Component: mxcli (v0.16.0) — MDL extraction / DESCRIBE MICROFLOW
Current behavior
DESCRIBE MICROFLOW correctly renders explicit commit $X ... activities, but for create and change activities it never reveals whether the Studio Pro "Commit" checkbox is ticked. The MDL output is indistinguishable between:
$NewObject = create Module.Entity (Attr = $Value); -- Commit = No
$NewObject = create Module.Entity (Attr = $Value); -- Commit = Yes
Both render identically. The same applies to change $X (...).
Why this matters for reliable bug analysis
Root-cause investigations of data-integrity bugs (missing/orphaned records, partial saves, unexpected rollback side effects) hinge entirely on knowing the exact commit boundaries in a microflow. Without the inline Commit flag, every such investigation has to fall back on manually opening Studio Pro and checking each activity by hand — which:
- Blocks automated/AI-assisted analysis from reaching a conclusion without human-in-the-loop verification at every step.
- Is error-prone under time pressure: it's easy to wrongly assume "no commit" (or vice versa) from bare MDL output and draw an incorrect conclusion about a bug's cause.
- Doesn't scale — a microflow chain with dozens of create/change activities across multiple sub-microflows requires checking each one individually in the modeler, which is slow and easy to get wrong or skip.
- Makes automated tooling (lint rules, impact analysis, SHOW IMPACT OF, transaction-boundary detection) fundamentally incomplete, since none of them can currently reason about where data actually lands in the database versus staying in-memory/uncommitted.
This is a recurring, confirmed pain point: an investigation into an object-orphaning bug was only resolved after manually confirming Commit flags in Studio Pro for each activity — mxcli output alone was insufficient and actively misleading during the analysis.
Proposed solution
Extract and render the Commit flag on create and change activities in DESCRIBE MICROFLOW output, e.g.:
$NewObject = create Module.Entity (Attr = $Value) commit;
or an equivalent explicit keyword/flag (commit refresh, matching the syntax already used for standalone commit statements). This should also flow through to SHOW CALLERS OF / SHOW IMPACT OF and any future transaction-boundary/lint tooling, so autocommit-risk patterns (uncommitted create that's only committed via an associated object, never explicitly) can eventually be flagged automatically (e.g. a lint rule: "created object with Commit=No has no reachable explicit commit path").
Acceptance criteria
DESCRIBE MICROFLOW shows the Commit flag for every create and change activity, matching what Studio Pro shows.
Round-trip safe: re-importing the described MDL preserves the Commit flag.
mxcli check --references (or equivalent) does not regress on existing scripts.
Component: mxcli (v0.16.0) — MDL extraction / DESCRIBE MICROFLOW
Current behavior
DESCRIBE MICROFLOW correctly renders explicit commit $X ... activities, but for create and change activities it never reveals whether the Studio Pro "Commit" checkbox is ticked. The MDL output is indistinguishable between:
$NewObject = create Module.Entity (Attr = $Value); -- Commit = No
$NewObject = create Module.Entity (Attr = $Value); -- Commit = Yes
Both render identically. The same applies to change $X (...).
Why this matters for reliable bug analysis
Root-cause investigations of data-integrity bugs (missing/orphaned records, partial saves, unexpected rollback side effects) hinge entirely on knowing the exact commit boundaries in a microflow. Without the inline Commit flag, every such investigation has to fall back on manually opening Studio Pro and checking each activity by hand — which:
This is a recurring, confirmed pain point: an investigation into an object-orphaning bug was only resolved after manually confirming Commit flags in Studio Pro for each activity — mxcli output alone was insufficient and actively misleading during the analysis.
Proposed solution
Extract and render the Commit flag on create and change activities in DESCRIBE MICROFLOW output, e.g.:
$NewObject = create Module.Entity (Attr = $Value) commit;
or an equivalent explicit keyword/flag (commit refresh, matching the syntax already used for standalone commit statements). This should also flow through to SHOW CALLERS OF / SHOW IMPACT OF and any future transaction-boundary/lint tooling, so autocommit-risk patterns (uncommitted create that's only committed via an associated object, never explicitly) can eventually be flagged automatically (e.g. a lint rule: "created object with Commit=No has no reachable explicit commit path").
Acceptance criteria
DESCRIBE MICROFLOW shows the Commit flag for every create and change activity, matching what Studio Pro shows.
Round-trip safe: re-importing the described MDL preserves the Commit flag.
mxcli check --references (or equivalent) does not regress on existing scripts.