| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
| |
Test a wider range of functionality.
|
| |
|
|
|
| |
No need to check the output - checking programmatically that the
produced streams are empty is enough.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The included test currently fails because `map` and `flatMap` do not
run in constant stack space on a stream returned by `Stream.withFilter`,
as I reported here:
https://groups.google.com/d/msg/scala-language/WqJR38REXnk/saaSiDdmyqoJ
Fix the problem and add a simple testcase.
Note that the stack space consumed when producing an element of this stream is
proportional to the number of elements failing the test before the next
success. The stack space consumed to produce the stream itself is the
space needed to produce the first element, that is, is proportional to
the number of failures before the first success.
|
| |\
| |
| | |
even more cleanup in Macros.scala
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Error reporting is moved to ContextErrors to disentangle stuff in Macros.scala.
With logics and error reporting intertwined it was an awful mess.
Exceptions are used for the same reason. Instead of threading failures through
the code polluting it with options/ifs, I outline the success path.
It worked much better for typedMacroBody, but I'm also happy with the resulting
code of macroExpand. To me a major factor towards applicability of exceptions
was that they are short-lived and that there might be max one error per domain,
after which we unconditionally bail.
|
| |\ \
| | |
| | | |
Fix for community build blocker.
|
| | |/
| |
| |
| |
| |
| |
| |
| | |
As two character bugfixes go, this was a doozy.
I will forego elaborating at length and offer generic
instructions for elucidation:
% git log --grep=tpeHK
|
| |/
|
|
|
|
| |
Previously I thought it's fixed because I didn't include the correct
testcase. Now it's the correct testcase and this prompted me to
change the code a bit to make it work properly.
|
| |\
| |
| | |
more cleanup in Macros.scala
|
| | | |
|
| | | |
|
| | |
| |
| |
| |
| |
| |
| | |
In separate compilation runs, the static field symbol in the
companion class of an object was not being recreated.
Given that the singleton object was compiled separately,
the static field symbol will be recreated on demand.
|
| | |
| |
| |
| |
| |
| |
| |
| | |
Disable @static for the REPL code.
The problem is that there are no companion classes generated for objects
that contain the top-level declarations in the REPL.
When there is no companion class, the selecting a @static field will translate
to a getter call, instead of to a field load.
|
| |\ \
| |/
|/| |
Fix SI-6208. mutable.Queue now returns mutable.Queue for collection met...
|
| | |
| |
| |
| | |
rather than MutableList.
|
| |\ \
| | |
| | | |
Fixes backend crash due to incorrect consumedTypes
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This started out as a compiler crash after Greg copied the
comprehension methods to List and made them final.
The culprit was the dead code elimination phase, which after
sweeping pieces of code was unable to restore the stack to its
original state, thus causing the ASM backend to crash notifying
the resulting bytecode is incorrect.
The dead code elimination phase uses the icode Instructions'
consumedTypes to determine what needs to be dropped from the
stack when an instruction is eliminated, but the consumedTypes
were only defined for a handful of instructions. So dce encountered
a DUP instruction for which it did not have consumedTypes defined
and did not restore the stack correctly.
The consumedTypes/producedTypes for icode instructions are
redundantly defined in 3 separate places:
- Opcodes.scala (consumedTypes/producedTypes)
- ICodeCheckers.scala (for checking icode)
- TypeFlowAnalysis.scala (for computing types on the stack at each
program point)
Since the Opcodes types are the only ones visible outside, I suggest
we use them in ICodeCheckers.scala and TypeFlowAnalysis.scala too. But
we should make such changes after the release, as we're chilling out
by the lake with a glass of good wine: SI-6234
The relevant discussion around it can be found at:
https://groups.google.com/forum/?fromgroups#!topic/scala-internals/qcyTjk8euUI[1-25]
Many thanks to Paul for his help!
Review by @magarciaEPFL or @paulp.
|
| | | |
| | |
| | |
| | |
| | | |
We no longer have multiple canonical signatures for macro implementations,
so it was time to put the List-List-List-based logic to rest.
|
| |/ /
| |
| |
| |
| |
| | |
mostly removes [Eugene] marks that I left back then and reviews related code
some of those tokens got left in place, because I don't know to how fix them
without imposing risks on 2.10.0
|
| |\ \
| | |
| | | |
SI-5940 impls are no longer in macro def pickles
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The immediate motivator for this change was the desire to keep things simple
during macro expansion.
Here's the story. When you expand a macro, you need to reflectively load
a macro implementation method and invoke it. However private methods in Scala
can sometime have their names mangled, so the reflector has to check multiple
variants of the same name (the normal one and the mangled one).
