modules, functions and variables are `snake_case`, classes `PascalCase`, constants `UPPER_SNAKE`; `ruff check` with the N rules passes. |
+| 2 | Names state purpose in the domain's language; no unexplained abbreviations, no single-letter names outside tiny scopes | Pass | Names state purpose (for example `retry`, `login`, `book_class`, `get_my_bookings`); no single-letter names outside comprehensions. |
+| 3 | Code is produced by the project's formatter and passes its linter with no unexplained suppressions | Pass | `ruff format` leaves 0 files changed and `ruff check` reports 0 findings; there is no suppression. |
+| 4 | Every function and method signature is type-annotated, including `-> None` | Pass | Every function and method has annotations (ruff ANN rules pass), including `-> None`. |
+| 5 | No bare `except:`, no swallowed exceptions; specific exceptions are raised and the cause is kept (`raise ... from`) | Pass | No bare `except`; specific exceptions are caught, and re-raised with `raise ... from` where the cause matters. |
+| 6 | No mutable default arguments and no shadowed builtins | Pass | No mutable default arguments; no builtin is shadowed (ruff B and A-style checks pass). |
+| 7 | Files, locks and connections are managed with context managers | Pass | Files and the browser are managed with `with`/`try-finally`; no unmanaged file handle exists. |
+| 8 | Public modules, classes and functions have docstrings that say what, not how | Pass | Public modules, classes and functions have Doxygen docstrings (`@brief`, `@param`, `@return`). |
+| 9 | Logging uses `logging`, not `print`; no secrets or personal data in log output | Pass | No print calls; messages go through `logging`. |
+| 10 | Classes and operations trace to the Design Class Diagram they implement; deviations are recorded | N-A | No Design Class Diagram exists; ADR-0001 records that design artifacts are skipped. Classes follow ADR-0002. |
+| 11 | Tests exist for new behaviour, are named for the behaviour, and do not depend on order or the network | Pass | 249 unit tests need no browser or network (new: entries outside the window, inclusive ends, missing class, unexpected entry, no results, New Year wrap-around, invalid dates); 21 container tests include a fixture with extra future bookings; coverage of the non-infrastructure modules is 100%. |
+| 12 | Type checker runs in strict mode without errors; `Any` is justified in a comment | N-A | No strict type checker is configured; annotations are complete and checked by ruff. |
+| 13 | Dependencies are declared and pinned in the project's dependency file, none unused | Pass | `pyproject.toml` declares Selenium (runtime) and pytest, pytest-cov, ruff (dev) with lower bounds; none is unused. |
+| 14 | Each piece of knowledge (a business rule, constant, format, validation or query) is defined in one place; code is merged only where it expresses the same knowledge, not where it merely looks alike | Pass | Each rule, constant and message exists once (constants in `constants.py`, rules in `domain/`). |
+| 15 | Imports point inward: business-rule modules import no framework, database, UI or delivery-mechanism code, and there are no import cycles | Pass | `domain` and `application` import no Selenium; only `infrastructure` imports it; no import cycle exists. |
+| 16 | SOLID holds: a class or module has one reason to change; new behaviour is added by extension, not by editing a type test in several places; subclasses honour the contract of their base; protocols are narrow; high-level code depends on protocols or abstract types and receives its implementations | Pass | One reason to change per module; the adapter hides Selenium behind a port; fakes honour the port contract. |
+| 17 | Source is laid out in the layers and folders of the clean architecture folder structure (`domain`, `application`, `infrastructure`, `presentation` packages under `src/`, and `unit`, `integration` and `e2e` under `tests/`), in `snake_case`, and imports between the layers point inward | Pass | Layout follows `src/snack_and_lift/{domain,application,infrastructure}` and `tests/{unit,integration,e2e}`; there is no presentation layer for a command-line bot (ADR-0002). |
+| 18 | File and folder names are English, contain no space and no character outside letters, digits, `-`, `_` and `.`, and follow the casing of the language for files | Pass | All file and folder names are English lowercase with `_`, `-` and `.` only. |
+
+## Overall Verdict
+
+Go-with-conditions - the code of MIL-009 meets every Mandatory criterion of QC-PY-001 and there is no blocking finding, but part of the gate evidence cannot be produced by the stand-in. Gate criteria 1 to 3 are met in tests. Criterion 5 is not met yet: `docs/images/step-8.png` still shows the MISMATCH of the run before the fix, because a new image needs a run with S01's own account, which the stand-in may not log in to. The README and the wiki say that the image predates the fix. The review is not independent: the AI wrote and reviewed it as stand-in for S02, and S01 can challenge it in the end-of-run summary.
+
+## Action Items
+
+| Action | Owner | Due |
+| --- | --- | --- |
+| Rerun `python -m snack_and_lift --screenshot-dir docs/images` with your own account and tell the AI, which then commits the new images and removes the note | S01 | when S01 has time |
+
+---
+
+[MIL-009]: ../../milestones/mil-009-verification-fix.md
+[QC-PY-001]: ../../../framework/qc/code/qc-programming-python.md
+[52427ca]: https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/commit/52427ca69cc0193ba9734b548f31a33549606249
@@ -0,0 +1,65 @@
+# SQA Review Record
+
+## Metadata
+| Key | Value |
+| --- | --- |
+| ID | RC-025 |
+| CrossReference | [MIL-010], [QC-MIL-001] |
+
+## Version History
+| Date | Status | Author | Reviewer | Change | Commit |
+| --- | --- | --- | --- | --- | --- |
+| 2026-10-28 | Accepted | Jens Tirsvad Nielsen | AI as stand-in for S02 | Review record, full review | [434bdaf] |
+
+---
+
+## Artifact Under Review
+
+- Instance reviewed: [MIL-010]
+- Checklist used: [QC-MIL-001] (`QC-MIL-001`)
+- Scope: full review
+- Language and domain: en / it
+- Language reviewer: none
+- Mode: autonomous, not independent (the reviewer is `AI as stand-in for S02`)
+
+## Checklist Results
+
+| # | Criterion | Status | Evidence/Notes |
+| --- | --- | --- | --- |
+| 1 | A concrete deliverable is defined for every gate | Pass | Deliverable section names concrete files, runs or documents. |
+| 2 | Explicit Go/No-Go criteria are stated for each gate | Pass | Go/No-Go table has objectively checkable criteria with Go and No-Go columns. |
+| 3 | Dependencies on other milestones are explicitly mapped | Pass | Dependencies table names the preceding milestone and the reason. |
+| 4 | Each milestone is traceable to a Business Case objective or KPI | Pass | Traceability cites O2 and O7 of BC-001 and the concerns of SA-001. |
+| 5 | Milestone owner and approving reviewer are identified | Pass | Ownership names S01 as owner and S02 (AI as stand-in) as approving reviewer. |
+| 6 | Milestone has a defined target date consistent with project constraints | Pass | Target Date lies inside the window 2026-10-28 to 2026-11-06 of PP-001. |
+
+## Language and Domain Results
+
+| # | Criterion | Status | Evidence/Notes |
+| --- | --- | --- | --- |
+| 1 | The Metadata table has a `Language` row and a `Domain` row, and neither is a placeholder | Pass | Metadata has `Language` = en and `Domain` = it. |
+| 2 | `Language` is a BCP 47 code and `Domain` is a value from the registry's domain list | Pass | `en` is BCP 47; `it` is in the registry domain list. |
+| 3 | The content (prose and table cells) is written in the stated language | Pass | All prose and table cells of MIL-010 are English (stand-in assessment, see the verdict). |
+| 4 | The register matches the one the registry gives for the artifact type | Pass | Register of the type is given in the registry (Languages); the text keeps to it. |
+| 5 | Domain terms are the PO terms of the domain's dictionary, with no synonyms | Pass | Terms follow DICT-001: bot, booking, verification (stand-in assessment, see the verdict). |
+| 6 | Metadata keys, section headings, IDs and statuses are in English | Pass | Keys, headings, IDs and statuses are English. |
+| 7 | No translated twin (`<name>.<language>.md`) exists beside the document | Pass | No translated twin exists in `docs/`. |
+| 8 | A change of language or domain since the previous accepted version has a Version History row and was reviewed again | N-A | First version; no language change. |
+| 9 | A reviewer competent in the domain, and in the language, has confirmed that the domain terms are used correctly | Pass | The stand-in reviewer reads English and checked the IT terms in context (stand-in assessment, see the verdict). |
+| 10 | Abbreviations are spelled out on first use, in the stated language | Pass | Abbreviations are spelled out or defined where first used, or are the framework's own IDs. |
+
+## Overall Verdict
+
+Go — MIL-010 meets every Mandatory criterion. No blocking finding. The review is not independent: the AI authored and reviewed it as stand-in for S02, and S01 can challenge it in the end-of-run summary. Criteria 3, 5 and 9 of QC-LANG-001 were assessed by the stand-in only; the framework would hold them `Pending` for an independent language reviewer. S01 is asked to confirm them in the end-of-run summary.
+
+## Action Items
+
+| Action | Owner | Due |
+| --- | --- | --- |
+| None | S01 | - |
+
+---
+
+[MIL-010]: ../../milestones/mil-010-credentials-file.md
+[QC-MIL-001]: ../../../framework/qc/artifacts/qc-milestones-gateways.md
+[434bdaf]: https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/commit/434bdafefdf0a95eed175dc89f27fbd91b19c91f
@@ -0,0 +1,62 @@
+# SQA Review Record
+
+## Metadata
+| Key | Value |
+| --- | --- |
+| ID | RC-026 |
+| CrossReference | [MIL-010], [QC-PY-001] |
+
+## Version History
+| Date | Status | Author | Reviewer | Change | Commit |
+| --- | --- | --- | --- | --- | --- |
+| 2026-10-28 | Accepted | Jens Tirsvad Nielsen | AI as stand-in for S02 | Review record, full review of the source and tests of MIL-010 | [42f05bd] |
+
+---
+
+## Artifact Under Review
+
+- Instance reviewed: [MIL-010]
+- Checklist used: [QC-PY-001] (`QC-PY-001`)
+- Scope: full review of the source and tests of MIL-010
+- Language and domain: n/a
+- Language reviewer: none
+- Mode: autonomous, not independent (the reviewer is `AI as stand-in for S02`)
+
+## Checklist Results
+
+| # | Criterion | Status | Evidence/Notes |
+| --- | --- | --- | --- |
+| 1 | Packages, modules, functions, variables, classes and constants follow PEP 8 casing (`snake_case`, `PascalCase`, `UPPER_SNAKE`) | Pass | Modules, functions and variables are `snake_case`, classes `PascalCase`, constants `UPPER_SNAKE`; `ruff check` with the N rules passes. |
+| 2 | Names state purpose in the domain's language; no unexplained abbreviations, no single-letter names outside tiny scopes | Pass | Names state purpose (for example `retry`, `login`, `book_class`, `get_my_bookings`); no single-letter names outside comprehensions. |
+| 3 | Code is produced by the project's formatter and passes its linter with no unexplained suppressions | Pass | `ruff format` leaves 0 files changed and `ruff check` reports 0 findings; there is no suppression. |
+| 4 | Every function and method signature is type-annotated, including `-> None` | Pass | Every function and method has annotations (ruff ANN rules pass), including `-> None`. |
+| 5 | No bare `except:`, no swallowed exceptions; specific exceptions are raised and the cause is kept (`raise ... from`) | Pass | No bare `except`; specific exceptions are caught, and re-raised with `raise ... from` where the cause matters. |
+| 6 | No mutable default arguments and no shadowed builtins | Pass | No mutable default arguments; no builtin is shadowed (ruff B and A-style checks pass). |
+| 7 | Files, locks and connections are managed with context managers | Pass | Files and the browser are managed with `with`/`try-finally`; no unmanaged file handle exists. |
+| 8 | Public modules, classes and functions have docstrings that say what, not how | Pass | Public modules, classes and functions have Doxygen docstrings (`@brief`, `@param`, `@return`). |
+| 9 | Logging uses `logging`, not `print`; no secrets or personal data in log output | Pass | No print calls; messages go through `logging`. |
+| 10 | Classes and operations trace to the Design Class Diagram they implement; deviations are recorded | N-A | No Design Class Diagram exists; ADR-0001 records that design artifacts are skipped. Classes follow ADR-0002. |
+| 11 | Tests exist for new behaviour, are named for the behaviour, and do not depend on order or the network | Pass | 273 unit tests need no browser or network (new: the admin account, half-set variables, the env file parser, precedence of the environment, `--env-file`); an autouse fixture keeps every CLI test in `tmp_path`, so no test can read `.env` or `.env.gym`; 21 container tests pass; coverage of the non-infrastructure modules is 100%. |
