It's necessary because even for stable a.b.c.d we can't guarantee that
this reference will always point to the same symbol because
different capture type instantiations generate different scopes
with different resulting symbol instances.
This is mostly a revert of 2f61a2f56f
There, we erroneously assumed that we may take captured types as equal
if they are based on the same-typed projections.
Each instance of capturing defines its own captured type,
that should not be equal to any other type captured in other place.
Initial motivation was brought by FP Ultimate, where a piece of code
from the new test was found that started working differently after
recent changes.
The most obvious consequence is the change in addAllProjection.fir.kt:
one cannot use an instance as an argument when expected type
is captured type based on the same instance.
Otherwise, it would lead to CCE if we allowed to put arbitrary charsequences
to the list that initially was a MutableList<String>
All other test data changes (but addAllProjection.fir.kt and differentCapturedTypes.kt)
are irrelevant and will be fixed in the subsequent commits
This directive anyway does not make test run twice with OI, and with NI
It only once run the test with specific settings (// LANGUAGE)
and ignores irrelevant (OI or NI tags)
To do so, inside the root cause of inapplicable candidate errors,
we will record expected/actual type of receiver, if any.
That will help identifying inapplicable calls on nullable receiver.
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
Previously helpers from checkType.kt was in special package, and
if directive was enabled then test runner (`AbstractDiagnosticTest`)
injected additional imports to test files and removed them after test
was completed.
It's very hard to support such behavior in new test infrastructure so
there was a decision about changing `CHECK_TYPE`:
1. All helpers from `checkType.kt` now stays in default package
2. `CHECK_TYPE` only adds `checkType.kt` to set of analyzed files
and don't modify their content
For test which are written in default package (most of tests actually)
there are no changes. On the other hand if there is a test where dev
want to use checkType functions in testfile with some package then he
should explicitly import functions which he needed (`checkSubtype`,
`checkType`, `_`)
Fix compilation errors, revealed by this fix.
SmartCastManager is unnecessary for error reporting, intermediate diagnostics from the NI contain all required infromation.
When SmartCastManager is used it leads to missing unstable smart casts in case of expressions with captured types.
This happens, because data flow info is recorded for original expression without captured types, which is used as a key.
DataFlowValues created from receivers with captured types can't be used to retrieve that info.
^KT-39010 Fixed
isLocalVariable added. Assignment / initialization analysis.
Control whether a variable is changed in a loop at the beginning and at the end of the loop.
Control whether a variable is captured in a closure. #KT-3175 Fixed. #KT-2266 Fixed.
Tests for variable null safety and for variables is/as operations.
Loop / closure / variable property tests are included.
Old tests changed in accordance with KT-3175. In particular, all three of testSmartcastImpossible were fixed.
Denis.Zharkov