iobroker.alias-workbench 0.9.8 → 0.9.10

package/LICENSE

@@ -1,6 +1,6 @@
 MIT License
 
-Copyright (c) 2026 Ricardo Hipp
+Copyright (c) 2026 Ricardo Hipp <ricardo.hipp@googlemail.com>
 
 Permission is hereby granted, free of charge, to any person obtaining a copy
 of this software and associated documentation files (the "Software"), to deal

package/README.md

@@ -529,8 +529,10 @@ from the library was the main weakness of the existing alias manager.
 
 ## Development
 
+After cloning, fetch the dependencies the usual way for a Node project, then
+run the build script:
+
 ```bash
-npm install
 npm run build      # bundles @iobroker/type-detector into admin/detector.js
 ```
 
@@ -562,9 +564,32 @@ imported, and that both language files carry the same keys.
 
 ## Changelog
 
+### 0.9.10 (2026-09-13)
+* Four things from the full test run. **A channel identifier** was not sanitised in the multi-output path: from the channel name “Licht, Bar” the js-controller quietly made `Licht_ Bar`, while the workbench displayed “Licht, Bar” and wrote it into the enumerations — an identifier that never existed, and after deleting the alias it stayed there as a leftover. `ausgangName` now sanitises itself, and where two channels produce the same name the second gets a `_2`. **The caption** differed between the two views: where the point's name equals the row name, “Edit alias” showed it and “Create alias” an empty field. **The row list** now counts like the tree — `POWER2` before `POWER10`. And **an enumeration the workbench created itself** now also disappears when it becomes empty through a change of function, not only when the alias is deleted
+* Tidying: "turn an alias object into a draft row" was written out five times — each with the same write-source rule and slightly different surroundings. Four bugs in this version came from exactly that: the lost write source on the roller shutter, a switching point that became a read-only one, the missing caption on `alias.0.Solar.Netz`, and a deviation the workbench had produced itself. The reading now lives in one place; what the four callers do with it — compare, create, overwrite — stays with them. Behaviour is unchanged
+* Three small things spotted during the run. **RSSI** was pre-ticked in the five Tasmota and measuring-point templates but not in the eleven Homematic ones — the same device property, two answers depending on the vendor; it is now unticked everywhere, like the rest of the diagnostics. **The `MAP_URL` slot** in the vacuum pattern got no role because its expression starts without `^` and the derivation demanded both anchors; a missing start anchor no longer counts against a slot. **And the "guessed" tag** in the swap dialog looked like text stuck to the point name — its styling hung on three parent levels that do not exist there. It is now a pill like in the state list and says in its tooltip whether the guess came from the point name or from the role
+* The save button in the template dialog now also checks when pressed, not only when drawn. It is re-evaluated when a field is left — typing an open bracket into the name hint and clicking "Save template" without leaving the field left you with a blue button and saved a template with a broken expression. The five lock reasons now live in one place that drawing and saving share
+* Three places now say why they cannot. **The save button in the template dialog** was greyed out and silent — five reasons can lock it and none of them was shown; every applicable one now appears on the button and as a box in the dialog. **The error dialog when deleting** only gave a count ("1 operations failed") and not which id is stuck; writing has had that list all along. **And the reason a send point could not be created** hung on a button that the same callback threw away 120 ms later — it now sits at the top of the dry run and survives the rebuild
+* If a point's type does not match its pattern slot, all three places now say so — not just the expanded row. If the slot was **required**, the pattern stops applying altogether, and the whole list showed up tinted: on the Bastelzimmer_Licht, switching SET from `boolean` to `mixed` was enough for eight of eight rows to count as "no slot", while the pattern field reported "✕ SET missing" — SET was right there. It now reads "✕ SET: type mixed instead of boolean", and above the list: "The light pattern does not apply right now: SET requires boolean, but mixed is set. That is why no row has a slot." Only the row that causes it is tinted now — the whole list used to light up, which showed the consequence instead of the cause.
+* A late answer to `system.config` now really does carry the expert mode over. The tab gives up after 2.5 seconds and continues with the browser language; if the answer arrives later, the language stays as it is (switching after the labels are drawn would be worse), the expert mode is adopted and the tree is redrawn once. That had long been built — but above it sat a guard from the time before, which bailed out in exactly this case. Without the mode, every node under `system.*` and `enum.*` was missing, with nothing saying why
+* The swap dialog notices when one level up fits better. The picker lets you choose any node with data points underneath — it has to, because on a flat adapter the channel is the device. With MQTT, though, `stat`, `tele` and `cmnd` are topic folders: landing there finds fewer points, and the lock would ask you to remove some. The workbench now also computes the parent level and says so — "One level up (…) finds a target for 7 of 7 points, here only 5" — with a button that switches the selection. Nothing is forbidden
+* The device list in the swap dialog keeps its position. Scrolling down a long list and picking a device used to jump back to the top, with the freshly picked entry out of sight — picking rebuilds the dialog, and the tree is a different element afterwards. If the selected row still ends up outside, it is brought into view. A new search and a freshly opened dialog still start at the top
+* When swapping the source, a write target no longer counts if it coincides with the read source. The search tries the bare point name as its last form, so `cmnd.POWER` found a status point that also ends in `POWER` — the alias would then have written to its own read source. Guessing already caught this, matching by path did not; it showed when the new source was picked as a sub-node instead of the device