+| 12 | Type checker runs in strict mode without errors; `Any` is justified in a comment | N-A | No strict type checker is configured; annotations are complete and checked by ruff. |
+| 13 | Dependencies are declared and pinned in the project's dependency file, none unused | Pass | `pyproject.toml` declares Selenium (runtime) and pytest, pytest-cov, ruff (dev) with lower bounds; none is unused. |
+| 14 | Each piece of knowledge (a business rule, constant, format, validation or query) is defined in one place; code is merged only where it expresses the same knowledge, not where it merely looks alike | Pass | Each rule, constant and message exists once (constants in `constants.py`, rules in `domain/`). |
+| 15 | Imports point inward: business-rule modules import no framework, database, UI or delivery-mechanism code, and there are no import cycles | Pass | `domain` and `application` import no Selenium; only `infrastructure` imports it; no import cycle exists. |
+| 16 | SOLID holds: a class or module has one reason to change; new behaviour is added by extension, not by editing a type test in several places; subclasses honour the contract of their base; protocols are narrow; high-level code depends on protocols or abstract types and receives its implementations | Pass | One reason to change per module; the adapter hides Selenium behind a port; fakes honour the port contract. |
+| 17 | Source is laid out in the layers and folders of the clean architecture folder structure (`domain`, `application`, `infrastructure`, `presentation` packages under `src/`, and `unit`, `integration` and `e2e` under `tests/`), in `snake_case`, and imports between the layers point inward | Pass | Layout follows `src/snack_and_lift/{domain,application,infrastructure}` and `tests/{unit,integration,e2e}`; there is no presentation layer for a command-line bot (ADR-0002). |
+| 18 | File and folder names are English, contain no space and no character outside letters, digits, `-`, `_` and `.`, and follow the casing of the language for files | Pass | All file and folder names are English lowercase with `_`, `-` and `.` only. |
+
+## Overall Verdict
+
+Go — the code of MIL-010 meets every Mandatory criterion of QC-PY-001. No blocking finding. The review is not independent: the AI wrote and reviewed it as stand-in for S02, and S01 can challenge it in the end-of-run summary.
+
+## Action Items
+
+| Action | Owner | Due |
+| --- | --- | --- |
+| None | S01 | - |
+
+---
+
+[MIL-010]: ../../milestones/mil-010-credentials-file.md
+[QC-PY-001]: ../../../framework/qc/code/qc-programming-python.md
+[42f05bd]: https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/commit/42f05bdd0f5c300fbd162e8cf50075fc98964e1b
@@ -0,0 +1,69 @@
+# SQA Review Record
+
+## Metadata
+| Key | Value |
+| --- | --- |
+| ID | RC-027 |
+| CrossReference | [MIL-008], [QC-WF-001] |
+
+## Version History
+| Date | Status | Author | Reviewer | Change | Commit |
+| --- | --- | --- | --- | --- | --- |
+| 2026-10-29 | Accepted | Jens Tirsvad Nielsen | AI as stand-in for S02 | Review record, the workflow sync-milestones.yml after the framework update to 0.18.0 and the removal of publish-wiki.yml | [2cfcd02] |
+
+---
+
+## Artifact Under Review
+
+- Instance reviewed: [MIL-008]
+- Checklist used: [QC-WF-001] (`QC-WF-001`)
+- Scope: the workflow sync-milestones.yml after the framework update to 0.18.0 and the removal of publish-wiki.yml
+- Language and domain: n/a
+- Language reviewer: none
+- Mode: autonomous, not independent (the reviewer is `AI as stand-in for S02`)
+
+## Checklist Results
+
+| # | Criterion | Status | Evidence/Notes |
+| --- | --- | --- | --- |
+| 1 | The file meets the YAML checklist ([QC-YML-001]) | Pending | `yamllint` is not installed on the stand-in's machine, so the YAML checklist QC-YML-001 was not run; the files parse (actionlint reads them) and use 2-space indentation. |
+| 2 | The file is in `.gitea/workflows/` or `.github/workflows/` with the extension `.yml` or `.yaml`, and the same workflow does not exist in both directories | Pass | `sync-milestones.yml` is in `.gitea/workflows/` with the extension `.yml` and not in `.github/workflows/`. `publish-wiki.yml` was removed (see the verdict). |
+| 3 | The file declares `name`, `on` and `jobs`; `on` lists explicit events and, where it matters, `branches` or `paths` | Pass | The file declares `name`, `on` and `jobs`; `on` lists `push` on `main` with `paths`, and `workflow_dispatch`. |
+| 4 | A header comment states the purpose, the triggers, the inputs, the secrets it needs, what it changes outside the repository, and the decision or milestone it implements | Pass | The header states the purpose, triggers, secrets, what changes outside the repository and the installation steps (from the framework template); the decision is MIL-008 task 5. The added step has its own comment. |
+| 5 | Every job and step has a descriptive `name`, unique within its job | Pass | Every step has a descriptive name. |
+| 6 | The file passes `actionlint` (or the project's workflow validation test) and the tool version is stated | Pass | `actionlint` 1.7.12 reports no finding (exit code 0). |
+| 7 | Every job declares `permissions` explicitly with the least access that works (`contents: read` or none by default); write access only on the job that needs it | Pass | The job sets `contents: read`. |
+| 8 | Every action used is pinned to a full commit SHA or an exact release tag, never to a branch; a moving major tag (`@v1`) is accepted only for an action published by the same organisation under a stated release policy | Pass | `actions/checkout@v4.2.2` is an exact release tag. |
+| 9 | No secret is written in the file; a secret is read as `secrets.NAME`, passed to a step through `env:`, never placed on a command line or echoed | Pass | The file reads `secrets.TOKEN_FOR_GITEA_ISSUES` only; no secret value is in the repository. The token reaches git through `GIT_CONFIG_*` environment variables, never a command line. |
+| 10 | Untrusted text (`inputs.*`, event fields such as a title, branch name or label, step outputs derived from them) is never interpolated into a `run:` script; it is passed through `env:` and used as a quoted variable | Pass | No `run:` block interpolates `inputs.*` or event text; `DRY_RUN` is passed through `env:`. |
+| 11 | A trigger that gives a run secrets or a write token on code from a fork or an untrusted branch (`pull_request_target`, `workflow_run`) does not check out or execute that code | Pass | The workflow does not use `pull_request_target` or `workflow_run`. |
+| 12 | A workflow that changes state outside the repository (release, tag, publish, sync, deploy) is started by hand or by a protected branch, and a manual run defaults to a dry run with an explicit input to apply | Pass | It runs on a push to `main` or by hand; the manual run defaults to a dry run. |
+| 13 | The token the workflow uses has only the scope the job needs, and where the host's built-in token is enough it is used instead of a personal token | Pending | The scope of the token is set by S01; the dry run on 2026-10-28 got HTTP 403 on the framework repository, so the token lacks read access to it. |
+| 14 | Every job sets `runs-on` to an explicit label and `timeout-minutes` | Pass | The job sets `runs-on: ubuntu-latest` and `timeout-minutes: 15`. |
+| 15 | `concurrency` is set where two runs would conflict: a check cancels the older run (`cancel-in-progress: true`); a publishing job queues (`cancel-in-progress: false`) | Pass | `concurrency` is set with `cancel-in-progress: false`. |
+| 16 | Triggers are filtered (`branches`, `paths`, `types`) so the workflow runs only for the changes it concerns | Pass | The triggers are filtered by `branches: [main]` and `paths`. |
+| 17 | Tools and dependencies are installed at pinned versions; a version is a workflow input with a default or a value in one place | Pass | The workflow installs nothing itself; `curl` and `jq` are checked by the framework script. |
+| 18 | A dependency cache is used where the install is slow enough to matter | N-A | No dependency is installed that a cache would speed up. |
+| 19 | A failing step names the tool, the file and the reason in its log (annotation or `::error::` line) | Pending | Not verified: the log messages come from the framework script and the reusable workflow, which the stand-in did not run. |
+| 20 | A `run:` block is within the limit on embedded program text of criterion 7 of [QC-YML-001]; a longer script is a file (`tools/`, or a shared repository fetched at a pinned ref) that the step calls with one command, and the workflow passes its settings through `env:` | Pass | Each `run:` block is one command; the logic is in `framework/scripts/ci-publish-milestones.sh`. |
+| 21 | Logic used by more than one workflow or repository is a reusable workflow (`workflow_call`) or a composite action; each input has a `description`, a `type` and a `default` where one is sensible | Pass | The workflow calls one framework script; the input of `workflow_dispatch` has a description, type and default. |
+| 22 | Every workflow feature used is verified on the instance that runs it, and the host and version are recorded; a feature the target hosts differ on is noted in the header | Pending | A dry run on the instance (2026-10-28, run 3034, branch `fix-sync-milestones-submodule`) passed the HTTPS rewrite and then failed with 403 on the framework repository, so the feature is not yet verified end to end. |
+| 23 | A reusable workflow states, in its header, how a caller pins it and the script it fetches, so the workflow and its script are the same release | N-A | The file is not a reusable workflow. |
+| 24 | A static test of the repository's workflows exists and runs on every pull request: valid YAML, required keys, explicit permissions and runner, no inline script above the limit | N-A | No static test of the workflows exists in this project; the checks are `actionlint` on request. |
+
+## Overall Verdict
+
+Go-with-conditions - the milestone workflow is the framework template (0.18.0) plus one step that fetches the framework submodule over HTTPS, and it passes `actionlint`. `publish-wiki.yml` is removed because the framework retired the caller: the instance publishes `docs/wiki/` with its own scoped workflow (RC-022 reviewed the old caller). Mandatory criteria 1 (`yamllint` was not run) and 22 (feature checked on the instance) are Pending: the token lacks read access to the framework repository. The review is not independent: the AI wrote and reviewed it as stand-in for S02, and S01 can challenge it in the end-of-run summary.
+
+## Action Items
+
+| Action | Owner | Due |
+| --- | --- | --- |
+| Give the token `TOKEN_FOR_GITEA_ISSUES` read access to the repository TirSystem/SQA-QC-Framework, create the secret in the organization and run the workflow once by hand with dry_run on | S01 | before the next merge to main |
+| Run `yamllint` on `.gitea/workflows/` | S01 | when convenient |
+
+---
+
+[MIL-008]: ../../milestones/mil-008-documentation.md
+[QC-WF-001]: ../../../framework/qc/code/qc-workflow.md
+[2cfcd02]: https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/commit/2cfcd0258897aeb90cea48baf59a8af258df19dd
@@ -23,7 +23,7 @@ Tracks backward and forward links between artifact instances so that the Busines
 | --- | --- | --- | --- | --- | --- | --- |
 | [BC-001] | Business Case | en | it | - | [SA-001], [PP-001] | RC-001 |
 | [SA-001] | Stakeholder Analysis | en | it | [BC-001] | [PP-001] | RC-002 |
-| [PP-001] | Project Plan | en | it | [BC-001], [SA-001] | [MIL-001] to [MIL-008] | RC-003 |
+| [PP-001] | Project Plan | en | it | [BC-001], [SA-001] | [MIL-001] to [MIL-010] | RC-003 |
 | [DICT-001] | Domain Dictionary | en | it | [BC-001], [SA-001] | - | RC-004 |
 | [ADR-0001] | ADR | - | - | [BC-001] | [PP-001] | RC-005 |
 | [ADR-0002] | ADR | - | - | [BC-001], [SA-001] | code from MIL-002 onward | RC-006 |
@@ -36,11 +36,15 @@ Tracks backward and forward links between artifact instances so that the Busines
 | [MIL-006] | Milestone | en | it | [BC-001], [PP-001] | [MIL-007] | RC-013 |
 | [MIL-007] | Milestone | en | it | [BC-001], [PP-001] | [MIL-008] | RC-014 |
 | [MIL-008] | Milestone | en | it | [BC-001], [PP-001] | - | RC-015 |
+| [MIL-009] | Milestone | en | it | [BC-001], [MIL-006] | - | RC-023 |
+| [MIL-010] | Milestone | en | it | [BC-001], [MIL-003] | - | RC-025 |
 | Source and tests of [MIL-002] | Python source code | - | - | [MIL-002], [ADR-0002], [ADR-0003] | - | RC-016 |
 | Source and tests of [MIL-003] | Python source code | - | - | [MIL-003], [ADR-0002], [ADR-0003] | - | RC-017 (Go-with-conditions) |
 | Source and tests of [MIL-004] | Python source code | - | - | [MIL-004], [ADR-0002] | - | RC-018 (Go-with-conditions) |
 | Source and tests of [MIL-005] | Python source code | - | - | [MIL-005], [ADR-0002] | - | RC-019 (Go-with-conditions) |
 | Source and tests of [MIL-006] | Python source code | - | - | [MIL-006], [ADR-0002] | - | RC-020 (Go-with-conditions) |
+| Source and tests of [MIL-007] | Python source code | - | - | [MIL-007], [ADR-0002] | - | RC-021 (Go-with-conditions) |
+| Workflows and wiki of [MIL-008] | Workflow files, wiki pages | - | - | [MIL-008] | - | RC-022, RC-027 (Go-with-conditions) |

