Quick quiz:
Q: In a CFG, what does `a -> b -> c -> d` mean?
A: `a`, then `b`, then `c`, then `d`.
Q: In a CFG, what does `a -> b -> d; a -> c -> d` mean?
A: `a`, then `b` or `c`, then `d`.
Q: So how do you encode "a, then (b, then c) or (c, then b), then d`?
A: You can't.
Problem is, you need to, because that's what `a; run2({ b }, { c }); d`
does when `run2` has a contract that it calls both its lambda arguments
in-place: `shuffle(listOf(block1, block2)).forEach { it() }` is a
perfectly valid implementation for it, as little sense as that makes.
So that's what union nodes solve. When a node implements
`UnionNodeMarker`, its inputs are interpreted as "all visited in some
order" instead of the normal "one of the inputs is visited".
Currently this is used for data flow. It *should* also be used for
control flow, but it isn't. But it should be. But that's not so easy.
BTW, `try` exit is NOT a union node; although lambdas in one branch can
be completed according to types' of lambdas in another, data does not
flow between the branches anyway (since we don't know how much of the
`try` executed before jumping into `catch`, and `catch`es are mutually
exclusive) so a `try` expression is more like `when` than a function
call with called-in-place-exactly-once arguments. The fact that
`exitTryExpression` used `processUnionOfArguments` in a weird way
should've hinted at that, but now we know for certain.
In this commit we upgrade FIR builder inference logic from
the compiler version to 1.7. FIR-based compiler now works with
"don't use builder inference" flag always ON and supports switching
the flag "use builder inference only if needed". To do it,
ContraintSystemCompleter (FIR) and KotlinConstraintSystemCompleter (FE 1.0)
are made similar with extracting some common parts into
ConstraintSystemCompletionContext.
Test status: one BB test fails after this commit (KT-49285).
Also we have a crush in DFA logic in FIR bootstrap test and somehow
questionable behavior in FIR diagnostic test. However,
two BB tests were fixed, the 3rd case from KT-49925 were also fixed.
#KT-49925 Fixed
In this commit we add nullability to upper type of a substituted
constraint in the situation like (Stub<_L> <: SomeType),
where _L is fixed to nullable Stub<_L>?.
We have to change this constraint to L <: SomeType? and not to
L <: SomeType as before, otherwise nullability become broken
(direct substitution of Stub<_L> to L is illegal here).
#KT-50470 Fixed
Some existing tests start failing after previous commits adding @Exact
attribute to `?:`
They have a form:
var x: String? = nullable()
if (x == null) {
x = nullable() ?: "" // considering @Exact the whole elvis is inferred to nullable from expect type
}
x.length // should be smart cast
- Add utilities to add new attribute to ConeAttributes
- Get rid of FlexibleNullability attribute (it can be easily inferred
for any flexible type at any moment)
- Fix determining of EnhancedNullability attribute
Update includes:
- Changing syntax of `OI/`NI` tags from `<!NI;TAG!>` to `<!TAG{NI}!>`
- Fix some incorrect directives
- Change order of diagnostics in some places
- Remove ignored diagnostics from FIR test data (previously `DIAGNOSTICS` didn't work)
- Update FIR dumps in some places and add `FIR_IDENTICAL` if needed
- Replace all JAVAC_SKIP with SKIP_JAVAC directive
Before that commit we desugared `a ?: b` as
when (val elvis = a) {
null -> b
else -> elvis
}
It was incorrect, because `a` should be resolved in dependent mode,
but when it was `elvis` initializer it was resolved in independent
mode, so we can't infer type for `a` in some complex cases
Steven Schäfer