STATUS.md (6303B)
1 <!-- 2 This file is part of GNU Taler 3 (C) 2026 Taler Systems S.A. 4 5 GNU Taler is free software; you can redistribute it and/or modify it under the 6 terms of the GNU General Public License as published by the Free Software 7 Foundation; either version 3, or (at your option) any later version. 8 9 GNU Taler is distributed in the hope that it will be useful, but WITHOUT ANY 10 WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR 11 A PARTICULAR PURPOSE. See the GNU General Public License for more details. 12 13 You should have received a copy of the GNU General Public License along with 14 GNU Taler; see the file COPYING. If not, see <http://www.gnu.org/licenses/> 15 --> 16 17 # Where the findings stand 18 19 Checked on 2026-11-08 against `origin/master`, after rebasing this branch onto 20 it. **All twelve are fixed.** Nothing from rounds one to three is still open, 21 so nothing was filed on `bugs.gnunet.org`. 22 23 Every fix arrived in one squashed commit, `24f7b249 [pos] add support for mixed 24 payments and UI changes`, and eleven of the twelve are now pinned by a unit 25 test that names the exact case the simulator found. Those tests pass on this 26 branch: `:merchant-terminal:testSimUnitTest`, 27 `:taler-kotlin-android:testDebugUnitTest` and `:merchant-lib:test` are all 28 green. 29 30 | | Finding | How it is fixed | Pinned by | 31 | --- | --- | --- | --- | 32 | 1 | [QR tab crashes on the first camera frame](qr-config-tab-crashes-the-terminal.md) | `config/QrLuminance.kt` is new: it honours `rowStride`/`pixelStride` and rotates by writing to the destination index, so nothing calls `PlanarYUVLuminanceSource.rotateCounterClockwise` any more | `QrLuminanceTest` -- padded interleaved frames at all four rotations, plus a truncated-plane case | 33 | 2 | [A refund of `1.000.50` settles at 1](refund-amount-swallows-extra-decimal-points.md) | `Amount.fromString` now throws on a second separator: `if (valueSplit.size > 2) throw AmountParserException(...)`, and `isValidAmountStr` agrees | `AmountTest.rejectsExtraDecimalSeparators`, whose list begins `"1.0.0", "1.000.50"` | 34 | 3 | [A fresh instance cannot be used](fresh-instance-cannot-be-used.md) | `OrderManager.applyConfiguration` synthesises the "all products" and "uncategorized" categories and returns `null` unconditionally; an empty grid says `order_empty_inventory` instead of refusing the configuration | `ConfigRequestsTest.empty inventory loads and persists through the real unescaped endpoint` | 35 | 4 | [An expired session shows a protocol error](expired-session-shows-a-protocol-error.md) | `ConfigManager.handleUnauthorized` fires `sessionExpired` only for the active merchant's own private requests; `MainActivity` observes it and shows `session_expired_toast` | `ConfigSessionTest`, three cases including `real session rejection clears credentials and is consumed once` | 36 | 5 | [A silent backend hangs the login forever](login-hangs-forever-on-a-silent-backend.md) | the Ktor client now sets `requestTimeoutMillis`/`socketTimeoutMillis` to 30 s and `connectTimeoutMillis` to 15 s | `ConfigRequestsTest.unresponsive configuration times out with a network error and can be retried` | 37 | 6 | [The terminal asks for `/private%2Fpos`](pos-config-url-percent-encodes-the-separator.md) | the URL is built one segment at a time (`.appendPath("private")`), so the separator is not encoded | `ConfigRequestsTest` asserts the server saw `"/instances/test/private/pos"` | 38 | 7 | [A setup URI loses the backend's port](setup-uri-drops-the-port.md) | `merchantSetupConfig` keeps it: `val authority = if (data.port == -1) host else "$host:${data.port}"` | `MerchantSetupUriTest.setup links preserve custom ports and both credential formats` | 39 | 8 | [A failed config fetch strands the spinner](config-fetch-failure-strands-the-spinner.md) | the spinner is its own `ConfigFetcherFragment` with a Cancel button and an error branch | `ConfigRequestsTest.configuration cancellation stops pending requests...` | 40 | 9 | [Controls a screen reader cannot announce](controls-without-accessible-names.md) | all three now have names: the drawer button `navigation_menu_open`, the password toggle `password_show`/`password_hide`, and the checkbox row carries `Role.Checkbox` with `onCheckedChange = null` on the box itself | source only | 41 | 10 | [A cut connection is reported in OkHttp's words](transport-error-shows-the-library-wording.md) | `is IOException -> NETWORK_ERROR_MESSAGE`, with no `exception.message` fallback | `ResponseCancellationTest` asserts the message is "Network error: check your internet connection and merchant URL." | 42 | 11 | [A missing field is reported by its Kotlin type name](serialization-error-shown-to-the-cashier.md) | `is SerializationException, is ContentConvertException -> "The merchant server returned an invalid response. Please try again."`, and the `else` arm no longer quotes the exception either | `ResponseCancellationTest` feeds it `SerializationException("Field amount is required for type net.taler.internal.Type")` | 43 | 12 | [A form loses its buttons to the keyboard on a phone](refund-form-unusable-on-a-phone.md) | `RefundFragment` scrolls: `.imePadding().verticalScroll(rememberScrollState())` | source only | 44 45 ## The tracker 46 47 `bugs.gnunet.org` has one open issue in `merchant-pos-terminal (Android App)` -- 48 11908, "POS apps lacks persistant db inside of the app" -- which is not any of 49 these. The three others in that category are closed: 10210 ("Session expired 50 closes application", closed *fixed*, and the closest neighbour to finding 4), 51 8652 and 9207. 52 53 One caution for whoever searches next: `status=open` is accepted by 54 `GET /api/ai/issues` but does **not** filter -- the three closed issues above 55 came back from a query asking for open ones. Check each issue's own `status`. 56 57 ## What this does not say 58 59 The twelve were checked by reading `origin/master` and by running its unit 60 tests, not by driving the app. The simulator's own environment was gone when 61 this was written -- `tripwire` had rebooted and `/tmp/taler-pos-sim` with it -- 62 so there is no fresh emulator run behind this table. For the eleven with a 63 regression test naming the exact input, that is not much of a gap. For 9 and 12, 64 which are layout and accessibility properties that a unit test cannot hold, 65 a round four should confirm them on a device.