0.7.1: require ObjectState 3

[trissim]

Caller update for ObjectState 3.0.0 (OpenHCSDev/ObjectState#50), with formbridge 0.1.2 (OpenHCSDev/formbridge#9). Not tagged or released.

Version: 0.7.1

No public name of pyqt-reactive is removed, so this is a patch release. What goes away is private: AbstractManagerWidget._live_context_resolver (never called), the positional shape of ProvenanceNavigationMixin._compute_target_path, and the inline code of WidgetService._apply_context_behavior.

Sites changed (the 10 in the 3.0.0 caller ledger, plus the two it marked for review)

Each read of a config value now names which read it is: raw (what was set on the object), resolved (after inheritance), or preview (the resolved value with its source, while the field's own value stays unset).

• services/widget_service.py: build_context_stack(..., ancestor_objects=...) raises TypeError in ObjectState 3. The site did not need a context stack at all. It was a second resolver for a preview: it rebuilt the stack by hand, asked for placeholder text, then read the attribute again for the value, and used the default prefix instead of the form's. It now shows state.get_resolved_value(path) through one method, WidgetService.show_resolved_value_as_placeholder, and the two refresh paths in ParameterOpsService (which each repeated resolve-and-format) call the same method.
• widgets/shared/clickable_help_components.py (ledger: review): get_provenance can return (None, source_type) for a state without a scope id. The ledger expected the same-window branch to handle it; it did not, because the target path was computed first from ObjectStateRegistry.get_by_scope(None), which is the global state. The source state is now the widget's own state whenever the source scope is the state's own scope. The source is read once per action (_provenance_source), without re-checking the tuple's shape.
• widgets/shared/abstract_manager_widget.py (ledger: review): the unused LiveContextResolver instance is deleted.
• pyproject.toml: objectstate>=3.0,<4, formbridge>=0.1.2,<0.2.
• requirements-ci.txt: ObjectState at abe763c and formbridge at 1e73e6c until both are on PyPI.
• tests/test_preview_formatters.py: declares abbreviations with @abbreviation instead of writing to the removed GROUP_ABBREVIATIONS_REGISTRY. The line that gave a lazy class a different abbreviation is gone: that state cannot exist in ObjectState 3.
• tests/test_function_list_editor_tokens.py: no longer assigns the removed _atomic_* attributes.
• New tests/test_provenance_navigation.py: a value supplied by a sibling config scrolls the same window to the sibling field, with and without a scope id.

No hits in this repo for the other removed names (cache-clear functions, resolve_field_inheritance, legacy_entries, the tuple-shaped layer stack); checked again with one grep over src, tests, docs, examples and scripts.

Left as found, for the owner: formbridge.forms.parameter_form_service.ParameterFormService.get_placeholder_text has no caller after this change.

Checks run locally (CPython 3.12, ObjectState abe763c, formbridge 1e73e6c)

• pytest tests -n 2: 593 passed, 1 failed. The failure is test_release_dependency_declarations.py::test_published_runtime_is_not_overridden_by_a_source_candidate, which compares the installed zmqruntime with the declared bound; it fails the same way at 0.7.0 with ObjectState 2.1.1 in this environment (run there too).
• ruff check src/ --select E9,F63,F7,F822,F823: clean.
• scripts/check_ruff_added_lines.py --base origin/main --head HEAD: no diagnostics on added lines.
• Docs build, black and mypy (non-blocking in CI): not run.

Release order

ObjectState 3.0.0, formbridge 0.1.2, then this (empty requirements-ci.txt first), then OpenHCS.

🤖 Generated with Claude Code

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.