 ## Concern Coverage

@@ -50,8 +54,8 @@ Tracks backward and forward links between artifact instances so that the Busines
 | CN-002 | [PP-001], [MIL-001] | [ADR-0001] | Review records RC-001 onward | Covered |
 | CN-003 | [MIL-006], [MIL-007] | [ADR-0002] | Unit tests of `retry`; container chaos run | Planned |
 | CN-004 | [MIL-003] to [MIL-007] | [ADR-0002] | Code review against QC-PY-001 | Planned |
-| CN-005 | [MIL-002], [MIL-008] | - | `check-readme.sh` | Planned |
-| CN-006 | [MIL-008] | - | Page count per language | Planned |
+| CN-005 | [MIL-002], [MIL-008] | - | `check-readme.sh` | Covered |
+| CN-006 | [MIL-008] | - | Page count per language (12 pages in each of en, da, ms) | Covered |
 | CN-007 | [MIL-001] | - | Repository API response | Covered |
 | CN-008 | [MIL-002] | [ADR-0002] | `pyproject.toml` | Planned |
 | CN-009 | [MIL-001] | - | `.gitignore` lists `.env` | Covered |
@@ -73,4 +77,6 @@ Tracks backward and forward links between artifact instances so that the Busines
 [MIL-006]: ../milestones/mil-006-verify-and-time-travel.md
 [MIL-007]: ../milestones/mil-007-resilience.md
 [MIL-008]: ../milestones/mil-008-documentation.md
+[MIL-009]: ../milestones/mil-009-verification-fix.md
+[MIL-010]: ../milestones/mil-010-credentials-file.md
 [50e8331]: https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/commit/50e8331bd21910a72ae56b74c9ff6a3677a709b3
@@ -0,0 +1,69 @@
+[English](Getting-Started) · **Dansk** · [Bahasa Melayu](Getting-Started-ms)
+
+# Kom i gang
+
+## Krav
+
+Python 3.13 eller nyere, Chrome eller Chromium og (valgfrit) Docker.
+
+## Installation
+
+```bash
+python -m venv .venv
+```
+
+```bash
+source .venv/bin/activate
+```
+
+```powershell
+.venv\Scripts\Activate.ps1
+```
+
+```bash
+python -m pip install --upgrade pip
+python -m pip install -e ".[dev]"
+```
+
+## Din konto
+
+Botten indeholder ingen adgangskode. Opret en konto én gang i hånden på siden (kør med `--keep-open`, så mappen `chrome_profile` gemmer sidens data), og giv derefter botten kontoen:
+
+```bash
+export SNACK_AND_LIFT_EMAIL="yourname@test.com"
+export SNACK_AND_LIFT_PASSWORD="your password"
+```
+
+```powershell
+$env:SNACK_AND_LIFT_EMAIL = "yourname@test.com"
+$env:SNACK_AND_LIFT_PASSWORD = "your password"
+```
+
+## Kontofilen
+
+I stedet for at sætte variabler i hånden kan du skrive de samme linjer i en fil `.env.gym` i projektmappen (`SNACK_AND_LIFT_EMAIL=...` og `SNACK_AND_LIFT_PASSWORD=...`). Botten læser den ved start; `--env-file PATH` peger på en anden fil, og en rigtig miljøvariabel vinder over filen. Git ignorerer `.env.gym`: læg den aldrig i repositoriet. De valgfrie `SNACK_AND_LIFT_ADMIN_EMAIL` og `SNACK_AND_LIFT_ADMIN_PASSWORD` gemmer øvelsessidens admin-konto; botten indlæser dem, men bruger dem ikke endnu.
+
+## Kør og test
+
+```bash
+snack-and-lift
+```
+
+```bash
+python -m pytest
+```
+
+Nyttige indstillinger: `--headless`, `--keep-open`, `--first-tuesday-only` (trin 3 og 4), `--retries N`, `--profile-dir DIR`, `--screenshot-dir DIR`.
+
+## Exitkoder
+
+0 færdig, 1 et trin fejlede (også når alle forsøg fejlede), 2 loginoplysninger mangler, 3 siden My Bookings passer ikke med de bookinger der blev lavet.
+
+## Skærmbilleder (vedligeholder)
+
+Siden gemmer sine data i browserens profil, så en kørsel i en container kræver en konto der findes i den profil. Brug `--env-file` til loginoplysningerne (aldrig på kommandolinjen eller i repositoriet) og et volume til profilen og skærmbillederne:
+
+```bash
+docker build -t snack-and-lift .
+docker run --rm --env-file my.env -v snack-profile:/app/chrome_profile -v "$PWD/docs/images:/app/screenshots" snack-and-lift --headless --screenshot-dir /app/screenshots
+```
@@ -0,0 +1,69 @@
+[English](Getting-Started) · [Dansk](Getting-Started-da) · **Bahasa Melayu**
+
+# Mula menggunakan
+
+## Keperluan
+
+Python 3.13 atau lebih baharu, Chrome atau Chromium, dan (pilihan) Docker.
+
+## Pemasangan
+
+```bash
+python -m venv .venv
+```
+
+```bash
+source .venv/bin/activate
+```
+
+```powershell
+.venv\Scripts\Activate.ps1
+```
+
+```bash
+python -m pip install --upgrade pip
+python -m pip install -e ".[dev]"
+```
+
+## Akaun anda
+
+Bot tidak mengandungi kata laluan. Daftar sekali secara manual di laman (jalankan dengan `--keep-open` supaya folder `chrome_profile` menyimpan data laman), kemudian beri bot akaun itu:
+
+```bash
+export SNACK_AND_LIFT_EMAIL="yourname@test.com"
+export SNACK_AND_LIFT_PASSWORD="your password"
+```
+
+```powershell
+$env:SNACK_AND_LIFT_EMAIL = "yourname@test.com"
+$env:SNACK_AND_LIFT_PASSWORD = "your password"
+```
+
+## Fail akaun
+
+Daripada menetapkan pemboleh ubah secara manual, letakkan baris yang sama dalam fail `.env.gym` di folder projek (`SNACK_AND_LIFT_EMAIL=...` dan `SNACK_AND_LIFT_PASSWORD=...`). Bot membacanya semasa bermula; `--env-file PATH` menamakan fail lain, dan pemboleh ubah persekitaran sebenar mengatasi fail. Git mengabaikan `.env.gym`: jangan sekali-kali commit. `SNACK_AND_LIFT_ADMIN_EMAIL` dan `SNACK_AND_LIFT_ADMIN_PASSWORD` yang pilihan menyimpan akaun admin laman latihan; bot memuatkannya tetapi belum menggunakannya.
+
+## Jalankan dan uji
+
+```bash
+snack-and-lift
+```
+
+```bash
+python -m pytest
+```
+
+Pilihan berguna: `--headless`, `--keep-open`, `--first-tuesday-only` (langkah 3 dan 4), `--retries N`, `--profile-dir DIR`, `--screenshot-dir DIR`.
+
+## Kod keluar
+
+0 selesai, 1 satu langkah gagal (juga apabila semua percubaan gagal), 2 kelayakan tiada, 3 halaman My Bookings tidak sepadan dengan tempahan yang dibuat.
+
+## Mengambil tangkapan skrin (penyelenggara)
+
+Laman menyimpan datanya dalam profil pelayar, jadi larian dalam bekas memerlukan akaun yang wujud dalam profil itu. Gunakan `--env-file` untuk kelayakan (jangan pada baris arahan atau dalam repositori) dan volume untuk profil serta tangkapan skrin:
+
+```bash
+docker build -t snack-and-lift .
+docker run --rm --env-file my.env -v snack-profile:/app/chrome_profile -v "$PWD/docs/images:/app/screenshots" snack-and-lift --headless --screenshot-dir /app/screenshots
+```
@@ -0,0 +1,69 @@
+**English** · [Dansk](Getting-Started-da) · [Bahasa Melayu](Getting-Started-ms)
+
+# Getting started
+
+## Requirements
+
+Python 3.13 or newer, Chrome or Chromium, and (optional) Docker.
+
+## Install
+
+```bash
+python -m venv .venv
+```
+
+```bash
+source .venv/bin/activate
+```
+
+```powershell
+.venv\Scripts\Activate.ps1
+```
+
+```bash
+python -m pip install --upgrade pip
+python -m pip install -e ".[dev]"
+```
+
+## Your account
+
+The bot contains no password. Register once by hand on the site (run with `--keep-open`, so the `chrome_profile` folder keeps the site's data), then give the bot the account:
+
+```bash
+export SNACK_AND_LIFT_EMAIL="yourname@test.com"
+export SNACK_AND_LIFT_PASSWORD="your password"
+```
+
+```powershell
+$env:SNACK_AND_LIFT_EMAIL = "yourname@test.com"
+$env:SNACK_AND_LIFT_PASSWORD = "your password"
+```
+
+## The account file
+
+Instead of setting variables by hand, put the same lines in a file `.env.gym` in the project folder (`SNACK_AND_LIFT_EMAIL=...` and `SNACK_AND_LIFT_PASSWORD=...`). The bot reads it at start; `--env-file PATH` names another file, and a real environment variable wins over the file. Git ignores `.env.gym`: never commit it. The optional `SNACK_AND_LIFT_ADMIN_EMAIL` and `SNACK_AND_LIFT_ADMIN_PASSWORD` hold the practice site's admin account; the bot loads them but does not use them yet.
+
+## Run and test
+
+```bash
+snack-and-lift
+```
+
+```bash
+python -m pytest
+```
+
+Useful options: `--headless`, `--keep-open`, `--first-tuesday-only` (steps 3 and 4), `--retries N`, `--profile-dir DIR`, `--screenshot-dir DIR`.
+
+## Exit codes
+
+0 done, 1 a step failed (also when every retry failed), 2 credentials missing, 3 the My Bookings page does not match the bookings made.
+
+## Taking the screenshots (maintainer)
+
+The site keeps its data in the browser profile, so a container run needs an account that exists in that profile. Use `--env-file` for the credentials (never write them on the command line or into the repository) and a volume for the profile and the screenshots:
+
+```bash
+docker build -t snack-and-lift .
+docker run --rm --env-file my.env -v snack-profile:/app/chrome_profile -v "$PWD/docs/images:/app/screenshots" snack-and-lift --headless --screenshot-dir /app/screenshots
+```
@@ -0,0 +1,25 @@
+[English](Home) · **Dansk** · [Bahasa Melayu](Home-ms)
+
+# Snack and Lift-bot
+
+En Selenium-bot der booker hold på øvelsesgymmet [Snack and Lift](https://appbrewery.github.io/gym/). Den er løsningen på dag 49 i Udemys *100 Days of Code*. Hver side nedenfor forklarer hvad ét trin i lektionen lærer.
+
+Botten logger ind, booker hvert tirsdags- og torsdagshold kl. 18 (eller kommer på ventelisten), bekræfter resultatet på siden My Bookings og prøver igen når siden fejler en forespørgsel.
+
+- [Kom i gang](Getting-Started-da): `.venv`, din konto, kørsel og test.
+- De ni trin, nedenfor.
+- [Fejlfinding](Troubleshooting-da).
+
+Kildekode og sager: [repository](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift).
+
+## De ni trin
+
+1. [Trin 1: Opsætning, Chrome-profil og navigation](Step-1-da)
+2. [Trin 2: Automatisk login](Step-2-da)
+3. [Trin 3: Book næste tirsdagshold](Step-3-da)
+4. [Trin 4: Tjek om et hold allerede er booket](Step-4-da)
+5. [Trin 5: Tællere og et resumé](Step-5-da)
+6. [Trin 6: Alle tirsdags- og torsdagshold](Step-6-da)
+7. [Trin 7: Bekræft på siden My Bookings](Step-7-da)
+8. [Trin 8: Kvalitetssikring med tidsrejse](Step-8-da)
+9. [Trin 9: Robusthed mod netværksfejl](Step-9-da)
@@ -0,0 +1,25 @@
+[English](Home) · [Dansk](Home-da) · **Bahasa Melayu**
+
+# Bot Snack and Lift
+
+Bot Selenium yang menempah kelas di gim latihan [Snack and Lift](https://appbrewery.github.io/gym/). Ia penyelesaian untuk hari 49 *100 Days of Code* Udemy. Setiap halaman di bawah menerangkan apa yang diajar oleh satu langkah kuliah.
+
+Bot log masuk, menempah setiap kelas Selasa dan Khamis pukul 6 petang (atau menyertai senarai menunggu), mengesahkan hasilnya di halaman My Bookings dan mencuba semula apabila laman gagal memproses permintaan.
+
+- [Mula menggunakan](Getting-Started-ms): `.venv`, akaun anda, menjalankan dan menguji.
+- Sembilan langkah, di bawah.
+- [Menyelesaikan masalah](Troubleshooting-ms).
+
+Kod sumber dan isu: [repository](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift).
+
+## Sembilan langkah
+
+1. [Langkah 1: Persediaan, profil Chrome dan navigasi](Step-1-ms)
+2. [Langkah 2: Log masuk automatik](Step-2-ms)
+3. [Langkah 3: Tempah kelas Selasa seterusnya](Step-3-ms)
+4. [Langkah 4: Semak sama ada kelas sudah ditempah](Step-4-ms)
+5. [Langkah 5: Pembilang dan ringkasan](Step-5-ms)
+6. [Langkah 6: Semua kelas Selasa dan Khamis](Step-6-ms)
+7. [Langkah 7: Sahkan di halaman My Bookings](Step-7-ms)
+8. [Langkah 8: Jaminan kualiti dengan perjalanan masa](Step-8-ms)
+9. [Langkah 9: Ketahanan terhadap kegagalan rangkaian](Step-9-ms)
@@ -0,0 +1,25 @@
+**English** · [Dansk](Home-da) · [Bahasa Melayu](Home-ms)
+
+# Snack and Lift bot
+
+A Selenium bot that books classes at the practice gym [Snack and Lift](https://appbrewery.github.io/gym/). It is the solution to day 49 of Udemy's *100 Days of Code*. Each page below explains what one lecture step teaches.
+
+The bot logs in, books every Tuesday and Thursday 6pm class (or joins the waitlist), verifies the result on the My Bookings page and retries when the site fails a request.
+
+- [Getting started](Getting-Started): the `.venv`, your account, running and testing.
+- The nine steps, below.
+- [Troubleshooting](Troubleshooting).
+
+Source and issues: [repository](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift).
+
+## The nine steps
+
+1. [Step 1: Setup, Chrome profile and navigation](Step-1)
+2. [Step 2: Automated login](Step-2)
+3. [Step 3: Book the next Tuesday class](Step-3)
+4. [Step 4: Check if a class is already booked](Step-4)
+5. [Step 5: Counters and a summary](Step-5)
+6. [Step 6: Every Tuesday and Thursday class](Step-6)
+7. [Step 7: Verify on the My Bookings page](Step-7)
+8. [Step 8: Time travel quality assurance](Step-8)
+9. [Step 9: Resilience against network failures](Step-9)
@@ -0,0 +1,17 @@
+[English](Step-1) · **Dansk** · [Bahasa Melayu](Step-1-ms)
+
+# Trin 1: Opsætning, Chrome-profil og navigation
+
+## Hvad du får ud af det
+
+Du lærer at starte Chrome fra Selenium med en egen profilmappe (`chrome_profile`), så sidens data overlever mellem kørsler, og med indstillingen `detach`, så vinduet bliver åbent når scriptet slutter. Botten åbner forsiden og venter på login-knappen i stedet for at sove.
+
+## Hvor det er i koden
+
+`src/snack_and_lift/infrastructure/browser.py`
+
+## Skærmbillede
+
+![Trin 1: Opsætning, Chrome-profil og navigation](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-1.png)
+
+[Forside](Home-da) · [Næste: 2](Step-2-da)
@@ -0,0 +1,17 @@
+[English](Step-1) · [Dansk](Step-1-da) · **Bahasa Melayu**
+
+# Langkah 1: Persediaan, profil Chrome dan navigasi
+
+## Apa yang anda peroleh
+
+Anda belajar memulakan Chrome daripada Selenium dengan folder profil sendiri (`chrome_profile`) supaya data laman kekal antara larian, dan dengan pilihan `detach` supaya tetingkap kekal terbuka apabila skrip tamat. Bot membuka halaman utama dan menunggu butang log masuk, bukannya tidur.
+
+## Di mana dalam kod
+
+`src/snack_and_lift/infrastructure/browser.py`
+
+## Tangkapan skrin
+
+![Langkah 1: Persediaan, profil Chrome dan navigasi](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-1.png)
+
+[Utama](Home-ms) · [Seterusnya: 2](Step-2-ms)
@@ -0,0 +1,17 @@
+**English** · [Dansk](Step-1-da) · [Bahasa Melayu](Step-1-ms)
+
+# Step 1: Setup, Chrome profile and navigation
+
+## What you gain
+
+You learn to start Chrome from Selenium with its own profile folder (`chrome_profile`), so the site's data survives between runs, and with the `detach` option, so the window stays open when the script ends. The bot opens the home page and waits for the login button instead of sleeping.
+
+## Where it is in the code
+
+`src/snack_and_lift/infrastructure/browser.py`
+
+## Screenshot
+
+![Step 1: Setup, Chrome profile and navigation](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-1.png)
+
+[Home](Home) · [Next: 2](Step-2)
@@ -0,0 +1,17 @@
+[English](Step-2) · **Dansk** · [Bahasa Melayu](Step-2-ms)
+
+# Trin 2: Automatisk login
+
+## Hvad du får ud af det
+
+Du lærer eksplicitte ventetider (`WebDriverWait`) og at udfylde en formular. `login()` skriver e-mail og adgangskode fra miljøvariabler, indsender og venter til timeplanen vises. En forkert adgangskode vises via sidens fejlmeddelelse.
+
+## Hvor det er i koden
+
+`src/snack_and_lift/application/session.py`
+
+## Skærmbillede
+
+![Trin 2: Automatisk login](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-2.png)
+
+[Forrige: 1](Step-1-da) · [Forside](Home-da) · [Næste: 3](Step-3-da)
@@ -0,0 +1,17 @@
+[English](Step-2) · [Dansk](Step-2-da) · **Bahasa Melayu**
+
+# Langkah 2: Log masuk automatik
+
+## Apa yang anda peroleh
+
+Anda belajar menunggu secara eksplisit (`WebDriverWait`) dan mengisi borang. `login()` menaip e-mel dan kata laluan daripada pemboleh ubah persekitaran, menghantar, dan menunggu sehingga jadual kelas dipaparkan. Kata laluan salah dilaporkan melalui mesej ralat laman.
+
+## Di mana dalam kod
+
+`src/snack_and_lift/application/session.py`
+
+## Tangkapan skrin
+
+![Langkah 2: Log masuk automatik](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-2.png)
+
+[Sebelumnya: 1](Step-1-ms) · [Utama](Home-ms) · [Seterusnya: 3](Step-3-ms)
@@ -0,0 +1,17 @@
+**English** · [Dansk](Step-2-da) · [Bahasa Melayu](Step-2-ms)
+
+# Step 2: Automated login
+
+## What you gain
+
+You learn explicit waits (`WebDriverWait`) and how to fill a form. `login()` types the e-mail and password from environment variables, submits, and waits until the class schedule is shown. A wrong password is reported by the site's error message.
+
+## Where it is in the code
+
+`src/snack_and_lift/application/session.py`
+
+## Screenshot
+
+![Step 2: Automated login](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-2.png)
+
+[Previous: 1](Step-1) · [Home](Home) · [Next: 3](Step-3)
@@ -0,0 +1,17 @@
+[English](Step-3) · **Dansk** · [Bahasa Melayu](Step-3-ms)
+
+# Trin 3: Book næste tirsdagshold
+
+## Hvad du får ud af det
+
+Du lærer at finde et element relativt til et andet: næste tirsdags dagsgruppe findes med et XPath-opslag på forfædre, derefter 18-kortet og dets knap. `book_class()` klikker på knappen og venter til teksten ændres.
+
+## Hvor det er i koden
+
+`src/snack_and_lift/application/booking.py`
+
+## Skærmbillede
+
+![Trin 3: Book næste tirsdagshold](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-3.png)
+
+[Forrige: 2](Step-2-da) · [Forside](Home-da) · [Næste: 4](Step-4-da)
@@ -0,0 +1,17 @@
+[English](Step-3) · [Dansk](Step-3-da) · **Bahasa Melayu**
+
+# Langkah 3: Tempah kelas Selasa seterusnya
+
+## Apa yang anda peroleh
+
