| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |\
| |
| | |
Test case closes SI-4818
|
| | |
| |
| |
| | |
Fixed between 2.10.0 M2 and M3, with both the old and new pattern matcher.
|
| |\ \
| | |
| | | |
SI-5318 Make implicit divergence checking PolyType aware.
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Replaces the two active subclasses of `SymCollector` with
direct use of traversal methods of `Type`.
Wildcard free class type parameters, not just method type parameters,
when stripping the core type of candidate implicits. The spec doesn't
make any such distinction, and the enclosed test, t5318c, crashes
without this change.
|
| |\ \ \
| | | |
| | | | |
Closes SI-5821.
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This was an interesting one. Basically an erroneous
import was creating an erroneous symbol for Array (similary
for other symbols that were 'found' in this import) which was
leading to all sorts of inconsistencies and spurious errors.
This wasn't a bug in ContextErrors but rather something that
existed for ages and was hidden from the general audience.
Review by @paulp.
|
| | |_|/
|/| |
| | |
| | |
| | | |
we don't handle switches with guards, whereas the old patmat did
to ease the transition, let's not error out and see how we can resolve this
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Move now-passing SI-963 test into neg.
Test for partial specialization.
Pending test for SI-5008.
Pending test for SI-4649.
Abstract array type test.
|
| | | | |
|
| |\ \ \
| | | |
| | | | |
virtpatmat: treemaker approximation refactorings and exhaustivity
|
| | |/ /
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
We check exhaustivity by representing a match as a formula in
finite-domain propositional logic (FDPL) that is false when
the match may fail. The variables in the formula represent
tested trees in the match (type tests/value equality tests).
The approximation uses the same framework as the CSE analysis.
A matrix of tree makers is turned into a DAG, where sharing
represents the same value/type being tested.
We reduce FDPL to Boolean PL as follows.
For all assignments, V_i = c_i_j, we introduce a proposition
P_i_j that is true iff V_i is equal to the constant c_i_j,
for a given i, and all j, P_i_j are mutually exclusive
(a variable cannot have multiple values).
If the variable's domain is closed, we assert that one of P_i_j
must be true for each i and some j. The conjunction of these
propositions constitutes the equality axioms.
After going through negational normal form to conjunctive normal
form, we use a small SAT solver (the DPLL algorithm) to find
a model under which the equational axioms hold but the match fails.
The formula: EqAxioms /\ -MatchSucceeds.
Note that match failure expresses nicely in CNF: the
negation of each case (which yields a disjunction) is anded.
We then turn this model into variable assignments
(what's the variable (not) equal to, along with recursive
assignments for its fields). Valid assignments (no ill-typed
field assignments) are then presented to the user as counter examples.
A counter example is a value, a type test, a constructor call
or a wildcard. We prune the example set and only report the
most general examples. (Finally, we sort the output to yield stable,
i.e. testable, warning messages.)
A match is only checked for exhaustivity when the type of the
selector is "checkable" (has a sealed type or is a tuple
with at least one component of sealed type).
We consider statically known guard outcomes, but generally back
off (don't check exhaustivity) when a match has guards
or user-defined extractor calls. (Sometimes constant folding
lets us statically decide a guard.)
We ignore possibly failing null checks (which are performed
before calling extractors, for example), though this could
be done easily in the current framework. The problem is false
positives. People don't usually put nulls in tuples or lists.
To improve the exhaustivity checks, we rewrite `List()` to Nil.
TODO: more general rewrite of List(a, b, ..., z) to `a :: b :: ... :: z`.
When presenting counter examples, we represent lists in the
user-friendly List(a,...,z) format. (Similarly for tuples.)
There are no exhaustivity checks for a match-defined PartialFunction.
misc notes:
- fix pure case of dpll solver
impure set (symbol that occurs both as a positive and negative literal)
was always empty since I was looking for literals (which are not equal if positivity is not equal)
but should have been looking for symbols
- FDPL -> BoolPL translation collects all syms in props
since propForEqualsTo generates an Or, must traverse the prop
rather than assuming only top-level Syms are relevant...
also, propForEqualsTo will not assume Or'ing a whole domain is equivalent to True
(which it isn't, since the type test may fail in general)
- improve counter example description
- treat as constructor call when we either have definite type information about a real class,
or we have no equality information at all, but the variable's type is a class
and we gathered constraints about its fields (typically when selector is a tuple)
- flatten a :: b :: ... :: Nil to List(a, b, ...)
