Release v7.141.0

The skills now teach where read models and graphs are stored: the default sinks, named sinks in a Screenplay model, [Sink("name")] in Chronicle, and the Kernel's sinks configuration.

Added

• The Screenplay read surface skill teaches where a read model is stored: sink <Name> for readmodels|graphs at the application root, a bare sink <Name> line in the read model, the rule that the line is a property unless the name is a declared sink, @sink <Type> for the property, and no sink on a passive read model. (#1049)
• The Screenplay projections skill shows a graph in a named sink (sink People for graphs) and says that a projection, reducer or graph projection writes to its target's sink and never names one.
• The Screenplay model authoring and toolchain skills list sink among the application-root declarations, the PLAY0931–PLAY0939 diagnostics with their severities, and the syntax-schema members ApplicationSyntax.sinks, SinkSyntax, ReadModelSyntax.sink and GraphSyntax.sink.
• The Chronicle read model and projection skills teach [Sink("name")] on read model and graph types, the client defaults DefaultSinkName and DefaultGraphSinkName, analyzer CHR0069 and the replay a sink change causes.
• The Chronicle hosting skill shows the Kernel sinks configuration with a MongoDB and a SQL sink, Neo4j graph sinks, the environment-variable form, and keeping read models in a different engine than the events.

Changed

• The render and gap-fill skill tells renderers to hash the sink metadata (Sinks, ReadModelSinks, GraphSinks) into their render input, because a sink change leaves the semantic revision unchanged.
• The model review checklist gains a sinks check: every declared sink is used and names differ ignoring case.
• Every Screenplay example is checked against Screenplay 6.4.0, and the UI composition evaluation expects an accent in the selected theme rather than in appearance (PLAY0929).