CRT/storebound
1
0
Fork 0
Code Issues Pull requests Releases 1 Packages 1 Activity Actions

--json prints only JSON on stdout (withJsonOutput around every command: logs, tables and prompts to stderr; printJson everywhere; test runs every --json command against the fake stores and parses stdout); prices ppp compares like with like: par = the store's own US-equivalent price (Play convertRegionPrices, Apple equalizations; offline base x fx x (1 + VAT) rounded up) scaled by the PPP price level; defaults only lower, only price levels <= 0.5, World Bank high-income at par, CN at par on Apple and skipped on Play, never above par (--only-decreases, --max-ratio, --min-ratio, --max-level, --include-high-income; --floor / --ceiling kept), top-10 revenue market drops > 20 % warned (sales reports, else a static list), unchanged regions out of the sheet; price changes of existing products only with --allow-price-changes, TTY or not (a terminal yes covered them), then their own question listing each old -> new, held section in dry runs (play products, prices import, play / apple subscriptions); pitfalls 61-63; tests

Browse source
This commit is contained in:
CRTkafa 2026-10-14 17:36:00 +08:00
parent bab3f63227
commit 357319a775
30 changed files with 865 additions and 209 deletions
Download patch file Download diff file Expand all files Collapse all files

3
docs/pitfalls.md
View file

@ -9,6 +9,9 @@ fails without the fix.
| # | date | lesson | guard |
|---|---|---|---|
| 63 | 2026-10 | STAR DASH: re-running `play products` / `prices import` to create 8 new IAPs also changed the live prices of existing products; `--only` was needed. In a terminal (no `--yes`) the one confirmation covered price changes too. | Price changes of existing products need `--allow-price-changes`, TTY or not; the terminal question covers creates and other updates only; with the flag a second question lists each old -> new (asked in a terminal even with `--yes`); dry runs show a "price changes on existing products (held)" section (`allowPriceChanges`, `heldSection` in `cli/lib/price-gate.ts`; `play products`, `prices import`, `play subscriptions`, `apple subscriptions`). `tests/price-gate.test.ts` (TTY and non-TTY). |
| 62 | 2026-10 | STAR DASH: `prices ppp` cut high-income markets (Germany EUR 3.59 instead of 5.99 for USD 4.99; UK, France, Japan cut too): it set base x PPP level x exchange rate, a net price, against VAT-inclusive store prices, and cut every region. | Par = the store's own conversion of the US price (Play convertRegionPrices, Apple equalizations; offline base x fx x (1 + VAT) rounded up); share = clamp(PPP level, 0.3, 1); defaults: only decreases, PPP only where the level is <= 0.5, World Bank high-income economies and CN at par, Play skips CN, top-10 revenue market drops > 20 % warned (from `.storebound/reports/` sales reports, else a static list); unchanged regions are left out of the sheet (`shareOf`, `offlinePar`, `planRows`, `topMarketDrops`, `topMarketsFromReports` in `cli/lib/ppp.ts`). `tests/ppp.test.ts`. |
| 61 | 2026-10 | STAR DASH: `prices ppp --json` printed log lines on stdout before the JSON, so parsing stdout failed. | `withJsonOutput` (`cli/lib/log.ts`) wraps every `--json` run in `cli/index.ts`: console.log / info go to stderr, `printJson` writes stdout, prompts use `promptOut()`; commands that printed JSON with console.log use `printJson`. `tests/json-stdout.test.ts` runs every command that reads `--json` against the fake stores and parses stdout. |
| 55 | 2026-10 | The Play feature graphic (1024 x 500) cut the phone off at the bottom: the game template's phone was 232 px wide at top 36 and rotated 7 deg, so its outline spanned about y 24..552 and the game's bottom bar was lost; the app template bled too, and `shots.feature` only took frame + palette. Nothing measured the device against the canvas. | `featureLayout` / `rotatedBox` (`cli/shots/lib.ts`) size the device so its rotated box fits the canvas with a 24 px margin (`shots.feature.fit` 'contain', the default; 'bleed' keeps the crop on purpose; `scale` 0.5-1.5), both templates use it. The templates record each device's measured layout box (`devices` in report.json) and `shots check` fails a feature graphic whose device leaves the canvas unless fit is 'bleed', and warns for a screenshot's front device off a side or the top (`deviceIssues`, `cli/shots/report.ts`). `tests/feature-fit.test.ts`. |
| 54 | 2026-10 | The second `apple build` of a Capacitor game archived fine, then `xcodebuild -exportArchive` failed with "exportArchive Copy failed" (exit 70). The IDEDistribution log had the real cause: `rsync error: syntax or usage error (code 1)` in IDEDistributionCreateIPAStep, because a Homebrew rsync 3.x, installed and linked that day by another session's release tool, came first on PATH. A tool changed a global binary on a shared machine as a side effect. | `apple build` runs xcodebuild with Xcode's and Apple's tool folders first on PATH (`xcodePath`) and prints a note when a non-Apple rsync is first (`foreignRsync`); an export failure shows the error lines of the newest IDEDistribution*.log (`distributionErrors`, `cli/lib/xcode.ts`, `cli/apple/build.ts`). `storebound doctor` flags a non-Apple rsync first on PATH. `tools/desktop-release.ts publish` calls rsync by absolute path (`/usr/bin/rsync` without `--chmod`, modes set on the server; else the Homebrew keg), never from PATH (`publishSteps`). Rule: tools never install or relink global binaries as a side effect. `tests/apple-build-export.test.ts`. |
| 53 | 2026-10 | After that failed export, `apple upload build/ios/export/App.ipa --yes` uploaded nothing: it read the stale IPA of build 1, found build 1 already VALID on App Store Connect and exited 0 ("build 1 is VALID"), so `build && upload` looked green. The export also wrote into the same folder every time. | `apple build` exports each build into its own folder `build/ios/export/<version>-<build>/` (removed first) and exits non-zero on any export failure or a missing IPA (`exportIpa`). `apple upload` fails with exit 1 when the IPA's build number is already on App Store Connect ("this ipa is build 1, which is already on App Store Connect (VALID): did the last build fail?"); only `--allow-existing` re-checks it. `--build N` picks / requires that build's IPA; the release and beta flows use `apple upload --no-wait --build ${BUILD} --allow-existing`, so a resumed flow still passes and a stale IPA never does (`existingBuild`, `exportedIpas`, `cli/apple/upload.ts`; `tests/apple-build-export.test.ts`). |