- don't print tuple constructor names, so instead of "Tuple2(a, b)", say "(a, b)"
- filter out more statically impossible subtypes
the static types convey more information than is actually checkable at run time
this is good, as it allows us to narrow down the subtypes of a sealed type,
however, when modeling the corresponding run-time checks we need to "erase" the
uncheckable parts (using existentials so that the types are still well-kinded),
so that we can use static subtyping as a sound model for dynamic type checks
- experimental java enum handling seals enum class
before, we created a refinement class as the placeholder for the sealed children
it seems more direct to use the enum class for that
this way, the new pattern matcher's exhaustiveness checker just works for java enums
|
| |\ \ \
| | | |
| | | | |
Closes SI-5735. Review by @adriaanm.
|
| | | | |
| | | |
| | | |
| | | | |
was experiencing in IDE but could not reproduce it.
|
| |\ \ \ \
| | | | |
| | | | | |
Don't hop to the first enclosing, non-silent context typing refinements
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Closes SI-5305.
This reverts a few lines of e5cfe47a19, which was a remedy for SI-3614 and SI-3856*. I added a test case for the former, the
latter was already tested. Both tests pass after this change, and they do so with the old and new pattern matcher.
But does this change still "avoid trees with null types in presentation compiler"? What was the intent of the context hopping in the first place?
* Based on this comment: https://issues.scala-lang.org/browse/SI-3614?focusedCommentId=50477&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-50477, which elaborates further than the commit comment of e5cfe47a19.
|
| | |_|/ /
|/| | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Yin and yang would be pleased: A fix in two parts.
1. Use the name of the imported symbol, rather than the alias, in the generated `Select(qual, name)` tree.
2. Do the opposite in `isQualifyingImplicit`, which performs one part of the shadowing check.
But there is still work to do. The second part of the shadowing check, `nonImplicitSynonymInScope`,
fails to notice this case (irrespective of aliased imports).
// Expecting shadowing #2. Alas, none is cast!
object Test1 {
object A { implicit val x: Int = 1 }
import A.x
def x: Int = 0
implicitly[Int]
}
I'm hitching the residual problem to SI-4270's wagon.
|
| | |_|/
|/| |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Added a clarification to DoubleDefError for when the previous symbol
was in the package object but current symbol is not.
This was actually supposed to be an opportunity to hack partest to
run the two-step failing compilation, but somebody beat me to it and my
rebase failed.
The next hacking opportunity might be to add .pt script files! The
possibilities are endless.
|
| |\ \ \
| | | |
| | | | |
fix SI-5044: better error message on cyclic error and named/default args
|
| | | | | |
|
| |\ \ \ \
| | | | |
| | | | | |
Fix SI-2488: allow named arg, in original position, before positionals
|
| | | |_|/
| |/| | |
|
| |\ \ \ \
| | | | |
| | | | | |
Fix SI-5801: error messages regression.
|
| | | | | | |
|
| | |/ / /
|/| | |
| | | |
| | | | |
Type-check annotations in a context with a localDummy owner
|
| |\ \ \ \
| | | | |
| | | | | |
Fix SI-4928
|
| | | |/ /
| |/| |
| | | |
| | | | |
better error message when a parameter is first defined positionally, then with a named argument.
|
| |/ / /
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Since ae5ff662, resetAttrs duplicates trees, which doesn't preserve ApplyConstructor.
My attempt to modify TreeCopier to do so proved trickier than expected.
In any case, ApplyConstructor is not long for this world, and is only used in tree printing
to distinguish `new X` from regular Apply trees, so this should suffice pending full surgery.
|
| | | |
| | |
| | |
| | | |
And undeprecated Positional.
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | | |
Prevents the wheels falling off during later compiler phases,
or, worse, during bytecode verification.
Closes SI-4098.
|
| |/ /
| |
| |
| | |
Catching typer errors in specialization.
|
| | | | |
| | \ | |
| |\ \ \
| | |/
| |/|
| | | |
'refs/pull/488/head'; commit 'refs/pull/489/head'; commit 'refs/pull/490/head' into develop
|
| | |/ |
|
| | | | |
| | \ | |
| | \ | |
| | \ | |
| |\ \ \ \
| | | |/
| | |/|
| | | | |
'refs/pull/481/head'; commit 'refs/pull/482/head'; commit 'refs/pull/483/head'; commit 'refs/pull/484/head'; commit 'refs/pull/485/head' into develop
|
| | |/ /
| | |
| | |
| | | |
What's the opposite of regression? Progression!
|
| |/ /
| |
| |
| |
| |
| |
| |
| | |
They exercise both joint and separate compilation.
