Daily Digest: 2026-09-26

[github-actions]

We are given a daily digest for 2026-09-26. We need to structure the response as:

1. Highlights (1-3 bullet points of the most important changes)
2. Code Changes (summarize commits by theme, not individually)
3. Pull Requests (merged, opened, pending review)
4. Issues (new, closed, total open)
5. Repository Health (stars, forks, CI status)
6. Action Items (anything that needs attention: failing CI, stale PRs, unanswered issues)

Let's break down the provided data:

Repository Stats

• Stars: 126
• Forks: 50
• Watchers: 6
• Open issues: 1 (but note: the issues section lists several opened and closed, so we'll see)

Commits (last 24h)

There are many commits, all merged via pull requests (as seen by the merge commit messages). We note that the commits are mostly merges of pull requests and some direct commits (like f1fd11c3 and 7c8619bb). However, the instruction says to summarize by theme.

Looking at the commit messages, we can group them by the type of fix or feature:

• Several commits related to "fix" and "lint" (e.g., f1fd11c3 for java-actions, 7c8619bb for lint, aaa56457 for lint, c1f8c899 for pages)
• Many merges of pull requests that are fixes (e.g., Consumed OData ConfigurationMicroflow service fails CE5111 on Mendix 11.11 (both engines) #690, feat(lint): expose XPath/expression AST to Starlark rules #688, docs(microflows): add NPE association retrieval and type conversion functions #687, etc.)

But note: the "Pull Requests Merged" section says "No PRs merged." This seems contradictory because the commits are merge commits of PRs. However, note that the commits listed are from the last 24h, and the "Pull Requests Merged" section might be referring to PRs that were merged in the last 24h but not listed as merged in the provided data? Actually, the provided data under "Commits" shows many merge commits (e.g., bf119f60 is a merge of PR #1204). But then the "Pull Requests Merged" section says none. This is confusing.

Looking again: the "Commits" section lists commits that are merges (like bf119f60 is "Merge pull request #1204"). However, the "Pull Requests Merged" section says "No PRs merged.". This might be an error in the provided data, or perhaps the "Pull Requests Merged" section is only for PRs that were merged and not via a merge commit? But that doesn't make sense.

Alternatively, note that the "Commits" section includes both merge commits and direct commits (like f1fd11c3 and 7c8619bb). The merge commits are the result of merging PRs, so we should consider that PRs were merged. However, the explicit "Pull Requests Merged" section says none. We must follow the data as given.

But wait: the "Pull Requests Merged" section is under the heading "Pull Requests Merged (last 24h)" and it says "No PRs merged.". However, the commits list shows many merge commits. This discrepancy might be because the merge commits are from PRs that were merged earlier but the merge commit happened in the last 24h? Or perhaps the data is inconsistent.

Given the instructions, we have to use the data as provided. So for the "Pull Requests" section, we will note that the "Pull Requests Merged" section says none, but we see in the commits that there are merge commits. However, the problem says: "summarize commits by theme", so we can ignore the PR merged count for the code changes and focus on the commit messages.

Let's look at the issues and PRs:

Pull Requests Opened (last 24h)

• No new PRs.

Currently Open PRs

• A list of 8 open PRs (from fix(lint): read an action activity's error handling from the action #1203 down to docs: add manual test case documentation for all MDL command areas #386)

Issues Opened (last 24h)

• Lint CONV013 flags handled Java/REST/web service calls, CONV014 misses "on error continue" on actions #1202, ALTER PAGE … SET ImageUrl = … ON <image> reports "Altered page" and changes nothing: the pluggable setter writes PrimitiveValue into a TextTemplate-kind property, so DESCRIBE, mx check and the runtime never see it #1201, Pluggable widgets built from the .mpk lose their <actionVariables>: the action's ValueType is written with an empty ActionVariables list, and mx check reports CE0463 #1200, File Uploader: an allowedfileformat in configMode: 'simple' passes check and exec, then fails mx check with CE0463 — hidden TextTemplates inside object-list items are never nulled #1198 -> 4 new issues

Issues Closed (last 24h)

• DESCRIBE of a File Uploader page does not re-execute: it prints a generic DataSource: that the builder then rejects as ambiguous (associatedFiles / associatedImages) #1199, Daily Digest: 2026-09-25 #1195, @caption on a while loop is dropped with no MDL042 — only for-each loops are warned #1187, documents with same not discernable in catalog #1185, Daily Digest: 2026-09-24 #1184, Catalog java actions return type confusing in case of typeparam named after primitive #1183, Starlark rule API: documented literals drift from what the linter emits — microflow_type has three spellings, two of which never occur #1178, RESET LAYOUT microflow option is missing from mxcli entirely #837 -> 8 closed issues

CI Status (latest runs on main)

• Three workflows:
  • Daily Project Digest (status not shown? It says " — " which might mean skipped or not run? But then two successes: Build, Test & Lint and Deploy Documentation)
  • Actually, the first one is listed as " — Daily Project Digest [2026-09-26T11:59:05Z]" which might be a placeholder? But then the next two are success.

