Rules list incomplete in 5.2.0.M5 – only 11 of 64 JS file rules shown

[miloit]

The problem

After updating from milestone M4 to M5, the Rules tile / Rules list no longer shows
all configured rules. The Rules overview displays a count of 11, while the same
installation showed 64 rules under M4.

It looks like not all rules are being loaded (or not rendered) in the Main UI under M5.

Expected behavior

All configured rules (64) should be listed and counted in the Rules tile, as was the
case in M4.

Steps to reproduce

1. Open the Main UI → Settings → Rules
2. Observe the rule count shown on the tile / the number of rules in the list

Screenshots

(M5 shows 11, M4 showed 64 – see attached screenshot)

Your environment

• openHAB version: 5.0.0.M5 (regression from 5.0.0.M4)
• Operating system: Debian 13
• Java runtime: Temurin 21
• Number of rules configured: 64 managed via rules files

[openhab-bot]

This issue has been mentioned on openHAB Community. There might be relevant details there:

https://community.openhab.org/t/openhab-5-2-milestone-discussion/168410/181

[Nadahar]

First, let's establish if it's a problem in OH itself or in the UI. Go to "Developer Tools ⇾ API Explorer ⇾ rules". Pick the first one under "rules", /rules and execute. In the output, are all the rules present, or only the 11?

[Nadahar]

On the forum, you indicated that this was JavaScript rules - but here you state that they are from .rules files. That would make them DSL, you can't have JavaScript in .rules files. Which of these it is decides what troubleshooting steps to take, if you confirm that they are indeed not loaded using the API Explorer.

[miloit]

Only 11

I have only js files and they are located in automation/js/***.js

The "." was just a mistake sorry

[Nadahar]

Ok, try to set org.openhab.automation.jsscripting to DEBUG logging level and restart OH to see if anything relevant shows up in the log.

[lolodomo]

Version is wrong in title and title could already help by indicating which kind of rules is concerned.
Can someone fix the title : Rules list incomplete in 5.2.0.M5 – only 11 of 64 JS file rules shown

[Nadahar]

Depending on what the problem is, it might need to be moved to the addons repo as well.

[miloit]

file.log
js.log

just to complete the picture

# maxdepth 1 -name '*.js' | wc -l
29
# find /etc/openhab/automation/js -maxdepth 1 -name '*.js' -printf '%f\n' | sort
astro.js
avl_abfall.js
blitzortung.js
deye.js
dwd.js  <<< last file openhab loads
energy.js
graph.js
haspmaster.js
hue.js
miflora.js
mopidy.js
mosmix.js
naming.js
network.js
nukihub.js
pihole.js
regenradar.js
RNTs-tlp3.js
rollershutters.js
scenes.js
soleforcast.js
switch.js
system.js
testing.js
testnew.js
transport_tracker.js
watering.js
wled.js
zigbee.js

the logs ar attached nothing really happens there unnormal

[lolodomo]

@florian-h05 for information as the issue was created in the wrong repo.

[florian-h05]

I don't think this is in the wrong repo.
File watching is done by core.

The logs indicate that once a script is provided to JS Scripting, it is provided properly. I see a single ERROR for deye.js, but no more issues.

@miloit
Unfiltered logs would be helpful. As long as there are no "(Re-)Loading script ..." from the AbstractScriptFileWatcher, file loading does not happen in core and the add-on can do nothing about.

Looking through file.log, the WatchService seems very busy with two JSON files related to DWR radar.
I could imagine this exhausts the file watching (inotify) capacity the OS provides to openHAB.

Probably very similar to #4529.

[miloit]

events.log
jsscript.log
openhab.log
watchservice.log

You mean openhab.log? And events.log? see attached I remove the json files same behaviour

no filter is set

[Nadahar]

I see a single ERROR for deye.js, but no more issues.

Isn't it a bit suspicious that the exception is the very last thing that happens in the log?