This resolves SI-750 (which was somewhat unfairly merged with another ticket).
The error message itself could do with refinement: in particular instance of the much beloved: found X, required X.
|
| | | |
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This change should be transparent to anything using sourceFile,
unless it was drinking from the inheritance well too deeply. Rather
than squander the already allocated field for every ClassSymbol not
being compiled from source, I'm now populating it with the file
representing the class. This will make a broad range of things easier,
like debugging, issuing useful error messages, symbol invalidation,
signature verification, you name it.
def sourceFile - still returns only source code files
def binaryFile - returns only class files
def associatedFile - returns whatever is there, if anything
Performance: I may be mistaken, but I believe this is a zero-impact
change. No new fields are allocated; fields which were null now hold
a useful reference. The reference is to a file instance which was
already being allocated and already long-lived.
Compare error messages:
// Version 1
% scalac a.scala
error: type _$1 is defined twice
// Version 2
% scalac a.scala
error: type _$1 is defined twice
conflicting symbols both originated in file './foo/package.class'
Note: this may be due to a bug in the compiler involving wildcards in package objects
one error found
Bonus for people who read commit logs. Try this in the repl
after starting power mode.
ListClass.info.members groupBy (_.associatedFile) foreach {
case (k, vs) => println("%s\n %s\n".format(k, vs map (_.defString) mkString "\n "))
}
|
| |/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Closes SI-963, since it was one of my random 30 it won the prize.
The trick after adding the stability check (which has been sitting
there commented out for 3+ years) was that implicit search depended
on the wrongness, because memberWildcardType would create scopes
with members of the form
?{ val name: tp }
And since a def shouldn't match that, fixing it broke everything
until I flipped it around: memberWildcardType should be seeking
?{ def name: tp }
It could also search for a mutable value: the relevant quality
is that it not be stable so it doesn't have a tighter type than
the members it hopes to match.
|
| | |
|
| |
|
|
|
|
|
|
| |
Most are pattern matcher bugs fixed by virtpatmat. A few are reifier,
package object, or miscellaneous. I threw in an original test for
SI-2337, to go with those for SI-1697, SI-3705, SI-4415, and SI-1357,
all of which (in the interests of making sure this basket has all the
eggs) I am closing.
|
| |
|
|
| |
Closes SI-5728.
|
| |
|
|
|
|
| |
Eventually "-Xlint would have told you not to do that" will
be a catchphrase, like "I love it when a plan comes together"
or "respect mah authoritah."
|
| |
|
|
|
|
|
|
|
| |
to improve Array[T] java-interop with T[],
normalize Object with Object{} to Object
fix #SI-5688 by flattening refined types in parents
updated check files to reflect flattening of refined types and updated position for refined types
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
sort field accessors, necessary after typers -- apparently...
don't throw TypeError, use issueTypeError
don't run patmat phase when -Xoldpatmat
only virtualize matches when -Xexperimental
recycle cps type of match for re-typechecking:
when one of the internal cps-type-state annotations is present, strip all CPS annotations
a cps-type-state-annotated type makes no sense as an expected type (matchX.tpe is used as pt in translateMatch)
don't synth FunctionN impls during typer, only do this for PartialFunction
updated check now function synth for match is deferred until uncurry
patmat-transform try/catch with match in cps
cleanup in selective anf
remove TODO: can there be cases that are not CaseDefs -- nope
|
| |
|
|
|
|
| |
Closes SI-5666 again.
(I don't have the least clue how I'm supposed to issue the error.)
|
| |
|
|
|
|
|
|
| |
ResetAttrs shouldn't be side-effecting on the original tree,
since it can lead to NPEs in erroneous trees (or maybe even for valid ones?).
Review by @odersky
(Patch by plocinic, applied without his complicity by extempore)
|
| |
|
|
|
|
|
| |
All those wildcards in a default-scoped implicit are expensive,
they each lead to a typevar on every search. Restructured the
Tuple2/Tuple3 Zipped classes, they're better this way anyway.
This also gets all that Tuple[23] code out of genprod.
|
| |
|
|
|
| |
The ones which remain I'm not removing on purpose, as I know
from experience it's more trouble than it's yet worth.
|
| |
|
|
|
|
| |
This reverts commit 1d0372f84f9a7325a47beb55169cc454895ef74b.
I forgot about polymorphic dispatch. Have to seek another way.
|
| |\ |
|
