eventb-soton / emf_inclusion Goto Github PK

Although there is less reason to include a context it may still be useful (and solve the problem of contexts being seen from other projects?).
Also is there a case for prefix renaming of sets and constants? E.g. maybe two copies of an included machine need different configuration.

A warning is generated if more than one machine is included without prefixing.
I think there is always an ambiguity over Initialisation so maybe it should be an error?
Or could we implicitly include all initialisations without needing to use 'synchronises.
If included machines do not contain any collisions it might be nicer to avoid prefix?

When prefix re-naming an included refined machine, references to abstract variables (that have been replaced in the refinement and only appear in the invariants/witnesses) are not renamed. This means it is not possible to refine composite machines using inclusion of an abstract - refined component sequence.

When using machine inclusion, some times the generated Rodin files disappeared. This is one of the cause of the issue for CamilleX
eventB-Soton/XEventB#59

contexts cannot be seen from another project.
This means the context for an included machine has to be copied into the project of the composite machine even though the included machine may be in another project.