+Anda belajar mencari elemen berdasarkan elemen lain: kumpulan hari Selasa seterusnya dicari dengan carian XPath ancestor, kemudian kad pukul 6 petang dan butangnya. `book_class()` menekan butang dan menunggu teks berubah.
+
+## Di mana dalam kod
+
+`src/snack_and_lift/application/booking.py`
+
+## Tangkapan skrin
+
+![Langkah 3: Tempah kelas Selasa seterusnya](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-3.png)
+
+[Sebelumnya: 2](Step-2-ms) · [Utama](Home-ms) · [Seterusnya: 4](Step-4-ms)
@@ -0,0 +1,17 @@
+**English** · [Dansk](Step-3-da) · [Bahasa Melayu](Step-3-ms)
+
+# Step 3: Book the next Tuesday class
+
+## What you gain
+
+You learn to find an element relative to another one: the day group of the next Tuesday is found with an XPath ancestor lookup, then the 6pm card and its button. `book_class()` clicks the button and waits until its text changes.
+
+## Where it is in the code
+
+`src/snack_and_lift/application/booking.py`
+
+## Screenshot
+
+![Step 3: Book the next Tuesday class](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-3.png)
+
+[Previous: 2](Step-2) · [Home](Home) · [Next: 4](Step-4)
@@ -0,0 +1,17 @@
+[English](Step-4) · **Dansk** · [Bahasa Melayu](Step-4-ms)
+
+# Trin 4: Tjek om et hold allerede er booket
+
+## Hvad du får ud af det
+
+Du lærer at læse en knaps tilstand før du handler: "Book Class", "Join Waitlist", "Booked" eller "Waitlisted". Et hold der allerede er booket klikkes ikke igen, og botten udskriver en ✓-linje om hvad den gjorde.
+
+## Hvor det er i koden
+
+`src/snack_and_lift/domain/rules.py`
+
+## Skærmbillede
+
+![Trin 4: Tjek om et hold allerede er booket](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-4.png)
+
+[Forrige: 3](Step-3-da) · [Forside](Home-da) · [Næste: 5](Step-5-da)
@@ -0,0 +1,17 @@
+[English](Step-4) · [Dansk](Step-4-da) · **Bahasa Melayu**
+
+# Langkah 4: Semak sama ada kelas sudah ditempah
+
+## Apa yang anda peroleh
+
+Anda belajar membaca keadaan butang sebelum bertindak: "Book Class", "Join Waitlist", "Booked" atau "Waitlisted". Kelas yang sudah ditempah tidak ditekan semula, dan bot mencetak baris ✓ tentang apa yang dilakukannya.
+
+## Di mana dalam kod
+
+`src/snack_and_lift/domain/rules.py`
+
+## Tangkapan skrin
+
+![Langkah 4: Semak sama ada kelas sudah ditempah](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-4.png)
+
+[Sebelumnya: 3](Step-3-ms) · [Utama](Home-ms) · [Seterusnya: 5](Step-5-ms)
@@ -0,0 +1,17 @@
+**English** · [Dansk](Step-4-da) · [Bahasa Melayu](Step-4-ms)
+
+# Step 4: Check if a class is already booked
+
+## What you gain
+
+You learn to read the state of a button before acting: "Book Class", "Join Waitlist", "Booked" or "Waitlisted". A class that is already booked is not clicked again, and the bot prints a ✓ line that says what it did.
+
+## Where it is in the code
+
+`src/snack_and_lift/domain/rules.py`
+
+## Screenshot
+
+![Step 4: Check if a class is already booked](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-4.png)
+
+[Previous: 3](Step-3) · [Home](Home) · [Next: 5](Step-5)
@@ -0,0 +1,17 @@
+[English](Step-5) · **Dansk** · [Bahasa Melayu](Step-5-ms)
+
+# Trin 5: Tællere og et resumé
+
+## Hvad du får ud af det
+
+Du lærer at føre tællere mens en løkke kører og at udskrive en rapport: BOOKING SUMMARY viser nye bookinger, nye ventelisteplaceringer og hold der allerede var booket eller på venteliste.
+
+## Hvor det er i koden
+
+`src/snack_and_lift/domain/report.py`
+
+## Skærmbillede
+
+![Trin 5: Tællere og et resumé](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-5.png)
+
+[Forrige: 4](Step-4-da) · [Forside](Home-da) · [Næste: 6](Step-6-da)
@@ -0,0 +1,17 @@
+[English](Step-5) · [Dansk](Step-5-da) · **Bahasa Melayu**
+
+# Langkah 5: Pembilang dan ringkasan
+
+## Apa yang anda peroleh
+
+Anda belajar menyimpan pembilang semasa gelung berjalan dan mencetak laporan: BOOKING SUMMARY menyenaraikan tempahan baharu, penyertaan senarai menunggu baharu, serta kelas yang sudah ditempah atau dalam senarai menunggu.
+
+## Di mana dalam kod
+
+`src/snack_and_lift/domain/report.py`
+
+## Tangkapan skrin
+
+![Langkah 5: Pembilang dan ringkasan](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-5.png)
+
+[Sebelumnya: 4](Step-4-ms) · [Utama](Home-ms) · [Seterusnya: 6](Step-6-ms)
@@ -0,0 +1,17 @@
+**English** · [Dansk](Step-5-da) · [Bahasa Melayu](Step-5-ms)
+
+# Step 5: Counters and a summary
+
+## What you gain
+
+You learn to keep counters while a loop runs and to print a report: the BOOKING SUMMARY lists new bookings, new waitlist entries, and classes that were already booked or waitlisted.
+
+## Where it is in the code
+
+`src/snack_and_lift/domain/report.py`
+
+## Screenshot
+
+![Step 5: Counters and a summary](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-5.png)
+
+[Previous: 4](Step-4) · [Home](Home) · [Next: 6](Step-6)
@@ -0,0 +1,17 @@
+[English](Step-6) · **Dansk** · [Bahasa Melayu](Step-6-ms)
+
+# Trin 6: Alle tirsdags- og torsdagshold
+
+## Hvad du får ud af det
+
+Du lærer at gennemløbe alle matchende elementer: hvert tirsdags- og torsdagshold kl. 18 håndteres, og DETAILED CLASS LIST udskriver hvert hold med dato, tid og resultat (New Booking, New Waitlist, Already Booked, Already Waitlisted).
+
+## Hvor det er i koden
+
+`src/snack_and_lift/application/booking.py`
+
+## Skærmbillede
+
+![Trin 6: Alle tirsdags- og torsdagshold](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-6.png)
+
+[Forrige: 5](Step-5-da) · [Forside](Home-da) · [Næste: 7](Step-7-da)
@@ -0,0 +1,17 @@
+[English](Step-6) · [Dansk](Step-6-da) · **Bahasa Melayu**
+
+# Langkah 6: Semua kelas Selasa dan Khamis
+
+## Apa yang anda peroleh
+
+Anda belajar mengulang semua elemen yang sepadan: setiap kelas Selasa dan Khamis pukul 6 petang diuruskan, dan DETAILED CLASS LIST mencetak setiap kelas dengan tarikh, masa dan hasil (New Booking, New Waitlist, Already Booked, Already Waitlisted).
+
+## Di mana dalam kod
+
+`src/snack_and_lift/application/booking.py`
+
+## Tangkapan skrin
+
+![Langkah 6: Semua kelas Selasa dan Khamis](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-6.png)
+
+[Sebelumnya: 5](Step-5-ms) · [Utama](Home-ms) · [Seterusnya: 7](Step-7-ms)
@@ -0,0 +1,17 @@
+**English** · [Dansk](Step-6-da) · [Bahasa Melayu](Step-6-ms)
+
+# Step 6: Every Tuesday and Thursday class
+
+## What you gain
+
+You learn to loop over all matching elements: every Tuesday and Thursday 6pm class is handled, and the DETAILED CLASS LIST prints each class with its date, time and result (New Booking, New Waitlist, Already Booked, Already Waitlisted).
+
+## Where it is in the code
+
+`src/snack_and_lift/application/booking.py`
+
+## Screenshot
+
+![Step 6: Every Tuesday and Thursday class](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-6.png)
+
+[Previous: 5](Step-5) · [Home](Home) · [Next: 7](Step-7)
@@ -0,0 +1,17 @@
+[English](Step-7) · **Dansk** · [Bahasa Melayu](Step-7-ms)
+
+# Trin 7: Bekræft på siden My Bookings
+
+## Hvad du får ud af det
+
+Du lærer at kontrollere dit eget arbejde fra en anden kilde: `get_my_bookings()` læser siden My Bookings (bookinger og venteliste) og `verify_bookings()` sammenligner med det botten gjorde. Resultatet er SUCCESS eller MISMATCH (exitkode 3).
+
+## Hvor det er i koden
+
+`src/snack_and_lift/application/verification.py`
+
+## Skærmbillede
+
+![Trin 7: Bekræft på siden My Bookings](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-7.png)
+
+[Forrige: 6](Step-6-da) · [Forside](Home-da) · [Næste: 8](Step-8-da)
@@ -0,0 +1,17 @@
+[English](Step-7) · [Dansk](Step-7-da) · **Bahasa Melayu**
+
+# Langkah 7: Sahkan di halaman My Bookings
+
+## Apa yang anda peroleh
+
+Anda belajar menyemak kerja sendiri daripada sumber kedua: `get_my_bookings()` membaca halaman My Bookings (tempahan dan senarai menunggu) dan `verify_bookings()` membandingkannya dengan apa yang dilakukan bot. Hasilnya SUCCESS atau MISMATCH (kod keluar 3).
+
+## Di mana dalam kod
+
+`src/snack_and_lift/application/verification.py`
+
+## Tangkapan skrin
+
+![Langkah 7: Sahkan di halaman My Bookings](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-7.png)
+
+[Sebelumnya: 6](Step-6-ms) · [Utama](Home-ms) · [Seterusnya: 8](Step-8-ms)
@@ -0,0 +1,17 @@
+**English** · [Dansk](Step-7-da) · [Bahasa Melayu](Step-7-ms)
+
+# Step 7: Verify on the My Bookings page
+
+## What you gain
+
+You learn to check your own work from a second source: `get_my_bookings()` reads the My Bookings page (bookings and waitlist entries) and `verify_bookings()` compares it with what the bot did. The result is SUCCESS or MISMATCH (exit code 3).
+
+## Where it is in the code
+
+`src/snack_and_lift/application/verification.py`
+
+## Screenshot
+
+![Step 7: Verify on the My Bookings page](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-7.png)
+
+[Previous: 6](Step-6) · [Home](Home) · [Next: 8](Step-8)
@@ -0,0 +1,19 @@
+[English](Step-8) · **Dansk** · [Bahasa Melayu](Step-8-ms)
+
+# Trin 8: Kvalitetssikring med tidsrejse
+
+## Hvad du får ud af det
+
+Du lærer at datoer er den almindelige fejlkilde: sidens admin-knap "advance 3 days" flytter dens ur. Botten læser dagsetiketterne ("Today", "Tomorrow", "Thu, Aug 14") i stedet for at antage en dato, så den også virker efter et tidshop.
+
+## Hvor det er i koden
+
+`src/snack_and_lift/domain/day_label.py`
+
+## Skærmbillede
+
+![Trin 8: Kvalitetssikring med tidsrejse](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-8.png)
+
+> Siden My Bookings viser også bookinger fra tidligere kørsler uden for denne kørsels dage. Kontrollen sammenligner kun kørslens dage og skriver de øvrige som "Also booked (not checked)".
+
+[Forrige: 7](Step-7-da) · [Forside](Home-da) · [Næste: 9](Step-9-da)
@@ -0,0 +1,19 @@
+[English](Step-8) · [Dansk](Step-8-da) · **Bahasa Melayu**
+
+# Langkah 8: Jaminan kualiti dengan perjalanan masa
+
+## Apa yang anda peroleh
+
+Anda belajar bahawa tarikh ialah punca pepijat biasa: kawalan admin laman "advance 3 days" menggerakkan jamnya. Bot membaca label hari ("Today", "Tomorrow", "Thu, Aug 14") dan bukannya menganggap tarikh, jadi ia berfungsi selepas lompatan masa juga.
+
+## Di mana dalam kod
+
+`src/snack_and_lift/domain/day_label.py`
+
+## Tangkapan skrin
+
+![Langkah 8: Jaminan kualiti dengan perjalanan masa](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-8.png)
+
+> Halaman My Bookings turut menyenaraikan tempahan daripada larian terdahulu di luar hari larian ini. Semakan hanya membandingkan hari larian dan mencetak yang lain sebagai "Also booked (not checked)".
+
+[Sebelumnya: 7](Step-7-ms) · [Utama](Home-ms) · [Seterusnya: 9](Step-9-ms)
@@ -0,0 +1,19 @@
+**English** · [Dansk](Step-8-da) · [Bahasa Melayu](Step-8-ms)
+
+# Step 8: Time travel quality assurance
+
+## What you gain
+
+You learn that dates are the usual bug: the site's admin control "advance 3 days" moves its clock. The bot reads the day labels ("Today", "Tomorrow", "Thu, Aug 14") instead of assuming a date, so it works after a time jump too.
+
+## Where it is in the code
+
+`src/snack_and_lift/domain/day_label.py`
+
+## Screenshot
+
+![Step 8: Time travel quality assurance](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-8.png)
+
+> The My Bookings page also lists bookings from earlier runs outside the days of this run. The check compares only the days of the run and prints the others as "Also booked (not checked)".
+
+[Previous: 7](Step-7) · [Home](Home) · [Next: 9](Step-9)
@@ -0,0 +1,17 @@
+[English](Step-9) · **Dansk** · [Bahasa Melayu](Step-9-ms)
+
+# Trin 9: Robusthed mod netværksfejl
+
+## Hvad du får ud af det
+
+Du lærer at prøve igen: `retry(func, retries=7, description=None)` gentager et kald der kaster en fejl, venter mellem forsøgene, logger en advarsel og kaster RetriesExhaustedError til sidst. `login()`, `book_class()` og `get_my_bookings()` er pakket ind i den, så en side der taber forespørgsler ikke stopper kørslen.
+
+## Hvor det er i koden
+
+`src/snack_and_lift/application/retry.py`
+
+## Skærmbillede
+
+![Trin 9: Robusthed mod netværksfejl](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-9.png)
+
+[Forrige: 8](Step-8-da) · [Forside](Home-da)
@@ -0,0 +1,17 @@
+[English](Step-9) · [Dansk](Step-9-da) · **Bahasa Melayu**
+
+# Langkah 9: Ketahanan terhadap kegagalan rangkaian
+
+## Apa yang anda peroleh
+
+Anda belajar mencuba semula: `retry(func, retries=7, description=None)` mengulang panggilan yang gagal, menunggu antara percubaan, mencatat amaran dan melempar RetriesExhaustedError di akhir. `login()`, `book_class()` dan `get_my_bookings()` dibalut dengannya, jadi laman yang menjatuhkan permintaan tidak menghentikan larian.
+
+## Di mana dalam kod
+
+`src/snack_and_lift/application/retry.py`
+
+## Tangkapan skrin
+
+![Langkah 9: Ketahanan terhadap kegagalan rangkaian](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-9.png)
+
+[Sebelumnya: 8](Step-8-ms) · [Utama](Home-ms)
@@ -0,0 +1,17 @@
+**English** · [Dansk](Step-9-da) · [Bahasa Melayu](Step-9-ms)
+
+# Step 9: Resilience against network failures
+
+## What you gain
+
+You learn to retry: `retry(func, retries=7, description=None)` repeats a call that raises, waits between attempts, logs a warning and raises RetriesExhaustedError at the end. `login()`, `book_class()` and `get_my_bookings()` are wrapped with it, so a site that drops requests does not stop the run.
+
+## Where it is in the code
+
+`src/snack_and_lift/application/retry.py`
+
+## Screenshot
+
+![Step 9: Resilience against network failures](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift/raw/branch/main/docs/images/step-9.png)
+
+[Previous: 8](Step-8) · [Home](Home)
@@ -0,0 +1,12 @@
+[English](Troubleshooting) · **Dansk** · [Bahasa Melayu](Troubleshooting-ms)
+
+# Fejlfinding
+
+| Symptom | Årsag og løsning |
+| --- | --- |
+| Exitkode 2 | `SNACK_AND_LIFT_EMAIL` eller `SNACK_AND_LIFT_PASSWORD` er ikke sat. |
+| Login fejler 7 gange, derefter exit 1 | Forkert e-mail eller adgangskode, eller kontoen findes ikke i denne `chrome_profile` (siden gemmer konti i browseren). Opret den igen med `--keep-open`. |
+| Exitkode 3 og MISMATCH | Siden My Bookings viser ikke det botten bookede. Kør igen; tjek om et hold var fyldt. |
+| Advarsler "attempt 2 of 7" | Sidens fejlrate er slået til; det er retry der arbejder (trin 9). |
+| Chrome advarer om adgangskoden | Chrome markerer svage adgangskoder som `12346`; vælg en anden. |
+| Intet vindue i Docker | Containeren er headless med vilje; brug `--screenshot-dir`. |
@@ -0,0 +1,12 @@
+[English](Troubleshooting) · [Dansk](Troubleshooting-da) · **Bahasa Melayu**
+
+# Menyelesaikan masalah
+
+| Gejala | Punca dan penyelesaian |
+| --- | --- |
+| Kod keluar 2 | `SNACK_AND_LIFT_EMAIL` atau `SNACK_AND_LIFT_PASSWORD` tidak ditetapkan. |
+| Log masuk gagal 7 kali, kemudian keluar 1 | E-mel atau kata laluan salah, atau akaun tiada dalam `chrome_profile` ini (laman menyimpan akaun dalam pelayar). Daftar semula dengan `--keep-open`. |
+| Kod keluar 3 dan MISMATCH | Halaman My Bookings tidak menyenaraikan apa yang ditempah bot. Jalankan semula; semak sama ada kelas penuh. |
+| Amaran "attempt 2 of 7" | Kadar kegagalan laman dihidupkan; ini retry sedang berfungsi (langkah 9). |
+| Chrome memberi amaran tentang kata laluan | Chrome menandakan kata laluan lemah seperti `12346`; pilih yang lain. |
+| Tiada tetingkap dalam Docker | Bekas ini tanpa kepala (headless) secara sengaja; gunakan `--screenshot-dir`. |
@@ -0,0 +1,12 @@
+**English** · [Dansk](Troubleshooting-da) · [Bahasa Melayu](Troubleshooting-ms)
+
+# Troubleshooting
+
+| Symptom | Cause and fix |
+| --- | --- |
+| Exit code 2 | `SNACK_AND_LIFT_EMAIL` or `SNACK_AND_LIFT_PASSWORD` is not set. |
+| Login fails 7 times, then exit 1 | Wrong e-mail or password, or the account is not in this `chrome_profile` (the site keeps accounts in the browser). Register again with `--keep-open`. |
+| Exit code 3 and MISMATCH | The My Bookings page does not list what the bot booked. Run again; check whether a class was full. |
+| Warnings "attempt 2 of 7" | The site's failure rate is on; this is the retry at work (step 9). |
+| Chrome warns about the password | Chrome flags weak passwords such as `12346`; choose another. |
+| No window in Docker | The container is headless by design; use `--screenshot-dir`. |
@@ -0,0 +1 @@
+[Repository](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift) · [Snack and Lift](https://appbrewery.github.io/gym/) · AGPL-3.0
@@ -0,0 +1,11 @@
+**English**
+- [Home](Home) · [Getting started](Getting-Started) · [Troubleshooting](Troubleshooting)
+- [1](Step-1) · [2](Step-2) · [3](Step-3) · [4](Step-4) · [5](Step-5) · [6](Step-6) · [7](Step-7) · [8](Step-8) · [9](Step-9)
+
+**Dansk**
+- [Forside](Home-da) · [Kom i gang](Getting-Started-da) · [Fejlfinding](Troubleshooting-da)
+- [1](Step-1-da) · [2](Step-2-da) · [3](Step-3-da) · [4](Step-4-da) · [5](Step-5-da) · [6](Step-6-da) · [7](Step-7-da) · [8](Step-8-da) · [9](Step-9-da)
+
+**Bahasa Melayu**
+- [Utama](Home-ms) · [Mula menggunakan](Getting-Started-ms) · [Menyelesaikan masalah](Troubleshooting-ms)
+- [1](Step-1-ms) · [2](Step-2-ms) · [3](Step-3-ms) · [4](Step-4-ms) · [5](Step-5-ms) · [6](Step-6-ms) · [7](Step-7-ms) · [8](Step-8-ms) · [9](Step-9-ms)
@@ -3,10 +3,12 @@
 """