The problem is that since the previous commit we don't have an access to a
symbol of the macro impl, hence we cannot just use `impl.expandedName` like
we did before. Sure I could duplicate the logic of expandedName, but I have
a better suggestion.
Let's prohibit non-public macro implementations. This doesn't seem to hurt,
and it lets us avoid extra bit of complexity in Macros.scala.
If this measure turns out to be a hassle during the trial period of macros,
non-public macros can always be allowed. In fact, we can even have this feature
back for free when we migrate from Java reflection to Scala reflection for
invoking macro implementations (because Scala reflection knows how to account
for mangled private names). But that's the 2.10.x business.
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The first officially released version of macros persisted
macro def -> impl bindings across compilation runs using a neat trick.
The right-hand side of macro definition (which contains a reference to an impl)
was typechecked and then put verbatim into an annotation on macro definition.
This solution is very simple, but unfortunately it's also lacking. If we use it
then signatures of macro defs become transitively dependent on scala-reflect.jar
(because they refer to macro impls, and macro impls refer to
scala.reflect.macros.Context defined in scala-reflect.jar).
More details can be found in https://issues.scala-lang.org/browse/SI-5940.
Therefore we have to avoid putting macro impls into binding pickles and
come up with our own serialization format. Situation is further complicated
by the fact that it's not enough to just pickle impl's class and method names,
because macro expansion needs knowledge about the shape of impl's signature
(which we can't pickle). Hence we precompute necessary stuff (e.g. the layout
of type parameters) when compiling macro defs.
|
| |\ \ \
| | | |
| | | | |
Move side-effecting call out of debuglog(). Fixes SI-6223.
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This change fixes a situation in which method calls aren't being properly
specialized. All the existing tests pass; in addition, it likely fixes other
specialization bugs (and improves performance of specialized code).
Also includes a small regression test.
|
| |\ \ \ \
| |/ / /
|/| | | |
SI-6022 cleaner model of variable equality modulo <:
|
| | |/ /
| | |
| | |
| | |
| | |
| | |
| | | |
a more conservative "excludes": no need to reason about types
(TODO: check we don't get any spurious unreachability errors
in the eclipse build, which is a good canary for this kind of thing)
|
| |\ \ \
| | | |
| | | | |
Fixes SI-6214.
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
addSynthetics was calling typedStat that caused flushing error buffer. Typically that would result in errors being re-added but addSynthetics was the last thing to run in typedStats.
Instead of adding yet another check I used this as an opportunity to cleanup the once necessary workaround for typing stats. Since we are now chaining buffers of contexts manual handling is no longer necessary.
Review by @odersky (I recall that you were planning to cleanup that part at some point).
|
| |\ \ \ \
| | | | |
| | | | | |
Topic/unpickle resilience 2
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
The situation (I don't know how to make partest test this) is
package s
class A ; class S { def f(): A = ??? }
If one compiles this and removes A.class, should references to
class S cause the compiler to explode eagerly and fail to load S,
or explode lazily if and when it needs to know something about A?
This patch takes us from the former strategy to the latter.
Review by @xeno-by.
|
| |\ \ \ \ \
| |/ / / /
|/| | | | |
Ticket 6184 revised
|
| | | |/ /
| |/| |
| | | |
| | | |
| | | |
| | | |
| | | | |
this should fix the crash in asSeenFrom that resulted from calling
baseTypeSeq on a type that had an unbound type parameter in it
also, simplify widenToClass
|
| |\ \ \ \
| | | | |
| | | | | |
SI-6201 minor fixes in key points
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Fixes several oversights that led to 6201. RootPackage should have been static,
refactored implementation of hasSymbolWhich shouldn't have checked hasSymbol.
Full discussion is here:
http://groups.google.com/group/scala-internals/browse_thread/thread/9500348f273a8aa.
|
| |\ \ \ \ \
| | | | | |
| | | | | | |
Only create a HashTrieSet if necessary.
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | | |
yields a HashSet1 as it should
|
| |\ \ \ \ \ \
| |_|/ / / /
|/| | | | | |
Si 6198
|
| | |/ / / /
| | | | |
| | | | |
| | | | | |
This tests that ListSet.tail is O(1) as well as testing that a HashSetCollision1 is turned back into a HashSet1 when removing all but one element
|
| |\ \ \ \ \
| | | | | |
| | | | | | |
Better pattern matcher error message.
