| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
| |
|
|
|
|
|
| |
Properly specialize superclasses of a specialized class.
|
|
|
|
|
|
|
| |
Kicked the partest timeouts way up since apparently even 6 hours isn't
enough for windows to complete. Don't know what the deal is there, let's
see if 10 hours is enough. No review.
|
|
|
|
|
|
| |
Accumulate missing abstract member errors so they can all be printed
instead of only the first. Closes #2213, no review.
|
|
|
|
|
|
|
| |
Removed AnnotationInfos.ScalaSigBytes from reflect.generic, as it is an
implementation detail of the compiler and shouldn't be part of a public
API. Review by odersky.
|
| |
|
|
|
|
|
|
|
| |
Deprecated two members of RemoteActor which clearly should not be in the
public API. Removed unnecessary casts. Some more complete doc comments.
No review.
|
|
|
|
|
|
| |
[scaladoc] Closes #3302 (case class with private constructor). No
review.
|
|
|
|
|
|
| |
Closes #3310 (very large Scala class is compiled to invalid classfile
because Scala signature can't fit into constant pool). Review by dragos.
|
|
|
|
|
|
|
| |
to properly specialize on primitive types (scala.Int, instead of
scala.runtime.Int). Now closures should be indeed specialized. No
review.
|
|
|
|
|
|
|
| |
Took advantage of the sophisticated abstractions available in ant to cut
and paste the entirety of r21723 a little further down the build file.
No review.
|
|
|
|
|
|
| |
Removes pointless reference queue from the Symbol cache. Closes #3370,
no review.
|
|
|
|
|
|
|
| |
Reverted change that made partest fail with 'bin/javac: no such file'
when JAVA_HOME was not set (instead of using the javac on the PATH).
Reviewed by extempore, so no review.
|
|
|
|
|
|
|
| |
Fixed misaligning due to wide (double|long) parameters in the icode
reader. Improved constant folding through local variables. Closes #3191,
no review.
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
| |
Made internal helper methods in concurrent.ops private. Removed
protected tryCatch helper method from concurrent.TaskRunner. Review by
rompf.
|
|
|
|
|
|
| |
Removes scala.concurrent.AsyncInvokable, which is superseded by
scala.actors.CanReply. No review.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
| |
Fixed bug in Iterator.iterate which would lead to a runtime exception
under some circumstances due to inadequate laziness in calculating the
next element. No review.
|
|
|
|
|
|
|
|
| |
Some path-dependent type fiddling so power mode Trees don't come
back typed _5.compiler.Tree forSome { val _5: scala.tools.nsc.Interpreter }
or similar. No review.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Added classes/continuations-plugin to the plugin build classpath.
Explanation: as things stood the plugin step of the overall build fails
constantly even when no files have been touched. The reason is that if
one checks out an older version of the repository and then returns to
the master, a subset of continuation source files will have more recent
modification dates than their corresponding classes, and ant sees that
as a reason to rebuild them. But without continuations-plugin on the
classpath, it cannot see the classfiles of those which were not changed.
IOW, if a project has A.scala B.scala C.scala and tries to rebuild
only A.scala and C.scala, B.class must be somewhere it can be seen. We
can resolve this differently if desired but this is how the rest of
the compiler does it. (Try removing quick from quick's classpath and
rebuilding after some changes.)
Review by rompf.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
| |
Fixed scaladoc output for several types and members. No review.
|
|
|
|
|
| |
A small error message improvement suggested at #3092. No review.
|
|
|
|
|
|
| |
Added size hints to builders where possible without introducing new
methods. Closes #3331, review by community.
|
| |
|
|
|
|
|
|
| |
Various refinements and polishing to do with method signature
completion. No review.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
Created TypeDiagnostics trait and have begun opportunistically moving
code into it. Along the way, some improvements to error messages.
The situation described in ticket #2206 has always had an applicable
error message, but it wasn't making it out to the user. More kinds of
ambiguity are disambiguated, see the test cases. And overload errors are
printed with some formatting so one has some hope of parsing. Review by
odersky.
|
|
|
|
|
|
| |
Added test for several tickets that were fixed earlier this week. no
review.
|
|
|
|
|
|
| |
Fixed construction of specialized classes in the presence of
side-effects and non-trivial initializers. Review by odersky.
|
|
|
|
|
|
| |
Fixed crash when calling super in a closure (see #3312). Closes #3325.
No review.
|
|
|
|
|
|
|
|
|
|
| |
fixed #3349 : method symbol cached too aggressively
cooking raw types changes a symbol's info, but the change was masked by
caching in MethodSymbol
review by odersky
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Mostly finishing the ball that got rolling in r21679.
scala> String.cop<tab>
scala> String.copyValueOf<tab>
def copyValueOf(Array[Char]): String
def copyValueOf(Array[Char], Int, Int): String
scala> Nil.mkString<tab><tab>
def mkString(sep: String): String
def mkString(start: String, sep: String, end: String): String
def mkString: String
Lines which are not simply delimited don't work yet, so don't go
expecting List(1).<tab> to do the right thing. Yet. No review.
|
| |
|
|
|
|
|
|
| |
[scaladoc] Removed commented-out items in the css, which should have
just been removed in the first place. No review.
|
|
|
|
|
|
| |
[scaladoc] Moved object/template icons to the left of the link in the
index. Review by dubochet.
|
|
|
|
|
|
| |
[scaladoc] Added TODO comments for later fixing issue that sourceless
templates are not documented. No review.
|
| |
|