Skip to content

docs: correct my own misdiagnosis — main was never broken (CDATA was innocent) #2

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
GothwadTech merged 1 commit into from
Oct 15, 2026
Merged

docs: correct my own misdiagnosis — main was never broken (CDATA was innocent) #2

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 11 additions & 1 deletion docs/AUDIT-2026-10-14.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -295,5 +295,15 @@ Commits: `086370d` (P0/P1 pass) → `bb3b832` (batches 1–3: app lock, DiffUtil
| W10 | 🟡 partial | `scripts/quality-gate.sh` (13 mechanical checks, all passing) + `.github/workflows/quality.yml` (gate, unit tests, compile, lint, APK build path untouched); first JVM unit tests: `ConfigRulesTest` pins the live-tab cap, input-mode resolution, zoom clamping, desktop-UA detection and privacy defaults | Robolectric tests for `PageZoomController`/`ShortcutMgr`/`TabsModel`/`PageTranslate`; APK-size-per-ABI recording in `release.yml` |
| W11 | 🟡 blocked on owner | README About-screen URL/e-mail corrected to the real remote, no e-mail shown in-app | The licence contradiction in §0 is a legal decision, not a code change: either publish under GPL-3.0-or-later with TV Bro attribution, or re-derive the affected modules |

**Verification status (updated after the first real CI run).** The code was authored in an environment with no JDK and no Android SDK, so it went through CI before it went anywhere else: `.github/workflows/quality.yml` now reports **green** — `:app:compileDebugKotlin` + `:app:common:compileDebugKotlin`, `:app:common:testDebugUnitTest` (the 11 new `ConfigRulesTest` cases) and `:app:lintDebug` all pass. The compile earned its keep: it caught 6 defects the static checks could not see — `ActivityManager.MemoryInfo` vs `isLowRamDevice()`, three string resources with an unescaped apostrophe, one with a Kotlin `\u2026` escape, a `return@withContext` label left behind by a moved guard, two missing extension imports, and two JVM property/setter signature clashes. **Two of them were already on `main`** (`<![CDATA[…]]>` in `version_and_updates`, both the default and `zh-rTW` — `:app:mergeDebugResources` fails on it), which is what a build gate on every PR is for.
**Verification status (updated after the first real CI run).** The code was authored in an environment with no JDK and no Android SDK, so it went through CI before it went anywhere else: `.github/workflows/quality.yml` reports **green** on `main` — `:app:compileDebugKotlin` + `:app:common:compileDebugKotlin`, `:app:common:testDebugUnitTest` (the 11 new `ConfigRulesTest` cases) and `:app:lintDebug`. The compile earned its keep: it caught six defects the static checks could not see — `ActivityManager.isLowRamDevice()` (not `MemoryInfo.isLowRamDevice`), three string resources with an unescaped apostrophe, a `return@withContext` label left behind by a moved guard, two missing extension imports, and two JVM property/setter signature clashes.

> **One of my own diagnoses there was wrong and is recorded so it does not get repeated.** The first
> resource failure (`Can not extract resource from ParsedResource` ×3) I blamed on the
> `<![CDATA[Version & updates]]>` strings and wrote that `main` had not built since `3c3f210`. It
> had: `build_apk.yml` completed successfully for `3c3f210` *and* for the merge of PR #1. The three
> unextractable resources were the three apostrophe strings (the count matched all along); aapt2
> accepts both CDATA and `\uXXXX`. The CDATA→`&amp;` edit is therefore cosmetic (identical rendered
> text), and the gate keeps a CDATA check for the one *real* trap: inside CDATA, `\n`/`\'` are taken
> literally, so escapes silently stop working.

Still owed before this is called production-ready: a real phone (portrait 360dp) + real TV-box run of the §6 checklist in `AGENTS.md`, and an APK-size measurement to back the README number.
7 changes: 5 additions & 2 deletions scripts/quality-gate.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -123,8 +123,11 @@ EOF
)
report "string resources escape apostrophes and avoid unicode escapes" "$esc"

# CDATA inside a <string> is a hard aapt2 failure ("Can not extract resource from ParsedResource"),
# so escaped entities are the only legal way to carry & or < in a resource.
# Prefer entities over CDATA in value resources: aapt2 accepts CDATA, but it also disables aapt's
# escape handling inside it (a \n or \' in a CDATA block reaches the user literally), which is a
# silent-trap in a file where the other 660 strings rely on escapes. Found while chasing a real
# "Can not extract resource" failure — which turned out to be unescaped apostrophes, not CDATA.
# (Yes, the first version of this comment blamed the wrong thing; keep the diagnosis honest.)
cdata=$(grep -rn 'CDATA' app/src/main/res/values* app/common/src/main/res/values* 2>/dev/null || true)
report "no CDATA in resource values" "$cdata"

Expand Down
Loading