| Commit message (Collapse) | Author | Age | Files | Lines |
|\
| |
| | |
SI-7499 Additional test case for SI-7319
|
| |
| |
| |
| | |
From a duplicate ticket.
|
|\ \
| | |
| | | |
SI-3425 erasure crash with refinement members.
|
| | |
| | |
| | |
| | | |
Probably overshot the mark a little.
|
| |/
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Checking that a refinement class symbol does not override
any symbols does mean it will have to be invoke reflectively;
but the converse is not true. It can override other symbols
and still have to be called reflectively, because the
overridden symbols may also be defined in refinement classes.
scala> class Foo { type R1 <: { def x: Any } ; type R2 <: R1 { def x: Int } }
defined class Foo
scala> typeOf[Foo].member(TypeName("R2")).info.member("x": TermName).overrideChain
res1: List[$r.intp.global.Symbol] = List(method x, method x)
scala> res1 filterNot (_.owner.isRefinementClass)
res2: List[$r.intp.global.Symbol] = List()
And checking that "owner.info decl name == this" only works if
name is not overloaded.
So the logic is all in "isOnlyRefinementMember" now, and
let's hope that suffices for a while.
|
|\ \
| | |
| | | |
SI-6811 Deprecate scala.util.parsing.json
|
| | | |
|
|\ \ \
| |/ /
|/| | |
SI-7436 Varargs awareness for super param aliasing.
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Don't consider a super class parameter accessor to be
an alias if it is a repeated. Parameter aliases are used
to avoid retaining redundant fields in the subclass; but
that optimization is out of the question here.
|
|\ \ \
| |_|/
|/| | |
Moves AbstractFileClassLoader to scala-reflect.jar
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Its string name was used in ReflectionUtils and became broken after repl
got factored out. This hints that that classloader belongs to where it is
used, i.e. to scala-reflect.jar.
Moreover AbstractFile is defined in scala-reflect.jar, so it's only
logical to also define a derived classloader in scala-reflect.jar.
|
|\ \ \
| | | |
| | | | |
sanitation of Infer.
|
| | | |
| | | |
| | | |
| | | |
| | | | |
Oh exprTypeArgs, you will send us to the poorhouse with
your spendy ways. So many tuples have been dying needlessly.
|
| | | | |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Only to discover that it's really hard to move isCoercible
anywhere because it wants to call inferView which, despite its
suggestive name, is not visible in Infer. So I did what I
could and documented it a little.
|
| | | |
| | | |
| | | |
| | | | |
He's a good kid, even if not the most hygenic.
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Yes checkAccessible, you patiently waited your turn.
Do you like your new clothes? Oh, still not clean enough?
I understand, but we must leave time for the other methods.
Now run along and check something.
|
| | | |
| | | |
| | | |
| | | | |
To streamline the essential logic.
|
| | | |
| | | |
| | | |
| | | | |
It has been hardcoded to return false for a long while anyway.
|
| | | |
| | | |
| | | |
| | | | |
A lot more useful than Unit.
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Because nobody should have to guess a return type like
(List[Type], Array[Int], Boolean). In fact nobody should
ever experience such a return type at all.
|
| | | | |
|
| | | |
| | | |
| | | |
| | | | |
It's not exactly 2006-era code but it's still a little ripe.
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Yes, 2006. Perhaps we should have a "sunset provision"
which removes commented-out code after seven years. Or even
slightly sooner.
|
| | | |
| | | |
| | | |
| | | | |
Now you can take inferPolyAlternatives home to meet your mother.
|
| | | |
| | | |
| | | |
| | | | |
So much dust and debris, so little time.
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Swabbing the decks of isAsSpecific and isApplicableSafe.
Wow isApplicableSafe, you don't really need to declare and
allocate a two-argument case class in order to manage a
second method call.
|
|/ / /
| | |
| | |
| | |
| | | |
I'd really like these methods to receive greater scrutiny.