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
For the common case when someone hasn't quite grokked
the significance of lower case in a pattern match. I'd
like to make all the unreachables errors, not warnings,
but there may be a bug or two to clear out first.
class A {
def badEquals(x: Any, y: Any) = x match {
case y => true
case _ => false
}
}
a.scala:3: warning: patterns after a variable pattern cannot match (SLS 8.1.1)
If you intended to match against parameter y of method badEquals, you must use backticks, like: case `y` =>
case y => true
^
a.scala:4: warning: unreachable code due to variable pattern 'y' on line 3
case _ => false
^
two warnings found
|
| |\ \ \ \ \ \
| | | | | | |
| | | | | | | |
fixes to existential-related reification problems
|
| | | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
When reifying certain trees/types reifier might decide that it needs to
introduce auxiliary symbols to represent non-locatable components of types
For example, when reifying a refined type we need to reify the symbols
that correspond to its members (aka RefinedType.decls). Since these symbols
are not stored in pickles, we can't do the usual staticClass(scala.Int) stuff,
and are forced to actually create corresponding symbols inside the reificode
(by "create" I mean literally create, by using newXXXSymbol methods).
It looks like a generally good idea to not only create those symbols naked,
but to also export their type signatures and annotations. However this brings
problems with type inference.
Say, a type of a method A depends on a type tag. To get a type of a type tag,
the compiler needs to expand the corresponding macro (when the macro is being
expanded, symbol A is marked as LOCKED). While reification macro expands,
it might want to reify certain symbol definitions (as explained above). If one
of these definitions is a class that contains method A, we're in trouble,
since reifying the corresponding ClassInfoType will eventually call A.info,
which will produce a cyclic reference error, because A is LOCKED.
An obvious solution is to check whether a reified symbol definition is locked.
If the symbol is locked, then the reifier bails (e.g. by reifying NoType
instead of the actual type signature of the symbol).
Being obvious this solution is also incorrect as illustrated by SI-6204.
Sure we can check whether a symbol itself is locked, but we cannot check
whether or not we reify someone who refers to the locked symbol.
As of such I'm telling the reifier to bail whenever it needs to reify a symbol
definition for a symbol that is not yet complete. Therefore reification
can no longer cause inference of method result types, which eliminates the
underlying problem.
This is a harsh measure, but taking into account that we have an RC planned
within a week, stability trumps completeness.
|
| | | |/ / / /
| |/| | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Unlike module classes that are going to be pickled (which are reified
as `<their module reification>.moduleClass), module classes of
local modules need to be reified as symbol defs.
Otherwise we get a stack overflow:
1) Local modules are deemed to be free terms,
2) All free symbols are reified together with their type signature (so that they
can be a) inspected by anyone interested, b) compiled at runtime),
3) Reifying a type signature of a module involves reifying its module class,
4) Reifying a module class involves reifying a module and calling its module class,
<ad infinitum>
This stack overflow doesn't happen for locatable modules, because they don't need
to have their type signatures reified (these signatures can later be loaded from
pickles if it becomes necessary).
|
| |\ \ \ \ \ \
| |_|_|/ / /
|/| | | | | |
Moved inline logic before pickler.
|
| | | |/ / /
| |/| | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
The test case failed due to separate compilation. The problem was
that we don't pickle the fact that the field was made public.
Original patch by @odersky. Cleaned up by me. Changes I made:
* removed stale test-case
* reduced whitespace changes
Supersedes #1089.
Review by @odersky and @moors.
|
| |\ \ \ \ \
| | | | | |
| | | | | | |
SI-5739 (bis) vals for subpatterns when -g:patmatvars
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
To facilitate debugging pattern matches, we store the values for
sub-patterns of extractor (synthetic or user-defined) patterns in local variables.
When performing an optimized build, and when possible, we don't do store but inline them directly.
For soundness, SI-5158, SI-6070, we must always store the values of mutable case class fields.
(Specifying -optimize is the only way to suppress emitting these local variables.
An unoptimized build will always generate them, which was deemed the right default during the meeting.)
(updated flags for t4425 to get consistent runs on optimized and non-optimized partest runs
by always passing -optimize)
|
| |\ \ \ \ \ \
| | | | | | |
| | | | | | | |
SI-6205 make pt fully defined before inferTypedPattern
|
| | | |_|_|/ /
| |/| | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
refines my fix for SI-2038 (#981) by making pt fully defined before calling inferTypedPattern,
instead of making the result of inferTypedPattern fully defined
I finally realized my mistake by diffing the -Ytyper-debug output of compiling the variants with:
```
x match {case Holder(k: A[kt]) => (k: A[kt])}
```
and
```
(x: Any) match {case Holder(k: A[kt]) => (k: A[kt])}
```
|
| |\ \ \ \ \ \
| |_|_|/ / /
|/| | | | | |
SI-6145 lax typing of args to synthetic case-labels
|