 import logging
+import time
 from collections.abc import Callable

 from snack_and_lift import constants
 from snack_and_lift.application.ports import GymPort
+from snack_and_lift.application.retry import Sleep, retry
 from snack_and_lift.domain.errors import GymError
 from snack_and_lift.domain.models import ButtonState, ClassCard, ClassResult
 from snack_and_lift.domain.rules import (
@@ -41,12 +43,68 @@ def book_class(gym: GymPort, card: ClassCard) -> ClassResult:
    return ClassResult(card, outcome)


-def book_next_tuesday(gym: GymPort) -> ClassResult | None:
+def _read_current_card(gym: GymPort, card: ClassCard) -> ClassCard:
+    """@brief Read a class again from the page, to see what an earlier attempt did.
+
+    @param gym The gym site.
+    @param card The class, as it was read before.
+    @return The class with the state the page shows now.
+    @exception GymError The schedule could not be read or the class is no longer on it.
+    """
+    for current in gym.list_class_cards():
+        if current.card_id == card.card_id:
+            return current
+    raise GymError(constants.MESSAGE_CLASS_NOT_ON_SCHEDULE.format(card_id=card.card_id))
+
+
+def book_class_with_retry(
+    gym: GymPort,
+    card: ClassCard,
+    retries: int = constants.RETRY_COUNT,
+    sleep: Sleep = time.sleep,
+) -> ClassResult:
+    """@brief Book a class, repeating it when the site drops the request.
+
+    Before each repeat the class is read from the page again. When the button now shows the
+    state that the click should give, the click did go through although its confirmation
+    failed, so the class counts as newly booked or waitlisted and is not clicked again.
+
+    @param gym The gym site.
+    @param card The class, as read from the page.
+    @param retries The most attempts to make.
+    @param sleep Waits between two attempts.
+    @return The class and what happened.
+    @exception RetriesExhaustedError Every attempt failed.
+    """
+    attempts = 0
+
+    def attempt() -> ClassResult:
+        nonlocal attempts
+        attempts += 1
+        if attempts == 1:
+            return book_class(gym, card)
+        current = _read_current_card(gym, card)
+        action = decide_outcome_action(card)
+        if CONFIRMED_STATE.get(action) is current.state:
+            outcome = OUTCOME_OF_ACTION[action]
+            logger.info(format_outcome(outcome, card))
+            return ClassResult(card, outcome)
+        return book_class(gym, current)
+
+    description = constants.RETRY_DESCRIPTION_BOOK.format(name=card.name, day=card.day_label)
+    return retry(attempt, retries, description, sleep=sleep)
+
+
+def book_next_tuesday(
+    gym: GymPort, retries: int = constants.RETRY_COUNT, sleep: Sleep = time.sleep
+) -> ClassResult | None:
    """@brief Book the next Tuesday class at the target time.

    @param gym The gym site.
+    @param retries The most attempts to book the class.
+    @param sleep Waits between two attempts.
    @return The result, or None when the schedule has no such class.
-    @exception GymError The page could not be read or the booking failed.
+    @exception GymError The page could not be read or the booking failed on every attempt.
    """
    card = find_next_class(gym.list_class_cards(), constants.NEXT_CLASS_WEEKDAY)
    if card is None:
@@ -57,11 +115,14 @@ def book_next_tuesday(gym: GymPort) -> ClassResult | None:
            )
        )
        return None
-    return book_class(gym, card)
+    return book_class_with_retry(gym, card, retries, sleep)


 def book_target_classes(
-    gym: GymPort, on_result: Callable[[ClassResult], None] | None = None
+    gym: GymPort,
+    on_result: Callable[[ClassResult], None] | None = None,
+    retries: int = constants.RETRY_COUNT,
+    sleep: Sleep = time.sleep,
 ) -> list[ClassResult]:
    """@brief Book every Tuesday and Thursday class at the target time, in page order.

@@ -69,8 +130,10 @@ def book_target_classes(

    @param gym The gym site.
    @param on_result Called after each class was handled, for example to take a screenshot.
+    @param retries The most attempts to book one class.
+    @param sleep Waits between two attempts.
    @return What happened to each class; empty when the schedule has none.
-    @exception GymError The page could not be read or a booking failed.
+    @exception GymError The page could not be read or a booking failed on every attempt.
    """
    cards = find_target_classes(gym.list_class_cards())
    if not cards:
@@ -78,7 +141,7 @@ def book_target_classes(
        return []
    results: list[ClassResult] = []
    for card in cards:
-        result = book_class(gym, card)
+        result = book_class_with_retry(gym, card, retries, sleep)
        results.append(result)
        if on_result is not None:
            on_result(result)
@@ -30,6 +30,34 @@ def load_credentials(environment: Mapping[str, str]) -> Credentials:
    if not email or not password:
        raise MissingCredentialsError(
            f"Set {constants.ACCOUNT_EMAIL_ENV} and {constants.ACCOUNT_PASSWORD_ENV} "
-            "in the environment (they must not be empty)"
+            "in the environment or the env file (they must not be empty)"
        )
    return Credentials(email=email, password=password)
+
+
+def load_admin_credentials(environment: Mapping[str, str]) -> Credentials | None:
+    """@brief Read the optional admin account from the environment.
+
+    @param environment The environment variables, for example `os.environ`.
+    @return The credentials, or None when neither variable is set.
+    @exception MissingCredentialsError Only one of the two variables is set; the message
+        names the missing variable and never a value.
+    """
+    email = environment.get(constants.ADMIN_EMAIL_ENV, "").strip()
+    password = environment.get(constants.ADMIN_PASSWORD_ENV, "")
+    if not email and not password:
+        return None
+    if not email:
+        raise MissingCredentialsError(_admin_missing(constants.ADMIN_EMAIL_ENV))
+    if not password:
+        raise MissingCredentialsError(_admin_missing(constants.ADMIN_PASSWORD_ENV))
+    return Credentials(email=email, password=password)
+
+
+def _admin_missing(name: str) -> str:
+    """@brief Build the message for a half-set admin account.
+
+    @param name The variable that is missing or empty.
+    @return The message; it names the variable only.
+    """
+    return constants.MESSAGE_ADMIN_HALF_SET.format(name=name)
@@ -0,0 +1,67 @@
+"""@file retry.py
+@brief Lecture step 9: repeat a step that failed because the site dropped a request.
+"""
+
+import logging
+import time
+from collections.abc import Callable
+
+from snack_and_lift import constants
+from snack_and_lift.domain.errors import GymError, RetriesExhaustedError
+
+logger = logging.getLogger(__name__)
+
+## Waits for a number of seconds; tests pass a recorder so that they never really sleep.
+Sleep = Callable[[float], None]
+
+
+def retry[T](
+    func: Callable[[], T],
+    retries: int = constants.RETRY_COUNT,
+    description: str | None = None,
+    *,
+    delay: float = constants.RETRY_DELAY_SECONDS,
+    sleep: Sleep = time.sleep,
+) -> T:
+    """@brief Call a function until it succeeds, at most `retries` times.
+
+    Only a `GymError` is repeated: it is what a dropped request looks like. Any other error,
+    such as missing credentials, is a defect and passes through at once.
+
+    @param func The step, called without arguments.
+    @param retries The most attempts to make, at least 1.
+    @param description What the step does, for the log (never put a secret in it).
+    @param delay Seconds to wait between two attempts.
+    @param sleep Waits for a number of seconds.
+    @return What the function returned on its first successful attempt.
+    @exception ValueError `retries` is below 1.
+    @exception RetriesExhaustedError Every attempt raised a `GymError`; the last one is the cause.
+    """
+    if retries < 1:
+        raise ValueError(constants.MESSAGE_RETRIES_INVALID.format(retries=retries))
+    name = description or constants.RETRY_DEFAULT_DESCRIPTION
+    for attempt in range(1, retries + 1):
+        try:
+            result = func()
+        except GymError as error:
+            if attempt == retries:
+                message = constants.MESSAGE_RETRIES_EXHAUSTED.format(
+                    description=name, retries=retries, error=error
+                )
+                raise RetriesExhaustedError(message) from error
+            logger.warning(
+                constants.MESSAGE_RETRY_FAILED.format(
+                    attempt=attempt,
+                    retries=retries,
+                    description=name,
+                    error=str(error).rstrip("."),
+                )
+            )
+            sleep(delay)
+        else:
+            if attempt > 1:
+                logger.info(
+                    constants.MESSAGE_RETRY_SUCCEEDED.format(description=name, attempts=attempt)
+                )
+            return result
+    raise AssertionError("unreachable: the last attempt returns or raises")  # pragma: no cover
@@ -3,9 +3,12 @@
 """

 import logging
+import time

+from snack_and_lift import constants
 from snack_and_lift.application.credentials import Credentials
 from snack_and_lift.application.ports import GymPort
+from snack_and_lift.application.retry import Sleep, retry

 logger = logging.getLogger(__name__)

@@ -19,3 +22,25 @@ def login(gym: GymPort, credentials: Credentials) -> None:
    """
    gym.log_in(credentials.email, credentials.password)
    logger.info("✓ Logged in: the class schedule is shown")
+
+
+def login_with_retry(
+    gym: GymPort,
+    credentials: Credentials,
+    retries: int = constants.RETRY_COUNT,
+    sleep: Sleep = time.sleep,
+) -> None:
+    """@brief Log in, repeating the login when the site drops the request.
+
+    @param gym The gym site.
+    @param credentials The account to log in with.
+    @param retries The most attempts to make.
+    @param sleep Waits between two attempts.
+    @exception RetriesExhaustedError Every attempt failed.
+    """
+    retry(
+        lambda: login(gym, credentials),
+        retries,
+        constants.RETRY_DESCRIPTION_LOGIN,
+        sleep=sleep,
+    )
@@ -3,10 +3,12 @@
 """

 import logging
+import time
 from collections.abc import Sequence

 from snack_and_lift import constants
 from snack_and_lift.application.ports import GymPort
+from snack_and_lift.application.retry import Sleep, retry
 from snack_and_lift.domain.booking_entry import BookingEntry, is_target_entry
 from snack_and_lift.domain.errors import NoBookingsFoundError
 from snack_and_lift.domain.models import ClassResult
@@ -30,18 +32,42 @@ def get_my_bookings(gym: GymPort) -> list[BookingEntry]:
    return entries


-def verify_bookings(gym: GymPort, results: Sequence[ClassResult]) -> VerificationResult:
+def get_my_bookings_with_retry(
+    gym: GymPort, retries: int = constants.RETRY_COUNT, sleep: Sleep = time.sleep
+) -> list[BookingEntry]:
+    """@brief Read the My Bookings page, repeating it when the page fails to load.
+
+    @param gym The gym site.
+    @param retries The most attempts to make.
+    @param sleep Waits between two attempts.
+    @return Every confirmed booking and waitlist entry.
+    @exception RetriesExhaustedError Every attempt failed.
+    """
+    return retry(
+        lambda: get_my_bookings(gym), retries, constants.RETRY_DESCRIPTION_BOOKINGS, sleep=sleep
+    )
+
+
+def verify_bookings(
+    gym: GymPort,
+    results: Sequence[ClassResult],
+    retries: int = constants.RETRY_COUNT,
+    sleep: Sleep = time.sleep,
+) -> VerificationResult:
    """@brief Compare what the bot booked with the My Bookings page.

    The page is not opened when the bot handled no class.

    @param gym The gym site.
    @param results What happened to each class.
+    @param retries The most attempts to read the page.
+    @param sleep Waits between two attempts.
    @return The comparison of the expected entries with the Tuesday and Thursday 6pm entries.
-    @exception NoBookingsFoundError Classes were handled but the page shows no entry.
-    @exception GymError The page could not be opened or read.
+    @exception RetriesExhaustedError The page could not be read on any attempt.
    """
    if not results:
        return build_verification(results, [])
-    entries = [entry for entry in get_my_bookings(gym) if is_target_entry(entry)]
+    entries = [
+        entry for entry in get_my_bookings_with_retry(gym, retries, sleep) if is_target_entry(entry)
+    ]
    return build_verification(results, entries)
@@ -6,14 +6,20 @@ import argparse
 import logging
 import os
 import sys
+import time
 from collections.abc import Callable, Sequence
 from pathlib import Path

 from snack_and_lift import constants
 from snack_and_lift.application.booking import book_next_tuesday, book_target_classes
-from snack_and_lift.application.credentials import Credentials, load_credentials
+from snack_and_lift.application.credentials import (
+    Credentials,
+    load_admin_credentials,
+    load_credentials,
+)
 from snack_and_lift.application.ports import GymPort
-from snack_and_lift.application.session import login
+from snack_and_lift.application.retry import Sleep
+from snack_and_lift.application.session import login_with_retry
 from snack_and_lift.application.verification import verify_bookings
 from snack_and_lift.domain.errors import GymError, LoginFailedError, MissingCredentialsError
 from snack_and_lift.domain.models import ClassResult
@@ -21,6 +27,7 @@ from snack_and_lift.domain.report import format_detail_list, format_summary
 from snack_and_lift.domain.verification import format_verification
 from snack_and_lift.infrastructure.browser import create_driver, default_profile_dir
 from snack_and_lift.infrastructure.console_recorder import ConsoleRecorder
+from snack_and_lift.infrastructure.env_file import read_env_file
 from snack_and_lift.infrastructure.selenium_gym import SeleniumGym

 ## Builds the gym site from (headless, profile folder, keep open).
@@ -40,6 +47,22 @@ def selenium_gym_factory(headless: bool, profile_dir: Path, keep_open: bool) ->
    return SeleniumGym(create_driver(headless, profile_dir, keep_open))


+def _at_least_one(text: str) -> int:
+    """@brief Read a whole number of at least 1 for an option.
+
+    @param text The text of the option.
+    @return The number.
+    @exception argparse.ArgumentTypeError The text is not a whole number of at least 1.
+    """
+    try:
+        number = int(text)
+    except ValueError:
+        number = 0
+    if number < 1:
+        raise argparse.ArgumentTypeError(f"{text!r} is not a whole number of at least 1")
+    return number
+
+
 def parse_args(argv: Sequence[str] | None = None) -> argparse.Namespace:
    """@brief Read the command line.

@@ -50,7 +73,8 @@ def parse_args(argv: Sequence[str] | None = None) -> argparse.Namespace:
        prog="snack-and-lift",
        description="Books the Tuesday and Thursday 6pm classes at the Snack and Lift gym.",
        epilog=(
-            "exit codes: 0 done, 1 a step failed, 2 credentials missing, "
+            "exit codes: 0 done, 1 a step failed (also when every retry failed), "
+            "2 credentials missing in the environment or the env file, "
            "3 the My Bookings page does not match the bookings made"
        ),
    )
@@ -82,6 +106,26 @@ def parse_args(argv: Sequence[str] | None = None) -> argparse.Namespace:
        action="store_true",
        help="book only the next Tuesday 6pm class (lecture steps 3 and 4); skips the verification",
    )
+    parser.add_argument(
+        "--retries",
+        type=_at_least_one,
+        default=constants.RETRY_COUNT,
+        metavar="N",
+        help=(
+            "attempts for the login, each booking and the My Bookings page when the site "
+            f"drops a request (default: {constants.RETRY_COUNT}, at least 1)"
+        ),
+    )
+    parser.add_argument(
+        "--env-file",
+        type=Path,
+        default=Path(constants.ENV_FILE_NAME),
+        metavar="PATH",
+        help=(
+            "file with the gym accounts as KEY=VALUE lines "
+            f"(default: ./{constants.ENV_FILE_NAME}); a real environment variable wins"
+        ),
+    )
    args = parser.parse_args(argv)
    if args.keep_open is None:
        args.keep_open = not args.headless
@@ -121,7 +165,11 @@ def _save_console(gym: GymPort, screenshot_dir: Path | None, text: str, name: st


 def _run(
-    gym: GymPort, args: argparse.Namespace, recorder: ConsoleRecorder, credentials: Credentials
+    gym: GymPort,
+    args: argparse.Namespace,
+    recorder: ConsoleRecorder,
+    credentials: Credentials,
+    sleep: Sleep,
 ) -> bool:
    """@brief Open the site, log in, book the classes and verify them, saving the step images.

@@ -133,15 +181,16 @@ def _run(
    @param args The parsed options.
    @param recorder The recorded console lines, drawn for step 4.
    @param credentials The account to log in with.
+    @param sleep Waits between two attempts of a retried step.
    @return False when the My Bookings page does not match the bookings made, else True.
-    @exception GymError A step failed.
+    @exception GymError A step failed on every attempt.
    """
    gym.open_site()
    _save_step(gym, args.screenshot_dir, "step-1.png")
-    login(gym, credentials)
+    login_with_retry(gym, credentials, args.retries, sleep)
    _save_step(gym, args.screenshot_dir, "step-2.png")
    if args.first_tuesday_only:
-        book_next_tuesday(gym)
+        book_next_tuesday(gym, args.retries, sleep)
        _save_step(gym, args.screenshot_dir, constants.SCREENSHOT_STEP_3)
        _save_console(gym, args.screenshot_dir, recorder.text, constants.SCREENSHOT_STEP_4)
        return True
@@ -153,7 +202,7 @@ def _run(
            _save_step(gym, args.screenshot_dir, constants.SCREENSHOT_STEP_3)
            window_saved = True

-    results = book_target_classes(gym, save_window_once)
+    results = book_target_classes(gym, save_window_once, args.retries, sleep)
    if not window_saved:
        _save_step(gym, args.screenshot_dir, constants.SCREENSHOT_STEP_3)
    _save_console(gym, args.screenshot_dir, recorder.text, constants.SCREENSHOT_STEP_4)
@@ -163,22 +212,25 @@ def _run(
    detail = format_detail_list(results)
    logger.info("%s", detail)
    _save_console(gym, args.screenshot_dir, f"{summary}\n{detail}", constants.SCREENSHOT_STEP_6)
-    verification = verify_bookings(gym, results)
+    verification = verify_bookings(gym, results, args.retries, sleep)
    _save_step(gym, args.screenshot_dir, constants.SCREENSHOT_STEP_7)
    verification_text = format_verification(verification)
    logger.info("%s", verification_text)
    _save_console(gym, args.screenshot_dir, verification_text, constants.SCREENSHOT_STEP_8)
+    _save_console(gym, args.screenshot_dir, recorder.text, constants.SCREENSHOT_STEP_9)
    return verification.is_success


 def main(
    argv: Sequence[str] | None = None,
    gym_factory: GymFactory = selenium_gym_factory,
+    sleep: Sleep = time.sleep,
 ) -> int:
-    """@brief Start the bot: open the site, log in and book the next Tuesday 6pm class.
+    """@brief Start the bot: open the site, log in, book the classes and verify them.

    @param argv Arguments without the program name; None reads `sys.argv`.
    @param gym_factory Builds the gym site; tests pass a fake.
+    @param sleep Waits between two attempts of a retried step; tests pass a recorder.
    @return Process exit code: 0 done, 1 a step failed, 2 credentials missing, 3 mismatch.
    """
    logging.basicConfig(level=logging.INFO, format="%(message)s")
@@ -187,29 +239,34 @@ def main(
    recorder = ConsoleRecorder()
    root.addHandler(recorder)
    try:
-        return _main(argv, gym_factory, recorder)
+        return _main(argv, gym_factory, recorder, sleep)
    finally:
        root.removeHandler(recorder)


-def _main(argv: Sequence[str] | None, gym_factory: GymFactory, recorder: ConsoleRecorder) -> int:
+def _main(
+    argv: Sequence[str] | None, gym_factory: GymFactory, recorder: ConsoleRecorder, sleep: Sleep
+) -> int:
    """@brief The body of `main`, run while the recorder is attached.

    @param argv Arguments without the program name.
    @param gym_factory Builds the gym site.
    @param recorder The console recorder.
+    @param sleep Waits between two attempts of a retried step.
    @return Process exit code.
    """
    _use_utf8_output()
    args = parse_args(argv)
    try:
-        credentials = load_credentials(os.environ)
+        environment = {**read_env_file(args.env_file), **os.environ}
+        credentials = load_credentials(environment)
+        load_admin_credentials(environment)  # load-only for now; nothing uses it yet
    except MissingCredentialsError as error:
        logger.error("%s", error)
        return 2
    gym = gym_factory(args.headless, args.profile_dir, args.keep_open)
    try:
-        verified = _run(gym, args, recorder, credentials)
+        verified = _run(gym, args, recorder, credentials, sleep)
    except LoginFailedError as error:
        logger.error("Login failed: %s", error)
        return 1
@@ -18,6 +18,31 @@ WAIT_TIMEOUT_SECONDS: Final[int] = 10
 ## Attempts of the `retry` wrapper before it gives up (lecture step 9).
 RETRY_COUNT: Final[int] = 7

+## Seconds to wait between two attempts of the `retry` wrapper.
+RETRY_DELAY_SECONDS: Final[float] = 0.5
+
+## Name of a retried step that was given no description.
+RETRY_DEFAULT_DESCRIPTION: Final[str] = "operation"
+
+## Descriptions of the retried steps, as they appear in the log.
+RETRY_DESCRIPTION_LOGIN: Final[str] = "login"
+RETRY_DESCRIPTION_BOOK: Final[str] = "book {name} on {day}"
+RETRY_DESCRIPTION_BOOKINGS: Final[str] = "get my bookings"
+
+## Messages of the `retry` wrapper.
+MESSAGE_RETRY_FAILED: Final[str] = (
+    "⚠ Attempt {attempt}/{retries} failed: {description}: {error}. Retrying..."
+)
+MESSAGE_RETRY_SUCCEEDED: Final[str] = "✓ {description} succeeded after {attempts} attempts"
+MESSAGE_RETRIES_EXHAUSTED: Final[str] = "{description} failed after {retries} attempts: {error}"
+MESSAGE_RETRIES_INVALID: Final[str] = "retries must be at least 1, got {retries}"
+
+## Error raised when a class card is no longer on the schedule while a booking is repeated.
+MESSAGE_CLASS_NOT_ON_SCHEDULE: Final[str] = "class {card_id} not found on the schedule"
+
+## File name of the screenshot of lecture step 9 (the console with the retry warnings).
+SCREENSHOT_STEP_9: Final[str] = "step-9.png"
+
 ## Weekdays of the classes the bot books, as the site abbreviates them.
 TARGET_WEEKDAYS: Final[tuple[str, ...]] = ("Tue", "Thu")

@@ -45,6 +70,19 @@ ACCOUNT_EMAIL_ENV: Final[str] = "SNACK_AND_LIFT_EMAIL"
 ## Name of the environment variable that holds the account password.
 ACCOUNT_PASSWORD_ENV: Final[str] = "SNACK_AND_LIFT_PASSWORD"

+## Names of the environment variables that hold the admin account (read, not used yet).
+ADMIN_EMAIL_ENV: Final[str] = "SNACK_AND_LIFT_ADMIN_EMAIL"
+ADMIN_PASSWORD_ENV: Final[str] = "SNACK_AND_LIFT_ADMIN_PASSWORD"
+
+## Error for an admin account with one variable only; {name} is the missing variable.
+MESSAGE_ADMIN_HALF_SET: Final[str] = (
+    "Set {name} too: the admin account needs both "
+    "SNACK_AND_LIFT_ADMIN_EMAIL and SNACK_AND_LIFT_ADMIN_PASSWORD"
+)
+
+## Default file with the gym accounts, relative to the current directory.
+ENV_FILE_NAME: Final[str] = ".env.gym"
+
 ## Folder where screenshots of a container run are written.
 SCREENSHOT_DIR: Final[str] = "docs/images"

@@ -203,3 +241,26 @@ VERIFY_SUCCESS: Final[str] = "✅ SUCCESS: All bookings verified!"
 VERIFY_MISMATCH: Final[str] = "❌ MISMATCH: Missing {count} bookings"
 VERIFY_LINE_MISSING: Final[str] = "  ✗ Missing: {entry}"
 VERIFY_LINE_UNEXPECTED: Final[str] = "  ? Unexpected: {entry}"
+VERIFY_LINE_ALSO_BOOKED: Final[str] = "  i Also booked (not checked): {entry}"
+
+## Month abbreviations of the dates on the pages, January first.
+MONTH_ABBREVIATIONS: Final[tuple[str, ...]] = (
+    "Jan",
+    "Feb",
+    "Mar",
+    "Apr",
+    "May",
+    "Jun",
+    "Jul",
+    "Aug",
+    "Sep",
+    "Oct",
+    "Nov",
+    "Dec",
+)
+
+## Days of a year, with the 29th of February, used to compare dates that have no year.
+DAYS_IN_DATE_CYCLE: Final[int] = 366
+
+## A leap year, so that the 29th of February is a valid date.
+DATE_CYCLE_YEAR: Final[int] = 2000
@@ -23,6 +23,10 @@ class PageFormatError(GymError):
    """@brief A text read from the page (a day label or a time) has an unexpected format."""


+class RetriesExhaustedError(GymError):
+    """@brief A step failed on every attempt; the cause is the error of the last attempt."""
+
+
 class NoBookingsFoundError(GymError):
    """@brief The My Bookings page shows no booking where the bot expects some.

@@ -5,9 +5,11 @@
 from collections import Counter
 from collections.abc import Sequence
 from dataclasses import dataclass
+from datetime import date

 from snack_and_lift import constants
 from snack_and_lift.domain.booking_entry import BookingEntry
+from snack_and_lift.domain.errors import PageFormatError
 from snack_and_lift.domain.models import ClassResult, Outcome

 _WAITLISTED_OUTCOMES = (Outcome.NEW_WAITLIST, Outcome.ALREADY_WAITLISTED)
@@ -20,8 +22,11 @@ _Key = tuple[str, str, bool]
 class VerificationResult:
    """@brief What the My Bookings page shows compared with what the bot did.

-    `verified`, `missing` and `unexpected` hold display names, with the waitlist suffix for a
-    waitlist entry; `missing` and `unexpected` also say the day (`Spin Class on Tue, Aug 12`).
+    `verified`, `missing`, `unexpected` and `also_booked` hold display names, with the waitlist
+    suffix for a waitlist entry; all but `verified` also say the day (`Spin Class on Tue, Aug 12`).
+    Only entries from the earliest to the latest class the bot handled are checked: `found`
+    counts them, and `also_booked` lists the entries outside that window, which never cause
+    a mismatch.
    """

    expected: int
@@ -29,6 +34,7 @@ class VerificationResult:
    verified: list[str]
    missing: list[str]
    unexpected: list[str]
+    also_booked: list[str]

    @property
    def is_success(self) -> bool:
@@ -57,6 +63,46 @@ def _describe(key: _Key) -> str:
    return f"{_display_name(name, waitlisted)} on {day}"


+def _day_number(date_text: str) -> int:
+    """@brief The position of a date (`Tue, Aug 12`) in a year; the page shows no year.
+
+    @exception PageFormatError The text has no valid month and day.
+    """
+    month_text, _, day_text = date_text.partition(", ")[2].partition(" ")
+    try:
+        month = constants.MONTH_ABBREVIATIONS.index(month_text) + 1
+        day = date(constants.DATE_CYCLE_YEAR, month, int(day_text))
+    except ValueError as error:
+        raise PageFormatError(f"Unexpected date: {date_text!r}") from error
+    return day.timetuple().tm_yday
+
+
+def _offset(day: int, start: int) -> int:
+    """@brief The days from `start` to `day`, counting over the turn of the year."""
+    return (day - start) % constants.DAYS_IN_DATE_CYCLE
+
+
+def _window(keys: Sequence[_Key]) -> tuple[int, int] | None:
+    """@brief The first day and the length in days of the shortest span holding all the keys.
+
+    @return None when there is no key.
+    """
+    days = [_day_number(key[1]) for key in keys]
+    spans = [(max(_offset(day, start) for day in days), start) for start in days]
+    if not spans:
+        return None
+    span, start = min(spans)
+    return start, span
+
+
+def _is_in_window(key: _Key, window: tuple[int, int] | None) -> bool:
+    """@brief Tell whether the date of the key lies in the window, ends included."""
+    if window is None:
+        return False
+    start, span = window
+    return _offset(_day_number(key[1]), start) <= span
+
+
 def build_verification(
    results: Sequence[ClassResult], entries: Sequence[BookingEntry]
 ) -> VerificationResult:
@@ -64,11 +110,16 @@ def build_verification(

    @param results What happened to each class; waitlisted when the outcome is a waitlist one.
    @param entries The Tuesday and Thursday 6pm entries found on the page.
-    @return The comparison; the counts are the numbers of results and of entries.
+    @return The comparison; only the entries from the earliest to the latest class handled
+        are checked, the others are `also_booked`, and no class handled means no check.
    @exception PageFormatError An entry or a card has no valid date.
    """
    expected = _expected_keys(results)
-    remaining = Counter(_entry_keys(entries))
+    window = _window(expected)
+    keys = _entry_keys(entries)
+    checked = [key for key in keys if _is_in_window(key, window)]
+    also_booked = [_describe(key) for key in keys if not _is_in_window(key, window)]
+    remaining = Counter(checked)
    verified: list[str] = []
    missing: list[str] = []
    for key in expected:
@@ -78,7 +129,9 @@ def build_verification(
        else:
            missing.append(_describe(key))
    unexpected = [_describe(key) for key in remaining.elements()]
-    return VerificationResult(len(expected), len(entries), verified, missing, unexpected)
+    return VerificationResult(
+        len(expected), len(checked), verified, missing, unexpected, also_booked
+    )


 def _count_text(template: str, count: int) -> str:
@@ -99,10 +152,11 @@ def format_verification(result: VerificationResult) -> str:
    lines.append(constants.VERIFY_RESULT_TITLE)
    lines.append(_count_text(constants.VERIFY_EXPECTED, result.expected))
    lines.append(_count_text(constants.VERIFY_FOUND, result.found))
+    lines.extend(constants.VERIFY_LINE_ALSO_BOOKED.format(entry=t) for t in result.also_booked)
    if result.is_success:
        lines.append(constants.VERIFY_SUCCESS)
        return "\n".join(lines)
-    lines.append(constants.VERIFY_MISMATCH.format(count=result.expected - result.found))
+    lines.append(constants.VERIFY_MISMATCH.format(count=len(result.missing)))
    lines.extend(constants.VERIFY_LINE_MISSING.format(entry=text) for text in result.missing)
    lines.extend(constants.VERIFY_LINE_UNEXPECTED.format(entry=text) for text in result.unexpected)
    return "\n".join(lines)
@@ -0,0 +1,44 @@
+"""@file env_file.py
+@brief Reads `KEY=VALUE` lines from an env file (never logs a value).
+"""
+
+from pathlib import Path
+
+## Prefix a line may carry so the file can also be sourced by a shell.
+_EXPORT_PREFIX = "export "
+
+## Quote characters that may surround a value.
+_QUOTES = ("'", '"')
+
+
+def _unquote(value: str) -> str:
+    """@brief Strip one pair of matching surrounding quotes.
+
+    @param value The text after the `=`.
+    @return The value without the surrounding pair, otherwise unchanged.
+    """
+    if len(value) >= 2 and value[0] == value[-1] and value[0] in _QUOTES:
+        return value[1:-1]
+    return value
+
+
+def read_env_file(path: Path) -> dict[str, str]:
+    """@brief Read the variables of an env file.
+
+    Blank lines, `#` comments, lines without `=` and an empty key are skipped.
+
+    @param path The file.
+    @return The variables; empty when the file does not exist.
+    """
+    if not path.is_file():
+        return {}
+    variables: dict[str, str] = {}
+    for line in path.read_text(encoding="utf-8").splitlines():
+        stripped = line.strip()
+        if not stripped or stripped.startswith("#") or "=" not in stripped:
+            continue
+        key, _, value = stripped.partition("=")
+        key = key.strip().removeprefix(_EXPORT_PREFIX).strip()
+        if key:
+            variables[key] = _unquote(value.strip())
+    return variables
@@ -88,11 +88,18 @@ class SeleniumGym:
        @exception LoginFailedError The site shows a login error.
        @exception GymError The page did not behave as expected.
        """
-        self._wait.until(ec.element_to_be_clickable((By.ID, constants.NAV_LOGIN_BUTTON_ID))).click()
-        self._wait.until(
+        # After a failed login the form is still shown, so a repeated login skips the nav button.
+        if not self._driver.find_elements(By.ID, constants.LOGIN_EMAIL_INPUT_ID):
+            self._wait.until(
+                ec.element_to_be_clickable((By.ID, constants.NAV_LOGIN_BUTTON_ID))
+            ).click()
+        email_input = self._wait.until(
            ec.presence_of_element_located((By.ID, constants.LOGIN_EMAIL_INPUT_ID))
-        ).send_keys(email)
-        self._driver.find_element(By.ID, constants.LOGIN_PASSWORD_INPUT_ID).send_keys(password)
+        )
+        password_input = self._driver.find_element(By.ID, constants.LOGIN_PASSWORD_INPUT_ID)
+        for field, value in ((email_input, email), (password_input, password)):
+            field.clear()
+            field.send_keys(value)
        self._driver.find_element(By.ID, constants.LOGIN_SUBMIT_BUTTON_ID).click()
        self._wait.until(
            ec.any_of(
@@ -12,8 +12,19 @@
    // Dummy account of this fixture only; it is not a real account.
    const ACCEPTED_EMAIL = "fixture@example.invalid";
    const ACCEPTED_PASSWORD = "fixture-password";
+    // Imitates the network simulation: ?fail=N drops the first N submits after a short delay.
+    let failuresLeft = parseInt(new URLSearchParams(window.location.search).get("fail") || "0", 10);
    document.getElementById("login-form").addEventListener("submit", function (event) {
      event.preventDefault();
+      document.getElementById("message-area").innerHTML = "";
+      if (failuresLeft > 0) {
+        failuresLeft -= 1;
+        setTimeout(function () {
+          document.getElementById("message-area").innerHTML =
+            '<div id="error-message">Network request failed. Please try again.</div>';
+        }, 100);
+        return;
+      }
      const email = document.getElementById("email-input").value;
      const password = document.getElementById("password-input").value;
      if (email === ACCEPTED_EMAIL && password === ACCEPTED_PASSWORD) {
@@ -10,21 +10,37 @@
  </div>
  <script>
    // Imitates the site: a loading text, then the page for the state in the query
-    // (?state=empty or ?state=error; anything else shows two bookings and a waitlist entry).
-    const state = new URLSearchParams(window.location.search).get("state");
+    // (?state=empty or ?state=error; anything else shows two bookings and a waitlist entry;
+    // ?extra=future adds three more Tuesday and Thursday 6pm bookings after the others).
+    const params = new URLSearchParams(window.location.search);
+    const state = params.get("state");
+    // ?fail=N shows the error for the first N loads of the page in this tab (the network
+    // simulation); the count lives in sessionStorage because every load starts a new page.
+    const loads = parseInt(sessionStorage.getItem("loads") || "0", 10);
+    const dropped = loads < parseInt(params.get("fail") || "0", 10);
+    sessionStorage.setItem("loads", String(loads + 1));
    // The nav link keeps the state, so opening the page from here shows the same state again.
    document.getElementById("my-bookings-link").href = "my_bookings.html" + window.location.search;
+    function futureBookings() {
+      const whens = ["Thu, Aug 21, 6:00 PM", "Tue, Aug 26, 6:00 PM", "Thu, Aug 28, 6:00 PM"];
+      return whens.map(function (when, i) {
+        return '<div id="booking-card-' + (13 + i) + '" data-booking-status="confirmed">' +
+          '<h3 id="booking-class-name-' + (13 + i) + '">Spin Class</h3>' +
+          '<p><strong>When:</strong> ' + when + '</p></div>';
+      }).join("");
+    }
    const page = document.getElementById("my-bookings-page");
    const content = document.getElementById("content");
    setTimeout(function () {
      document.getElementById("loading-placeholder").remove();
      page.dataset.loading = "false";
-      if (state === "error") {
+      if (state === "error" || dropped) {
        content.innerHTML = '<div id="error-message">Failed to load bookings</div>';
      } else if (state === "empty") {
        content.innerHTML = '<div id="empty-bookings-state">You have no bookings yet</div>';
      } else {
-        page.dataset.bookingsCount = "2";
+        const extra = params.get("extra") === "future" ? futureBookings() : "";
+        page.dataset.bookingsCount = params.get("extra") === "future" ? "5" : "2";
        page.dataset.waitlistCount = "1";
        content.innerHTML =
          '<div id="confirmed-bookings-section">' +
@@ -34,6 +50,7 @@
          '<div id="booking-card-12" data-booking-status="confirmed">' +
          '<h3 id="booking-class-name-12">Yoga Class</h3>' +
          '<p><strong>When:</strong> Wed, Aug 13, 7:00 AM</p></div>' +
+          extra +
          '</div>' +
          '<div id="waitlist-section">' +
          '<div id="waitlist-card-21" data-booking-status="waitlisted">' +
@@ -46,7 +46,7 @@
        <p id="class-time-301"><strong>Time:</strong> 7:00 AM</p>
        <button id="book-button-301">Book Class</button>
      </div>
-      <div id="class-card-302" data-class-type="spin" data-user-booked="false" data-fail-once="true">
+      <div id="class-card-302" data-class-type="spin" data-user-booked="false" data-fail-times="1">
        <h3 id="class-name-302">Spin Class</h3>
        <p id="class-time-302"><strong>Time:</strong> 6:00 PM</p>
        <button id="book-button-302">Book Class</button>
@@ -55,6 +55,11 @@
  </div>
  <script>
    // Imitates the site: a busy text, then the new state; card 302 fails on its first click.
+    // ?fail=N makes every button fail on its first N clicks (the network simulation).
+    const failParam = parseInt(new URLSearchParams(window.location.search).get("fail") || "0", 10);
+    document.querySelectorAll("div[id^='class-card-']").forEach(function (card) {
+      card.dataset.failLeft = String(Math.max(failParam, parseInt(card.dataset.failTimes || "0", 10)));
+    });
    document.querySelectorAll("button[id^='book-button-']").forEach(function (button) {
      button.addEventListener("click", function () {
        const id = button.id.replace("book-button-", "");
@@ -66,8 +71,8 @@
        button.textContent = waitlist ? "Joining..." : "Booking...";
        button.disabled = true;
        setTimeout(function () {
-          if (card.dataset.failOnce === "true") {
-            card.dataset.failOnce = "false";
+          if (parseInt(card.dataset.failLeft, 10) > 0) {
+            card.dataset.failLeft = String(parseInt(card.dataset.failLeft, 10) - 1);
            button.textContent = oldText;
            button.disabled = false;
            const box = document.createElement("div");
@@ -8,10 +8,17 @@ from pathlib import Path

 import pytest

-from snack_and_lift.application.booking import book_target_classes
-from snack_and_lift.application.verification import get_my_bookings
+from snack_and_lift.application.booking import book_class_with_retry, book_target_classes
+from snack_and_lift.application.credentials import Credentials
+from snack_and_lift.application.session import login_with_retry
+from snack_and_lift.application.verification import get_my_bookings, get_my_bookings_with_retry
 from snack_and_lift.domain.booking_entry import BookingEntry
-from snack_and_lift.domain.errors import GymError, LoginFailedError, NoBookingsFoundError
+from snack_and_lift.domain.errors import (
+    GymError,
+    LoginFailedError,
+    NoBookingsFoundError,
+    RetriesExhaustedError,
+)
 from snack_and_lift.domain.models import Outcome
 from snack_and_lift.infrastructure.browser import create_driver
 from snack_and_lift.infrastructure.selenium_gym import SeleniumGym
@@ -119,13 +126,11 @@ def test_save_text_screenshot_writes_a_png_and_keeps_the_page(
 def test_book_target_classes_books_every_tuesday_and_thursday_class(
    schedule_gym: SeleniumGym,
 ) -> None:
-    with pytest.raises(GymError, match="Network error"):
-        book_target_classes(schedule_gym)
-
-    results = book_target_classes(schedule_gym)
+    # Card 302 drops its first click; the booking repeats it and the page confirms the second.
+    results = book_target_classes(schedule_gym, sleep=lambda _: None)

    assert [(r.card.card_id, r.outcome) for r in results] == [
-        ("201", Outcome.ALREADY_WAITLISTED),
+        ("201", Outcome.NEW_WAITLIST),
        ("302", Outcome.NEW_BOOKING),
    ]

@@ -170,6 +175,20 @@ def test_list_my_bookings_reads_confirmed_and_waitlist_entries(
    ]


+def test_get_my_bookings_returns_bookings_listed_after_the_others(
+    open_my_bookings_page: Callable[[str], SeleniumGym],
+) -> None:
+    entries = get_my_bookings(open_my_bookings_page("?extra=future"))
+
+    assert [entry.when_text for entry in entries][-4:] == [
+        "Thu, Aug 21, 6:00 PM",
+        "Tue, Aug 26, 6:00 PM",
+        "Thu, Aug 28, 6:00 PM",
+        "Thu, Aug 14, 6:00 PM",
+    ]
+    assert len(entries) == 6
+
+
 def test_list_my_bookings_returns_nothing_for_the_empty_state(
    open_my_bookings_page: Callable[[str], SeleniumGym],
 ) -> None:
@@ -188,3 +207,80 @@ def test_list_my_bookings_raises_with_the_text_of_the_error_box(
 ) -> None:
    with pytest.raises(GymError, match="Failed to load bookings"):
        open_my_bookings_page("?state=error").list_my_bookings()
+
+
+FIXTURE_LOGIN = (Path(__file__).parent / "fixtures" / "login.html").as_uri()
+
+
+@pytest.fixture
+def open_fixture(tmp_path: Path) -> Iterator[Callable[[str], SeleniumGym]]:
+    """@brief Opens a fixture page by its address, with the query that sets its failures."""
+    driver = create_driver(headless=True, profile_dir=tmp_path / "profile", keep_open=False)
+
+    def open_page(url: str) -> SeleniumGym:
+        driver.get(url)
+        return SeleniumGym(driver, base_url=url)
+
+    try:
+        yield open_page
+    finally:
+        driver.quit()
+
+
+def test_login_with_retry_logs_in_after_two_dropped_requests(
+    open_fixture: Callable[[str], SeleniumGym],
+) -> None:
+    gym = open_fixture(FIXTURE_LOGIN + "?fail=2")
+    sleeps: list[float] = []
+
+    login_with_retry(gym, Credentials(FIXTURE_EMAIL, FIXTURE_PASSWORD), sleep=sleeps.append)
+
+    assert len(sleeps) == 2
+
+
+def test_login_with_retry_gives_up_when_every_request_is_dropped(
+    open_fixture: Callable[[str], SeleniumGym],
+) -> None:
+    gym = open_fixture(FIXTURE_LOGIN + "?fail=5")
+
+    with pytest.raises(RetriesExhaustedError, match="Network request failed"):
+        login_with_retry(
+            gym, Credentials(FIXTURE_EMAIL, FIXTURE_PASSWORD), retries=3, sleep=lambda _: None
+        )
+
+
+def test_book_class_with_retry_books_after_two_dropped_clicks(
+    open_fixture: Callable[[str], SeleniumGym],
+) -> None:
+    gym = open_fixture(FIXTURE_SCHEDULE + "?fail=2")
+    card = next(card for card in gym.list_class_cards() if card.card_id == "302")
+
+    result = book_class_with_retry(gym, card, sleep=lambda _: None)
+
+    assert result.outcome is Outcome.NEW_BOOKING
+    assert next(c for c in gym.list_class_cards() if c.card_id == "302").button_text == "Booked"
+
+
+def test_book_target_classes_completes_when_every_class_drops_two_clicks(
+    open_fixture: Callable[[str], SeleniumGym],
+) -> None:
+    gym = open_fixture(FIXTURE_SCHEDULE + "?fail=2")
+
+    results = book_target_classes(gym, sleep=lambda _: None)
+
+    assert [(r.card.card_id, r.outcome) for r in results] == [
+        ("201", Outcome.NEW_WAITLIST),
+        ("302", Outcome.NEW_BOOKING),
+    ]
+
+
+def test_get_my_bookings_with_retry_reads_the_page_after_two_failed_loads(
+    open_fixture: Callable[[str], SeleniumGym],
+) -> None:
+    gym = open_fixture(FIXTURE_MY_BOOKINGS + "?fail=2")
+    sleeps: list[float] = []
+
+    entries = get_my_bookings_with_retry(gym, sleep=sleeps.append)
+
+    assert [entry.name for entry in entries] == ["Spin Class", "Yoga Class", "HIIT Class"]
+    assert sleeps
@@ -2,6 +2,8 @@
@brief A fake of the GymPort for the unit tests.
 """

+import random
+from dataclasses import replace
 from pathlib import Path

 from snack_and_lift import constants
@@ -32,11 +34,20 @@ def _entry_of(card: ClassCard) -> BookingEntry | None:
    return BookingEntry(card.name, f"{card.date_text}, {time_text}", flag)


+NETWORK_ERROR_TEXT = "Network request failed. Please try again."
+
+
 class FakeGym:
    """@brief Records the calls; can refuse the login, serve cards and fail a click.

