Call macros are to be parameterized with logicals and other parameters.
To replace combined expand/call macros causing confusion.
Context parameters to be passed explicitly.
Modified TYPECHECK query to include "checkingScope" query parameter:
it is a root node for which program is generated (for local typechecking it's local node).
Rules additionally match on a constraint `checkDecls(node<>)` bearing this node,
so while rules will be generated for all scopes, only those in applicable scopes will run.
Also split rule for EnumConstantDeclaration on declration checking and just typing rule.
Scheduler in rxjava2-swing seems to cause a peculiar bug
that is difficult to reproduce.
Seems like AtomicReference or one of its subclasses
gets itself as the value, which triggers infinite recursion.
UnitTypecheckingCacheHolder now holds the data needed for ChangesTrackers.
This data is updated by UpdatesRecorder: different models and nodes from different roots are tracked.
ExpressionLogicalVariable is replaced with ExpressionItem wrapping
the same expression.
Fixes ensure correct application of transformations.
Late expression can't be used as "placeholder" logical item.
This expression is to be used in those rare cases
when a logical var is to be passed to an external method.
Not available within EvaluationSetting (new marker interface).
Normally logical variable (as an instance of LogicalItem)
is only to be used within a logical context.
Logical item value expr only to contain a logical variable.
SNode is supported now directly as an argument to constraint/predicate.
Pointer operation is no longer necessary, as is node-ptr<> type for logical vars.
Previously there were two independent ChangeTracker-s in UpdatesRecorder and UnitTypecheckingCacheHolder.
Now, given more elaborate logic for tracking changes to types (which includes changes in different roots)
this information must be propagated from UpdatesRecorder to cache holder,
so that if there're changes relevant for recomputing types then types cache is also invalidated.
To avoid confusion, a rule defines "target ref", which is supposed
to indicate the source code location for possible errors.
The origin node is to be referred as "origin" as before.
Temporarily rename OriginSpec's alias to "target" to reflect change
in semantics.
Sometimes UpdatesRecorder can log changes of SNodes which have same SNodeId, but different identity.
Thus reversedRequirements map returned nothing on such nodes and these require-dependencies were not handled.
Such case is described in MPSCR-42 ticket.
LogicalItemValueExpression is to play the same role as valueOf(logical)
within an instance of EvaluationSubject.
LogicalItemExpression is to represent the logical variable itself.
This is an important change in coderules syntax: all expressions to be evaluated from
constraints/predicates are to be wrapped in parentheses and preceded by dollar sign.
So, instead of eval/backticks the new syntax is to be used: $(<expression>).
Also the expressions to be passed to "feedback reporting" predicates are to be wrapped similarly.
TypesIndex now invalidates report items from previous sessions not by node refs,
but by tags of rules which produced them. Set of invalidated tags is an addiitonal
output from incremental engine.
Now principal rule head must include only principal constraints or be empty.
Without this restriction journal may include stale chunks which have no information to invalidate them.
A term may indicate that it "extends" another open term.
An extended term has essentially the same form as the one
being extended, so that they can be unified.
TypecheckingQueries must track the types index used for reporting
the state of cache, since the index itself is built independently
of whether or not 'checkRecursively' is invoked.
Introduce SessionData, allow to keep arbitrary data with the session.
Extend RulesListTable with support for invalidation based on rule unique tags.
Introduce UpdatesRecorder to keep track of invalidated nodes.
Extend ApplyTemplatesStep with invalidate logic for typechecking.
Simplify design, drop unnecessary fields, minor refactorings.
The inferred type can be either capture or type variable.
This logic is covered by a new rule 'inferFromCapture'.
For examples see 'Bounds' test sample.
This inference is used in call arguments type inference.
It happens on cases when bounds arise from several places:
(a) bounds for inference variables from type vars declared in method signature (hasBound constraint)
(b) bounds of type vars used as arguments to call (typeVariableType term wiht bounds inside)
(c) bounds from captures on type variables from case (b), which come in the form of 'captureOf' term
CopyUtil didn't correctly map logicals nested inside DataForms which are bound to Logicals.
That's because CopyUtil relies on DataForm:variablesCount() and
LogicalDataForm:variablesCount didn't consider the case when bound value is itself DataForm.
Fedor Isakov