+* When swapping the source, a point that cannot come along now stops the swap instead of letting it fail halfway. A point with separate sources — reads from `stat.POWER`, writes to `cmnd.POWER` — needs both at the new source. If the write target is missing, ioBroker has **no** form that means "reads, writes nowhere": a plain id means "write to the status point", `{read: X}` without `write` is rejected by the js-controller, and `common.write: false` does not protect the source (measured: writing to such an alias overwrote the read source). So the user decides — the checkbox "remove this point when swapping" now appears in this case as well, and "Swap now" stays disabled until someone has decided; the reason shows on the button and as a line in the dialog. Before, the swap started, failed at the controller and left the alias half on the old and half on the new source
+* If a write fails during the swap, **nothing** is deleted. The deletions used to run anyway, so an alias that was already half migrated lost points on top of it. Moving an alias has followed that rule for a long time
+* "All data points of the device" no longer moves the storage location. For a device with several outputs the workbench creates a folder with one channel per output. The button rebuilds the draft and passed on only the **finished target**; `setzeZiel` reads its last segment as the device name — with several outputs that segment is the output, and it ended up in the folder as well. Every press pushed the location one level deeper, and that is where it was written, too: three clicks produced four folders instead of one (`…Garten-Ventilinsel.POWER2.POWER2.POWER2.POWER2`). The button now passes folder and name separately, the way the output selector next to it has always done. A single-output device was never affected
+
+### 0.9.9
+* The build source moved from `src/` to `src-admin/`. ioBroker reserves that name for frontend sources that are bundled into `admin/` and never reach the user, and the adapter checker skips such directories when it looks for missing dependencies. `@iobroker/type-detector` therefore stays in `devDependencies`, which is where it belongs: esbuild bakes it into `admin/detector.js` at build time and nothing loads it at runtime. Users receive the finished 71 KB file, not the 548 KB package (W5042)
+* The settings page now takes its translations from `admin/i18n/` — the dictionary ioBroker reads itself (`"i18n": true`) and that Weblate can work with. The eleven languages it already carried moved there unchanged; not a word was retranslated. The tab keeps its own dictionary, now at `admin/sprachen/`: 588 keys in German and English, with an unknown language falling back to English. Two dictionaries, two jobs — the name `i18n` is the one the adapter checker inspects, and it expects all eleven languages there (W5603, W5614, W5015)
+* The changelog entries for 0.9.7 and 0.9.8 are now complete in all eleven languages. Nine of them carried only a short summary — the English text for 0.9.7 ran to 3,000 characters with five headings and 22 items, the Russian one to 480 without any structure. A test now checks every entry for length, headings and item count against the English original (W1145)
+* Two developer-side additions, invisible in the adapter itself: `.vscode/settings.json` wires up the JSON schemas, so mistakes in `io-package.json` or `jsonConfig.json` show up while typing instead of in the adapter checker (S4036). And the GitHub Actions updates no longer run `monthly` — that would ask on the first of every month, at the same moment as everyone else; a cron schedule spreads the load while keeping the monthly rhythm (S8906)
+* The data type now follows the role. A pattern slot prescribes not only the role but also the type, and both sit in the same type-detector record — so picking a role sets the matching type as well. Where a slot allows several types nothing is set, and where it prescribes none any choice stays right. If the type does not fit the slot after all, the workbench says so instead of staying quiet: in the row ("expects number"), as a warning underneath, and in the pattern field, which now gives the reason — "✕ CONSUMPTION: type mixed instead of number" — rather than a bare "does not fit". Before, a state with the right role and the wrong type dropped out of the pattern silently, and the whole device was recognised as something lesser Where a role matches several slots that differ in type — the blind's travel direction sits on DIRECTION as a boolean and on DIRECTION_ENUM as a number — the workbench picks the slot whose type fits, and both forms count as right.
+* Saving a template no longer loses the write targets. Every actuator that keeps its own state writes to the same point it reads - a blind, a dimmer and a thermostat all use 1.LEVEL for both. Saving dropped the write path in exactly that case, and the resulting alias showed the position but could no longer be moved. Rows that only write, like the blind's STOP, disappeared altogether. With Tasmota it never showed, because there tele.STATE is read and cmnd.POWER is written. The device type in the template sheet was affected by a second slip: three patterns are named differently from their type (blinds/blind, mediaPlayer/media, levelSlider/slider), so a blind always claimed to have no type at all
+
 ### 0.9.8
 * The changelog only lists versions that actually exist on npm. The entries 0.9.1 to 0.9.6 were local intermediate steps and were never published; their content is now summarised under 0.9.7. Found by the repochecker (E2004, E1032)
-* The README no longer explains installation via “install from custom URL” — in the official repository that goes through the adapter list (E6013)
+* The README no longer describes how to add the adapter by hand — in the official repository it is picked from the adapter list (E6013)
 * `dependabot.yml` gets a seven-day cooldown for npm packages: what appears today is proposed a week later. If a compromised release is pulled within that time, it never arrives here (E8915). And `@types/node` stays on its major version — the types must match the Node release the adapter runs on (E8917)
 
 ### 0.9.7

package/admin/css/werkbank.css

@@ -553,9 +553,14 @@
   .erow .nm3 .geae { font-family:var(--sans); font-size:8px; font-weight:700; letter-spacing:.05em;
                      text-transform:uppercase; background:var(--warn-soft); color:var(--warn);
                      padding:0 6px; border-radius:999px; margin-left:6px; }
-  .erow .nm3 .rat { font-family:var(--sans); font-size:8px; font-weight:700; letter-spacing:.05em;
-                    text-transform:uppercase; background:var(--accent-soft); color:var(--accent);
-                    padding:0 6px; border-radius:999px; margin-left:6px; }
+  /* Die Marke gilt ueberall gleich, nicht nur in der Zustandsliste.
+     Mit den drei Ebenen davor griff sie im Tausch-Dialog nicht: dort
+     haengt das span unter .sn, und "vermutet" stand als blosser Text
+     am Punktnamen - ohne Abstand, ohne Farbe, in 12 statt 8 px
+     (Ricardo, 12.09.2026, an alias.0.Garten.FK_Badezimmer). */
+  .rat { font-family:var(--sans); font-size:8px; font-weight:700; letter-spacing:.05em;
+         text-transform:uppercase; background:var(--accent-soft); color:var(--accent);
+         padding:0 6px; border-radius:999px; margin-left:6px; white-space:nowrap; }
   .erow .nm3 .ratknopf { font-family:var(--sans); font-size:11px; cursor:pointer; margin-left:5px;
                          color:var(--ink-3); padding:0 4px; border-radius:var(--r-sm); }
   .erow .nm3 .ratknopf:hover { background:var(--surface-3); color:var(--ink); }