| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
|
|
|
|
|
|
|
| |
closes #1693: improve implicit conversion disambiguation by
incorporating the expected type of the member that triggered the
conversion
back-ported fix by Martin in embeddings branch.
review by odersky (just in case it shouldn't have been back-ported)
|
|
|
|
|
|
| |
Enabled partest to run entire directories for scalacheck test group.
Review by extempore.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
given the definitions below (for full context, see test file), before,
O0 would work but O would not, now both are accepted:
{{{
object O0 extends C {}
object O extends C { self => }
}}}
review by odersky
|
|
|
|
|
|
|
|
| |
closes #3692: make instantiateTypeVar more careful so it does not change
T's info to >: T <: T.
review by odersky
|
|
|
|
|
|
|
|
| |
patch contributed by Mark Harrah in
http://article.gmane.org/gmane.comp.lang.scala/20700
reviewed by moors and odersky
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Couldn't bear to see sbt seemingly so close but still unable to run
tests. Made some progress but bogged down in scalacheck when I realized
things were farther from working than I'd thought. Please forgive this
brief expression of frustration, because the despair is palpable:
the weeks of work I did on partest was apparently of no interest,
and now I'm in here having to fix the same partest bugs again, in a
more virulent form. They've evolved into some kind of ant-partest-sbt
artificial creature, more fearsome than any one of them could be under
its own power.
Did we not know we have a whole directory full of pretty code like this?
List(
"Scala compiler classes in: " + testBuildDir,
"Scala version is: " + nsc.Properties.versionMsg,
"Scalac options are: " + universe.scalacOpts,
"Java binaries in: " + javaBin,
"Java runtime is: " + javaInfoString,
"Java runtime options: " + (Process.javaVmArguments mkString " "),
"Javac options are: " + universe.javacOpts,
"Java options are: " + universe.javaOpts,
"Source directory is: " + src,
"Selected categories: " + (selectedCategories mkString " "),
""
) mkString "\n"
It's src/partest-alternative in case anyone wants to consider any of it.
At the very least there are individual files or functions which could be
cherry picked.
-- This patch contains: --
Readability efforts. Twelve parameter functions whose last six
parameters are "true,true,failedOnly,true,false,log" are pretty
challenging to work with.
Changes to the environment variables sbt was setting to look a bit more
like what partest expects and will crash without.
Sets partest.debug=true if the sbt logger level is at Debug.
On my machine all test groups except scalacheck appear to run to
completion.
Review by moix. (Hello, moix! I am available for assistance if you are
in the market for it. You have brought us long suffering ant-ers to the
brink of sbt goodness: and now, we enter the fray!)
|
|
|
|
|
|
| |
Tweaking sbt so it will stop jabbering about svn info failing when
there's a perfectly fine git repo yet to be tried. No review.
|
| |
|
|
|
|
|
|
|
|
|
| |
Interpreted absence of any objections in mailing list thread
http://www.scala-lang.org/node/7398 as implicit encouragement to
proceed. Adds complementary implicits to Predef such that the eight
primitive types are both boxed and unboxed as the occasion demands.
Review by odersky.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
| |
Proposed implementation of 'unlift' on Function1, the inverse function
of PartialFunction#lift. Review by rytz and other interested parties.
References #3825, but not closing until this is further considered.
|
|
|
|
|
|
| |
Unbroke the command line ~/test/partest tool, which was displeased at
being unable to find scalacheck.jar. No review.
|
| |
|
| |
|
|
|
|
|
|
| |
fix fix for (close #3791) and made deprecation for constants work. no
review.
|
| |
|
|
|
|
|
|
|
| |
long cycle (deprecate, remove, add with new behavior). I tried adding
Scaladoc for these members but failed for now, see my message in
internals.
|
|
|
|
|
|
| |
Some modifications to repl classloading to make it usable in managed
classloader environments. