Repository navigation
M2C: indexed IEC FCDA Global Data refresh with bounded MMS targets - #161
Conversation
… references Resolve watched DATA and DataSet DO/DA members through the indexed canonical IEC/MMS alias map instead of repeatedly scanning every node and comparing incompatible reference spellings. Deduplicate MMS targets, honor FC, hard-bound the requested read count, and fail closed on unmatched members. Coalesce observed UI value updates over 50ms without any cyclic MMS polling. Add large-model regression for IEC/MMS aliases, DO FC filtering, and bounded refresh.
…lidation Extend canonical indexed Global Data resolver for LD/LN$FC$DO references previously supported by literal prefix matching. Normalize to exact indexed IEC DO, enforce matching FC, and reject incomplete/ambiguous prefixes. Add regression for four MX children and FC mismatch negative case.
Replace forbidden QML timer with Qt.callLater(root.syncRows), coalescing hot canonical dataChanged/valuesChanged notifications without cyclic polling, new runner schedules, or background reads. Extend P6A contract to enforce event-driven refresh.
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 5a5fdd2e55
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| const auto [reference, fc] = splitDataSetMemberReference(raw); | ||
| if (reference.isEmpty()) continue; | ||
| QString effectiveFc = fc; | ||
| auto found = referenceIndex_.constFind(reference); |
There was a problem hiding this comment.
In the SCL-assisted-connect path, this exact lookup cannot resolve DataSet members: build_signal_reference() produces references such as IED/LD/LN.DO.DA [MX], while IedEngineeringContextController::dataSets() appends another FC suffix and the tree index uses the MMS domain form IEDLD/LN.DO.DA. splitDataSetMemberReference() removes only the final suffix, leaving a key that never exists, and the $ fallback does not apply, so refreshing an opened-SCL DataSet returns no read targets. Convert the canonical SCL reference to the indexed MMS-domain identity, or pass/index the member's existing mms_reference, so Open SCL and live discovery resolve through the same semantics.
AGENTS.md reference: AGENTS.md:L299-L301
Useful? React with 👍 / 👎.
1 participant
Baseline
Stacked after integrated M2B PR #160 on
main(d527c894). Native Workbench only; no changes to SMV/PTP/ESP32 or MMS protocol engine.Bug fixed
DataSets are now displayed via canonical IEC/MMS reference alias lookup, but Global Data → Refresh watched still used a linear scan requiring a literal
node.reference. IEC FCDA references (e.g.LD/LN.DO.DA [ST]) therefore could fail to produce actual MMS targets even when DataSets resolved correctly.Implementation
referenceIndex_from M2B to resolve watched DO/DA with exact FC filtering, including IEC and MMS spellings.Acceptance
Do not merge until exact-head CI success. This fixes data authority, not full q/t bit-level decoding or pixel-perfect visual parity.