Dependency Changes

• None.

Now, let's try to extract:

Highlights

We need 1-3 bullet points of the most important changes.

Looking at the commits, we see:

• A fix for java-actions: f1fd11c3 (primitive-named type parameters and list of a type parameter)
• Several lint fixes: 7c8619bb (name nanoflows and rules correctly), aaa56457 (pin documented Starlark field values), etc.
• A fix for pages: c1f8c899 (refuse an attribute binding with no object to bind to)

Also, note that there are many merges of fix PRs (like #690, #688, etc.) but we don't have the details of those PRs.

However, the issues opened recently include:

• Lint CONV013 flags handled Java/REST/web service calls, CONV014 misses "on error continue" on actions #1202: Lint CONV013 and CONV014 issues
• ALTER PAGE … SET ImageUrl = … ON <image> reports "Altered page" and changes nothing: the pluggable setter writes PrimitiveValue into a TextTemplate-kind property, so DESCRIBE, mx check and the runtime never see it #1201: ALTER PAGE ... SET ImageUrl issue
• Pluggable widgets built from the .mpk lose their <actionVariables>: the action's ValueType is written with an empty ActionVariables list, and mx check reports CE0463 #1200: Pluggable widgets losing actionVariables
• File Uploader: an allowedfileformat in configMode: 'simple' passes check and exec, then fails mx check with CE0463 — hidden TextTemplates inside object-list items are never nulled #1198: File Uploader issue

And closed issues include some related to lint and design properties.

Given the recent activity, a highlight could be the ongoing work on lint rules and fixes for various MDL issues.

But note: the repository has only 1 open issue according to the stats, but we see 4 new issues opened and 8 closed in the last 24h. The stats say "Open issues: 1", which might be the current total? However, the issues section lists 4 new and 8 closed, so the net change is -4, meaning if there were 5 open issues before, now there is 1. So the stat of 1 open issue is current.

Code Changes (by theme)

We can group the commits by the area they touch:

1. Java Actions:

   • f1fd11c3: fix(java-actions): primitive-named type parameters and list of a type parameter

2. Lint / Starlark Rules:

   • 7c8619bb: fix(lint): name nanoflows and rules correctly in shipped Starlark rules
   • aaa56457: fix(lint): pin documented Starlark field values to what the API emits
   • Also, note the commit 1ae6737f: fix: resolve every widget keyword's design properties through the $Type it writes (Ako) - this might be related to lint or design properties?
   • And c1f8c899: fix(pages): refuse an attribute binding with no object to bind to (MDL-WIDGET34) (Ako) - this is under pages but might be related to lint?

3. Pages:

   • c1f8c899: as above
   • Also, note several merges of PRs that are fixes for pages (e.g., feature/test infra governance track a #693, feat(lint): expose scheduled_events() query function in Starlark rules #692, Doctype mx-check gate (TestMxCheck_DoctypeScripts) runs only on the legacy engine — misses modelsdk regressions #691, etc.) but we don't have the details.

4. Design Properties:

   • Many commits mention "design-property" (e.g., Daily Digest: 2026-06-25 #695, docs(microflows): add NPE association retrieval and type conversion functions #687, docs: add write-microflows skill for MDL microflow authoring #686, etc.) and the commit 1ae6737f is about resolving widget keyword's design properties.

Given the volume, we can summarize:

• Java Actions: Fixed handling of primitive-named type parameters and lists of type parameters.
• Lint Rules: Fixed naming of nanoflows and rules in Starlark, and pinned documented field values to API output.
• Pages: Fixed attribute binding checks (refusing bindings without an object) and various design property-related fixes (label, styling, input binding, etc.).
• Design Properties: Multiple fixes for design property keyword mappings, altered styling, and label checks.

Pull Requests

• Merged (last 24h): None (as per the section, though the commits suggest otherwise, we go by the provided data)
• Opened (last 24h): None
• Currently Open: 8 PRs (listed: fix(lint): read an action activity's error handling from the action #1203, docs(lint): correct field value formats in write-lint-rules skill #1128, fix(new): guard unsupported MxToolset output paths #865, deps: bump typescript from 6.0.3 to 7.0.2 in /vscode-mdl #748, fix(init): make devcontainer postCreateCommand robust on Podman-on-Windows (#742) #744, feat(new): add --container-runtime flag (parity with mxcli init) #740, docs: add manual test case documentation for all MDL command areas #386)