    Unless `bookings` is given, the My Bookings page shows every card that is, or by the
-    click becomes, booked or waitlisted.
+    click becomes, booked or waitlisted. A successful click turns the card's button into its
+    new state, so a card read again shows what the click did.
+
+    `fail_first` is a deterministic failure schedule: it maps the name of a port method
+    (`log_in`, `press_class_button`, `open_my_bookings`, `list_my_bookings`) to the number of
+    its first calls that fail like a dropped request. With `click_goes_through`, a failing
+    `press_class_button` still turns the button (the click arrived, the answer did not).
    """

    def __init__(
@@ -47,6 +58,8 @@ class FakeGym:
        press_error: str | None = None,
        bookings: list[BookingEntry] | None = None,
        bookings_error: str | None = None,
+        fail_first: dict[str, int] | None = None,
+        click_goes_through: bool = False,
    ) -> None:
        self.login_error = login_error
        self.cards = cards or []
@@ -54,6 +67,8 @@ class FakeGym:
        self.press_error = press_error
        self.bookings = bookings
        self.bookings_error = bookings_error
+        self.click_goes_through = click_goes_through
+        self._failures_left = dict(fail_first or {})
        self.calls: list[str] = []
        self.logins: list[tuple[str, str]] = []
        self.screenshots: list[Path] = []
@@ -64,11 +79,29 @@ class FakeGym:
    def open_site(self) -> None:
        self.calls.append("open_site")

+    def _should_fail(self, name: str) -> bool:
+        """@brief Whether this call is one of the scheduled failures; counts it."""
+        left = self._failures_left.get(name, 0)
+        if left > 0:
+            self._failures_left[name] = left - 1
+            return True
+        return False
+
+    def _turn_button(self, card_id: str) -> str:
+        """@brief Turn the button of a card to the state its click gives; return the new text."""
+        index = next(i for i, card in enumerate(self.cards) if card.card_id == card_id)
+        card = self.cards[index]
+        new_text = _PRESS_RESULTS[card.button_text]
+        self.cards[index] = replace(card, button_text=new_text)
+        return new_text
+
    def log_in(self, email: str, password: str) -> None:
        self.calls.append("log_in")
        self.logins.append((email, password))
        if self.login_error is not None:
            raise LoginFailedError(self.login_error)
+        if self._should_fail("log_in"):
+            raise LoginFailedError(NETWORK_ERROR_TEXT)

    def list_class_cards(self) -> list[ClassCard]:
        self.calls.append("list_class_cards")
@@ -81,14 +114,21 @@ class FakeGym:
            raise GymError(self.press_error)
        if self.press_result is not None:
            return self.press_result
-        card = next(card for card in self.cards if card.card_id == card_id)
-        return _PRESS_RESULTS[card.button_text]
+        if self._should_fail("press_class_button"):
+            if self.click_goes_through:
+                self._turn_button(card_id)
+            raise GymError(NETWORK_ERROR_TEXT)
+        return self._turn_button(card_id)

    def open_my_bookings(self) -> None:
        self.calls.append("open_my_bookings")
+        if self._should_fail("open_my_bookings"):
+            raise GymError(NETWORK_ERROR_TEXT)

    def list_my_bookings(self) -> list[BookingEntry]:
        self.calls.append("list_my_bookings")
+        if self._should_fail("list_my_bookings"):
+            raise GymError(NETWORK_ERROR_TEXT)
        if self.bookings_error is not None:
            raise GymError(self.bookings_error)
        if self.bookings is not None:
@@ -107,3 +147,30 @@ class FakeGym:
    def close(self) -> None:
        self.calls.append("close")
        self.closed = True
+
+
+class ChaosGym(FakeGym):
+    """@brief A gym whose login, click and bookings read each fail with a set probability.
+
+    The site's network simulation fails up to 50% of these requests and writes nothing when
+    one fails. The random numbers come from a seeded generator, so a run is repeatable.
+    """
+
+    _FAILING_CALLS = frozenset({"log_in", "press_class_button", "list_my_bookings"})
+
+    def __init__(
+        self,
+        rng: random.Random,
+        failure_probability: float = 0.5,
+        cards: list[ClassCard] | None = None,
+    ) -> None:
+        super().__init__(cards=cards)
+        self._rng = rng
+        self._failure_probability = failure_probability
+
+    def _should_fail(self, name: str) -> bool:
+        return name in self._FAILING_CALLS and self._rng.random() < self._failure_probability
+
+
+def no_sleep(seconds: float) -> None:
+    """@brief A `sleep` that returns at once, so a retry test never waits."""
@@ -3,13 +3,13 @@
 """

 import pytest
-from tests.unit.fakes import FakeGym
+from tests.unit.fakes import FakeGym, no_sleep
 from tests.unit.test_booking_entry import make_entry
 from tests.unit.test_models import make_card

 from snack_and_lift.application.booking import book_target_classes
 from snack_and_lift.application.verification import get_my_bookings, verify_bookings
-from snack_and_lift.domain.errors import GymError, NoBookingsFoundError
+from snack_and_lift.domain.errors import GymError, NoBookingsFoundError, RetriesExhaustedError
 from snack_and_lift.domain.models import ClassResult, Outcome


@@ -53,6 +53,16 @@ def test_verify_bookings_ignores_entries_that_are_not_target_classes() -> None:
    assert verify_bookings(gym, results).is_success


+def test_verify_bookings_lists_target_entries_after_the_window_as_also_booked() -> None:
+    gym = FakeGym(bookings=[make_entry(), make_entry("Thu, Aug 21, 6:00 PM")])
+    results = [ClassResult(make_card("Tue, Aug 12"), Outcome.NEW_BOOKING)]
+
+    verification = verify_bookings(gym, results)
+
+    assert verification.is_success
+    assert verification.also_booked == ["Spin Class on Thu, Aug 21"]
+
+
 def test_verify_bookings_reports_a_missing_booking() -> None:
    gym = FakeGym(bookings=[make_entry("Wed, Aug 13, 7:00 AM")])
    results = [ClassResult(make_card("Tue, Aug 12"), Outcome.NEW_BOOKING)]
@@ -67,8 +77,10 @@ def test_verify_bookings_raises_when_classes_were_booked_but_the_page_is_empty()
    gym = FakeGym(bookings=[])
    results = [ClassResult(make_card("Tue, Aug 12"), Outcome.NEW_BOOKING)]

-    with pytest.raises(NoBookingsFoundError):
-        verify_bookings(gym, results)
+    with pytest.raises(RetriesExhaustedError) as caught:
+        verify_bookings(gym, results, sleep=no_sleep)
+
+    assert isinstance(caught.value.__cause__, NoBookingsFoundError)


 def test_verify_bookings_skips_the_page_when_nothing_is_expected() -> None:
@@ -5,7 +5,7 @@
 import logging

 import pytest
-from tests.unit.fakes import FakeGym
+from tests.unit.fakes import FakeGym, no_sleep
 from tests.unit.test_models import make_card

 from snack_and_lift.application.booking import (
@@ -147,4 +147,4 @@ def test_book_target_classes_propagates_a_failing_class() -> None:
    gym = FakeGym(cards=[make_card("Thu, Aug 14")], press_error="Network down")

    with pytest.raises(GymError, match="Network down"):
-        book_target_classes(gym)
+        book_target_classes(gym, sleep=no_sleep)
--- a/Show More
+++ b/Show More
Author	SHA1	Message	Date
Tirsvad	0f4c6989a4	Resolve pending commit links	2026-10-29 00:11:16 +08:00
Tirsvad	2cfcd02588	Update the framework to 150819a, follow the new milestone workflow template, review RC-027	2026-10-29 00:11:15 +08:00
Tirsvad	332d65255c	Merge pull request 'MIL-010: gym accounts from a credentials file' (#59 ) from mil-010-credentials-file into main Publish accepted milestones / publish (push) Failing after 4s Details TirSystem/git-workflows: Publish wiki / publish-wiki (push) Successful in 16s Details TirSystem/git-workflows: Sync GitHub mirror metadata / sync-metadata (push) Successful in 6s Details Publish wiki / publish-wiki (push) Successful in 14s Details Publish wiki / wiki (push) Successful in 14s Details	2026-10-28 15:26:25 +01:00
Tirsvad	da3f999c6f	Resolve pending commit links	2026-10-28 22:26:18 +08:00
Tirsvad	42f05bdd0f	Load the gym accounts from the .env.gym file Task: MIL-010#1 Task: MIL-010#2 Task: MIL-010#3	2026-10-28 22:26:03 +08:00
Tirsvad	434bdafefd	Plan MIL-010: gym accounts from a credentials file	2026-10-28 22:22:00 +08:00
Tirsvad	a248541e58	Merge pull request 'MIL-009: screenshots after the fix' (#55 ) from mil-009-screenshots into main TirSystem/git-workflows: Publish wiki / publish-wiki (push) Successful in 12s Details TirSystem/git-workflows: Sync GitHub mirror metadata / sync-metadata (push) Successful in 7s Details Publish wiki / publish-wiki (push) Successful in 10s Details Publish wiki / wiki (push) Successful in 11s Details	2026-10-28 14:48:45 +01:00
Tirsvad	034f792c1a	Add step 8 screenshot after the verification fix (MIL-009)	2026-10-28 21:48:36 +08:00
Tirsvad	fedbc39165	Merge pull request 'MIL-009: verification fix' (#54 ) from mil-009-verification-fix into main Publish accepted milestones / publish (push) Failing after 4s Details Publish wiki / publish-wiki (push) Successful in 15s Details Publish wiki / wiki (push) Successful in 15s Details TirSystem/git-workflows: Publish wiki / publish-wiki (push) Successful in 11s Details TirSystem/git-workflows: Sync GitHub mirror metadata / sync-metadata (push) Successful in 5s Details	2026-10-28 14:46:15 +01:00
Tirsvad	85f5eecd07	Resolve pending commit links	2026-10-28 21:46:06 +08:00
Tirsvad	52427ca69c	Add review record RC-024 and update the step 8 notes (MIL-009)	2026-10-28 21:46:04 +08:00
Tirsvad	353d157e22	Compare only the days of the run in the verification Task: MIL-009#1 Task: MIL-009#2	2026-10-28 21:45:48 +08:00
Tirsvad	3b9e381cd4	Plan MIL-009: verification fix (issue #50 )	2026-10-28 21:39:22 +08:00
Tirsvad	ca6a706abe	Merge pull request 'MIL-008: real screenshots' (#49 ) from mil-008-real-screenshots into main TirSystem/git-workflows: Sync GitHub mirror metadata / sync-metadata (push) Successful in 7s Details TirSystem/git-workflows: Publish wiki / publish-wiki (push) Successful in 36s Details Publish wiki / publish-wiki (push) Successful in 12s Details Publish wiki / wiki (push) Successful in 13s Details	2026-10-28 14:34:50 +01:00
Tirsvad	bd28943482	Add real screenshots of steps 1 to 9 (MIL-008)	2026-10-28 21:34:38 +08:00
Tirsvad	2c568f1fa2	Merge pull request 'MIL-008: add step images' (#48 ) from mil-008-screenshots into main TirSystem/git-workflows: Publish wiki / publish-wiki (push) Successful in 24s Details TirSystem/git-workflows: Sync GitHub mirror metadata / sync-metadata (push) Successful in 5s Details Publish wiki / publish-wiki (push) Successful in 15s Details Publish wiki / wiki (push) Successful in 15s Details	2026-10-28 12:04:24 +01:00
Tirsvad	cbdd553394	Add step 2 to 9 illustrations and label them (MIL-008)	2026-10-28 19:04:15 +08:00
Tirsvad	7a8412ad6b	Merge pull request 'MIL-008: documentation, wiki and release' (#47 ) from mil-008-documentation into main Publish wiki / publish-wiki (push) Successful in 36s Details TirSystem/git-workflows: Publish wiki / publish-wiki (push) Successful in 11s Details TirSystem/git-workflows: Sync GitHub mirror metadata / sync-metadata (push) Successful in 6s Details Publish accepted milestones / publish (push) Failing after 6s Details Publish wiki / wiki (push) Successful in 36s Details	2026-10-28 11:20:58 +01:00
Tirsvad	08ced00f62	Resolve pending commit links	2026-10-28 18:20:49 +08:00
Tirsvad	0301fd2f0b	Add wiki in three languages, README screenshot section and wiki workflows (MIL-008)	2026-10-28 18:20:47 +08:00
Tirsvad	c0aa0b4b2b	Merge pull request 'MIL-007: resilience against network failures' (#46 ) from mil-007-resilience into main TirSystem/git-workflows: Sync GitHub mirror metadata / sync-metadata (push) Successful in 6s Details	2026-10-28 11:14:26 +01:00
Tirsvad	6871e54333	Resolve pending commit links	2026-10-28 18:14:19 +08:00
Tirsvad	5d28e682b4	Add review record RC-021 for MIL-007	2026-10-28 18:14:18 +08:00
Tirsvad	37f3cc436a	Add retry wrapper and resilient login, booking and bookings read (step 9) retry(func, retries=7, description=None), RetriesExhaustedError, resilient login, book_class and get_my_bookings that re-read the card before a repeat and count a booking only when the page confirms it, --retries option, step-9 console image, seeded chaos tests at 50 percent failure and container tests against failing local fixtures. Task: MIL-007#1 Task: MIL-007#2 Task: MIL-007#3 Task: MIL-007#4	2026-10-28 18:14:13 +08:00
Tirsvad	a830e3ae5f	Merge pull request 'MIL-006: verify bookings and time travel QA' (#45 ) from mil-006-verify-and-time-travel into main TirSystem/git-workflows: Sync GitHub mirror metadata / sync-metadata (push) Successful in 7s Details	2026-10-28 11:03:00 +01:00
				`@@ -0,0 +1 @@`
				`[Repository](https://git.tirsystem.com/Tirsvad-Udemy-100-days-of-code/049-snack-and-lift) · [Snack and Lift](https://appbrewery.github.io/gym/) · AGPL-3.0`