| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
During large compilations runs, the large numbers of globally unique
fresh names for existentials captured from prefixes of `asSeenFrom`.
is a) somewhat wasteful (all these names are interned in the name table)
, and, b) form a pathological case for the current implementation of
`Names#hashValue`, which leads to overfull hash-buckets in the name table.
`hashValue` should probably be improved, but my attempts to do so have
shown a small performance degradation in some benchmarks. So this commit
starts by being more frugal with these names, only uniquely naming
within an `asSeenFrom` operation.
References scala/scala-dev#246
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
They remain ValDefs until then.
- remove lazy accessor logic
now that we have a single ValDef for lazy vals,
with the underlying machinery being hidden until the fields phase
leave a `@deprecated def lazyAccessor` for scala-refactoring
- don't skolemize in purely synthetic getters,
but *do* skolemize in lazy accessor during typers
Lazy accessors have arbitrary user code, so have to skolemize.
We exempt the purely synthetic accessors (`isSyntheticAccessor`)
for strict vals, and lazy accessors emitted by the fields phase
to avoid spurious type mismatches due to issues with existentials
(That bug is tracked as https://github.com/scala/scala-dev/issues/165)
When we're past typer, lazy accessors are synthetic,
but before they are user-defined to make this hack less hacky,
we could rework our flag usage to allow for
requiring both the ACCESSOR and the SYNTHETIC bits
to identify synthetic accessors and trigger the exemption.
see also https://github.com/scala/scala-dev/issues/165
ok 7 - pos/existentials-harmful.scala
ok 8 - pos/t2435.scala
ok 9 - pos/existentials.scala
previous attempt: skolemize type of val inside the private[this] val
because its type is only observed from inside the
accessor methods (inside the method scope its existentials are skolemized)
- bean accessors have regular method types, not nullary method types
- must re-infer type for param accessor
some weirdness with scoping of param accessor vals and defs?
- tailcalls detect lazy vals, which are defdefs after fields
- can inline constant lazy val from trait
- don't mix in fields etc for an overridden lazy val
- need try-lift in lazy vals: the assign is not seen in uncurry
because fields does the transform (see run/t2333.scala)
- ensure field members end up final in bytecode
- implicit class companion method: annot filter in completer
- update check: previous error message was tangled up with unrelated
field definitions (`var s` and `val s_scope`),
now it behaves consistently whether those are val/vars or defs
- analyzer plugin check update seems benign, but no way to know...
- error message gen: there is no underlying symbol for a deferred var
look for missing getter/setter instead
- avoid retypechecking valdefs while duplicating for specialize
see pos/spec-private
- Scaladoc uniformly looks to field/accessor symbol
- test updates to innerClassAttribute by Lukas
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This commit does not close SI-5900. It only addresses a regression
in 2.11 prereleases caused by SI-7886.
The fix for SI-7886 was incomplete (as shown by the last commit)
and incorrect (as shown by the regression in pos/t5900a.scala and
the fact it ended up inferring type parameters.)
I believe that the key to fixing this problem will be unifying
the inference of case class constructor patterns and extractor
patterns.
I've explored that idea:
https://gist.github.com/retronym/7704153
https://github.com/retronym/scala/compare/ticket/5900
But didn't quite get there.
|
|
|
|
|
|
|
| |
Rather than just the first.
For example, `foo(wizzle, wuzzle, woggle)` should report all three
not-found symbols.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
I tracked down what was behind the issue described here:
// TODO: fix the illegal type bound in pos/t602 -- type inference
// messes up before we get here:
/*override def equals(x$1: Any): Boolean = ...
// Span[Any] --> Any is not a legal type argument for Span!
val o5: Option[com.mosol.sl.Span[Any]] =
*/
...which led straight to the unsoundness seen in neg/t7886.
It is dangerous to have an expected type of "Any" because
the type system will blithely ignore kind errors, since "Any"
can absorb anything. The consequence in this instance was
that inferring the case constructor for a type like
Foo[T <: Bound]
if done with expected type Any, this would come up with Foo[Any].
I altered it to use expected type Foo[T], which lets the dummy
type parameter survive to carry the bound forward and restores
sense to the inference. The before/after output for -Xprint:patmat
on pos/t602.scala is:
15c15
< if (x1.isInstanceOf[com.mosol.sl.Span[Any]])
---
> if (x1.isInstanceOf[com.mosol.sl.Span[K]])
17c17
< <synthetic> val x2: com.mosol.sl.Span[Any] = \
(x1.asInstanceOf[com.mosol.sl.Span[Any]]: com.mosol.sl.Span[Any]);
---
> <synthetic> val x2: com.mosol.sl.Span[K] = \
(x1.asInstanceOf[com.mosol.sl.Span[K]]: com.mosol.sl.Span[K]);
19,20c19,20
< val low$0: Option[Any] = x2.low;
< val high$0: Option[Any] = x2.high;
---
> val low$0: Option[K] = x2.low;
> val high$0: Option[K] = x2.high;
A file in the library depended (needlessly) on the unsoundness.
It was easy to fix but reminds us this may affect existing code.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- when typing (and naming) a ValDef, tpt and rhs are now type checked
in the same context (the inner / ValDef context). this does not change
any behavior, but is more uniform (same as for DefDef). martin told me
(offline) that this change is desirable if it doesn't break anything.
(it doesn't).
- typeSig is now more uniform with a separate method for each case
(methodSig, valDefSig, etc). methodSig was cleaned up (no more variables)
and documented. the type returned by methodSig no longer contains /
refers to type skolems, but to the actual type parameters (so we don't
need to replace the skolems lateron).
- documentation on constructor contexts, type skolems
- more tests for SI-5543
|
|
Have to intercept trees which have a null type due to errors
before they leave the warm confines of 'def typed' because from
that point everything assumes tree.tpe != null.
|