To that end I'm scrutinizing them.
|
|\ \ \
| | | |
| | | | |
Toward more organized typer state.
|
| |\ \ \
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
* origin/master:
An attempt to make tests deterministic.
SI-7427 stop crashing under -Ydebug.
|
| |\ \ \ \
| | |_|/ /
| |/| | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
* commit 'd93826f278':
Incorporated reviewer feedback.
Incorporated reviewer feedback.
Refactored stabilize.
Eliminated HKmode.
Eliminated RETmode.
Eliminated SNDTRYmode.
Started eliminating modes.
Corralling Modes into a smaller pen.
Conflicts:
src/compiler/scala/tools/nsc/typechecker/Macros.scala
src/compiler/scala/tools/nsc/typechecker/RefChecks.scala
src/compiler/scala/tools/nsc/typechecker/Typers.scala
|
| | | | |
| | | | |
| | | | |
| | | | | |
Made things a little more consistent and self-apparent.
|
| | |\ \ \
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
* master:
SI-7469 Remove @deprecated scala.util.logging
SI-3943 Test case for already-fixed Java interop bug
Fix formatting for couple of docs in the compiler
SI-7476 Add documentation to GenericTraversableTemplate
SI-7469 Remove @deprecated scala.util.parsing.ast
SI-7469 Remove @deprecated MurmurHash elements
SI-7469 Remove deprecated elements in s.u.parsing.combinator
SI-7469 Make @deprecated elems in scala.concurrent private[scala]
removes duplication in inferImplicitValue
SI-7047 fixes silent for c.inferImplicitXXX
SI-7167 implicit macros decide what is divergence
macroExpandAll is now triggered in all invocations of typed
SI-5923 instantiates targs in deferred macro applications
SI-6406 Restore deprecated API
SI-6039 Harden against irrelevant filesystem details
Limit unnecessary calls to Type#toString.
fix typo in comment
SI-7432 add testcases
SI-7432 Range.min should throw NoSuchElementException on empty range
AbstractFile.getDirectory does not return null when outDir is "."
Conflicts:
src/compiler/scala/tools/nsc/typechecker/Typers.scala
|
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Made things a little more consistent and self-apparent.
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Mode elimination ramps up in difficulty now, so I pursued
other forms of code hygiene.
|
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
It becomes context mode "TypeConstructor".
|
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
It becomes context mode "ReturnExpr".
|
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
It becomes context mode "SecondTry".
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Consolidating the scattered typer state in Context, where it's
relatively easy to keep an eye on, rather than threaded throughout
the typer in sneaky/sticky bitmasks. The general pattern will be
what was once an explicitly passed around bit in Mode becomes
an implicitly propagated-as-appropriate bit in Context.
In this commit:
ALTmode becomes context mode "PatternAlternative"
STARmode becomes context mode "StarPatterns"
SUPERCONSTRmode becomes context mode "SuperInit"
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Attempting to reduce the frequency of low-level operations
with modes. I mean stuff like this:
if ((mode & (EXPRmode | LHSmode)) == EXPRmode)
THey don't make those ten line boolean guards any easier
to understand. Hopefully this will lead us toward eliminating
some of the modes entirely, or at least better isolating
their logic rather than having it interspersed at arbitrary
points throughout the typer.
Modes are in their entirety a leaked implementation detail.
Typing a tree requires a tree and optionally an expected type.
It shouldn't require a bucket of state bits. In subsequent
commits I will start eliminating them.
This commit also breaks adapt down into more digestible chunks.
|
|\ \ \ \ \ \
| | | | | | |
| | | | | | | |
Harden partest against duplicate paths.
|
|/ / / / / /
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
A serious issue: partest would launch the same test multiple
times if a test path was given in multiple forms (e.g. absolute
and relative paths.) Unfortunately all those tests would share
the same logfile, output directory, etc. which would predictably
lead to explosions.
Since overwriting classfiles while being loaded can lead to
jvm core dumps, it's possible this is involved in recent jvm
crashes and other test breakdowns.
This commit also alters the default partest verbosity to only
full print test transcripts under --verbose.
|
|\ \ \ \ \ \
| |_|_|_|/ /
|/| | | | | |
SI-7494 Each plugin must only be instantiated once.
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Exploit SI-6446 fix to move old pending tests for plugins to the
active pool.
Also, nuance the partest runner to allow the compiler to crash
with a FatalError if a check file is present.
Some of the plugin tests make phase assembly throw.
One of the legacy tests was moved back to pending:
There seems to be a race for who shall be deemed truly dependent.
Back to pending for you!
apm@mara:~/projects/snytt/test$ ./partest files/neg/t7494-cyclic-dependency/
Selected 1 tests drawn from specified tests
> starting 1 test in neg
ok 1 - neg/t7494-cyclic-dependency
1/1 passed (elapsed time: 00:00:04)
Test Run PASSED
apm@mara:~/projects/snytt/test$ ./partest files/neg/t7494-cyclic-dependency/
Selected 1 tests drawn from specified tests
> starting 1 test in neg
!! 1 - neg/t7494-cyclic-dependency [output differs]
> 0/1 passed, 1 failed in neg
>>>>> Transcripts from failed tests >>>>>
> partest files/neg/t7494-cyclic-dependency
% scalac t7494-cyclic-dependency/ThePlugin.scala
% scalac t7494-cyclic-dependency/sample_2.scala
error: Cycle in compiler phase dependencies detected, phase cyclicdependency1 reacted twice!
% diff files/neg/t7494-cyclic-dependency-neg.log files/neg/t7494-cyclic-dependency.check
@@ -1 +1 @@
-error: Cycle in compiler phase dependencies detected, phase cyclicdependency1 reacted twice!
+error: Cycle in compiler phase dependencies detected, phase cyclicdependency2 reacted twice!
0/1 passed, 1 failed (elapsed time: 00:00:04)
Test Run FAILED
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
The old behavior was that Plugin.loadAllFrom took a distinct
list of plugin locations.
The broken behavior incorrectly punted to the pruning
algorithm in Plugins.loadPlugins$pick.
The new behavior is to distinctify the list of class names
to load.
This is more robust; under the old scheme, specifying the
continuations.jar under two file names would fail as
described in the ticket.
Pruning a plugin by disabling (-Xplugin-disable) or because
it steps on someone else's phase name suffers the same
defect. There is no way to tell the plugin it has been
benched. And you know that global I gave you? Can I
have it back?
|
|\ \ \ \ \ \
| | | | | | |
| | | | | | | |
An attempt to make tests deterministic.
|
|/ / / / / /
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
The nondeterminism presently showing itself in
presentation/implicit-member is a consequence of the
presentation compiler tests relying on details of the
behavior of toString calls. We need to stomp this out,
but it will take a while. Based on the check file
changes enclosed with this commit, this will suffice
for the presentation compiler tests. A broader assault
will have to take place, but not yet.
|
|\ \ \ \ \ \
| |_|/ / / /
|/| | | | | |
SI-7427 stop crashing under -Ydebug.
|
|/ / / / /
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Change "dumpClassesAndAbort" to "devWarning". You can witness
it happen like so.
% scalac test/files/pos/t7427.scala -Ydebug -Xdev
...
[running phase cleanup on t7427.scala]
[running phase icode on t7427.scala]
warning: !!! PJUMP(method matchEnd4)/scala.tools.nsc.backend.icode.GenICode$PJUMP is not a control flow instruction
warning: !!! PJUMP(method case6)/scala.tools.nsc.backend.icode.GenICode$PJUMP is not a control flow instruction
[running phase inliner on t7427.scala]
[running phase inlinehandlers on t7427.scala]
Having now lived with this for months, I have no ambition to
unravel the actual problem, I just want it to stop crashing.
|
|\ \ \ \ \
| | | | | |
| | | | | | |
Merge 2.10.x
|