Issues

• New (last 24h): 4 (Lint CONV013 flags handled Java/REST/web service calls, CONV014 misses "on error continue" on actions #1202, ALTER PAGE … SET ImageUrl = … ON <image> reports "Altered page" and changes nothing: the pluggable setter writes PrimitiveValue into a TextTemplate-kind property, so DESCRIBE, mx check and the runtime never see it #1201, Pluggable widgets built from the .mpk lose their <actionVariables>: the action's ValueType is written with an empty ActionVariables list, and mx check reports CE0463 #1200, File Uploader: an allowedfileformat in configMode: 'simple' passes check and exec, then fails mx check with CE0463 — hidden TextTemplates inside object-list items are never nulled #1198)
• Closed (last 24h): 8 (DESCRIBE of a File Uploader page does not re-execute: it prints a generic DataSource: that the builder then rejects as ambiguous (associatedFiles / associatedImages) #1199, Daily Digest: 2026-09-25 #1195, @caption on a while loop is dropped with no MDL042 — only for-each loops are warned #1187, documents with same not discernable in catalog #1185, Daily Digest: 2026-09-24 #1184, Catalog java actions return type confusing in case of typeparam named after primitive #1183, Starlark rule API: documented literals drift from what the linter emits — microflow_type has three spellings, two of which never occur #1178, RESET LAYOUT microflow option is missing from mxcli entirely #837)
• Total Open: According to the repo stats, 1 open issue (but note: the issues section shows 4 new and 8 closed, so the current open issues should be the previous open minus 4 (closed) plus 4 (new) -> same as before? However, the stat says 1, so we'll use that).

Repository Health

• Stars: 126
• Forks: 50
• Watchers: 6
• CI Status:
  • Daily Project Digest: (status not clearly success, but the next two are success) -> we see two successes: Build, Test & Lint and Deploy Documentation. The first one might be a placeholder or skipped? We'll note that the main CI (Build, Test & Lint) is passing.

Action Items

We need to look for anything that needs attention:

• Failing CI: None reported (the two shown are success; the first one is unclear but not marked as failure).
• Stale PRs: We have open PRs, some of which are old (e.g., docs: add manual test case documentation for all MDL command areas #386 is from a long time ago). We should note that there are old PRs that might need attention.
• Unanswered issues: The new issues (Lint CONV013 flags handled Java/REST/web service calls, CONV014 misses "on error continue" on actions #1202, ALTER PAGE … SET ImageUrl = … ON <image> reports "Altered page" and changes nothing: the pluggable setter writes PrimitiveValue into a TextTemplate-kind property, so DESCRIBE, mx check and the runtime never see it #1201, Pluggable widgets built from the .mpk lose their <actionVariables>: the action's ValueType is written with an empty ActionVariables list, and mx check reports CE0463 #1200, File Uploader: an allowedfileformat in configMode: 'simple' passes check and exec, then fails mx check with CE0463 — hidden TextTemplates inside object-list items are never nulled #1198) are recent and may need attention.

Specifically:

• The open PR docs: add manual test case documentation for all MDL command areas #386 is from a long time ago (the number is low) and might be stale.
• Also, note that the issue stats say only 1 open issue, but we have 4 new issues opened in the last 24h and 8 closed. This implies that the 4 new issues were immediately closed? But the issues closed list does not include these new ones. Actually, the closed issues list includes DESCRIBE of a File Uploader page does not re-execute: it prints a generic DataSource: that the builder then rejects as ambiguous (associatedFiles / associatedImages) #1199 down to RESET LAYOUT microflow option is missing from mxcli entirely #837, which are older. So the 4 new issues are still open? Then why does the repo stats say 1 open issue? This is confusing.

Let me re-read the repo stats: "Open issues | 1". But the issues opened last 24h are 4 and closed last 24h are 8. So if we assume that before the last 24h there were X open issues, then:
X + 4 (new) - 8 (closed) = current open issues = 1
=> X = 1 + 8 - 4 = 5
So 5 open issues before, now 1.

Therefore, the 4 new issues are not all still open; some of them might have been closed? But the issues closed list does not include the new ones (they are #1199 and below, while the new ones are #1200+). So the 4 new issues are still open? Then we would have:
current open = (previous open) + 4 (new) - 8 (closed) = 1
=> previous open = 1 - 4 + 8 = 5
So the 4 new issues are open, and 8 old issues were closed, leaving 5 - 8 + 4 = 1? That doesn't add up.

Alternatively, the 8 closed issues include some of the new ones? But the numbers: new issues are #1198, #1200, #1201, #1202. The closed issues are #837, #1178, #1183, #1184, #1185, #1187, #1195, #1199. None of the new issue numbers are in the closed list. So the 4 new issues are still open.

Then the current open issues should be: (previous open) + 4 - 8 =

---

Daily digest generated automatically — workflow source