2026-06-01 06:45:00.799 [ERROR] [.automation.jsscripting.file.deye.js] - Failed to execute rule Lastmanagement: java.lang.IllegalStateException: Duplicate key event (attempted merging values Timer 991db4dd-2667-4f42-83ec-599f99fe1694 triggered. and Timer 991db4dd-2667-4f42-83ec-599f99fe1694 triggered.): Duplicate key event (attempted merging values Timer 991db4dd-2667-4f42-83ec-599f99fe1694 triggered. and Timer 991db4dd-2667-4f42-83ec-599f99fe1694 triggered.)
    at java.util.stream.Collectors.duplicateKeyException (Collectors.java:135:0)
    at java.util.stream.Collectors.lambda$uniqKeysMapAccumulator$1 (Collectors.java:182:0)
    at java.util.stream.ReduceOps$3ReducingSink.accept (ReduceOps.java:169:0)
    at java.util.HashMap$EntrySpliterator.forEachRemaining (HashMap.java:1858:0)
    at java.util.stream.AbstractPipeline.copyInto (AbstractPipeline.java:509:0)
    at java.util.stream.AbstractPipeline.wrapAndCopyInto (AbstractPipeline.java:499:0)
    at java.util.stream.ReduceOps$ReduceOp.evaluateSequential (ReduceOps.java:921:0)
    at java.util.stream.AbstractPipeline.evaluate (AbstractPipeline.java:234:0)
    at java.util.stream.ReferencePipeline.collect (ReferencePipeline.java:682:0)
    at _getTriggeredData (rules.js:470-481:20522-20814)
    at doExecute (rules.js:305:14532-14555)

Is it possible that this exception isn't caught and ends up crashing the thread that does the file watching? I'm a bit confused by which code logs the exception, and thus, whether it is "contained" or not.

[miloit]

Every rule with more than one trigger now throws on execution as soon as a GenericCronTrigger / TimeOfDayTrigger fires:

[ERROR] [.automation.jsscripting.file.deye.js] - Failed to execute rule Lastmanagement:
java.lang.IllegalStateException: Duplicate key event
(attempted merging values Timer 991db4dd-2667-4f42-83ec-599f99fe1694 triggered.
 and Timer 991db4dd-2667-4f42-83ec-599f99fe1694 triggered.)
    at java.util.stream.Collectors.duplicateKeyException (Collectors.java:135)
    ...
    at _getTriggeredData (rules.js:470-481)
    at doExecute (rules.js:305)

Root cause

openhab-js _getTriggeredData() collapses the raw context map and strips the module-ID prefix from each key, using Collectors.toMap without a merge function:

// src/rules/rules.js
const input = rawInput.entrySet().stream().collect(
  Collectors.toMap(
    entry => {
      const idx = entry.getKey().indexOf('.');
      return idx >= 0 ? entry.getKey().substring(idx + 1) : entry.getKey();
    },
    entry => entry.getValue()
  )
);

Under M5 the context now contains both a bare event entry and a module-prefixed <moduleId>.event entry that reference the same event object (note the identical TimerEvent UUID 991db4dd… in the merge message). After the prefix is stripped, both map to key event, and the 2-arg toMap has no resolver → IllegalStateException. This only surfaces with multiple triggers; single-trigger rules don't produce two entries that collapse to the same key, which is why most of my rules still run.

Minimal repro

const { rules, triggers } = require('openhab');
rules.JSRule({
  name: 'Duplicate key repro', id: 'DuplicateKeyRepro',
  triggers: [
    triggers.ItemStateChangeTrigger('SomeItem'),
    triggers.GenericCronTrigger('0 0/5 * * * ?')
  ],
  execute: () => console.log('fired')
});

Relation to this issue

This is an execution-time failure, whereas #5619 is about rules not being listed/counted, so it may be a distinct bug. But both appeared on the same M4→M5 step and both live in the JS rule path, so the changed trigger-/module-context handling in M5 looks like a plausible common root cause. Happy to split this into its own issue (likely against openhab-js) if you'd prefer.

Possible fixes

• core: avoid emitting both the bare and the module-prefixed context entry for the same output, or
• openhab-js: pass a merge resolver, e.g. Collectors.toMap(keyFn, valFn, (existing, replacement) => existing) — safe here since the colliding entries reference the same event.

Environment: openHAB 5.0.0.M5 (regression from M4), Debian 13, Temurin 21, file-based JS rules.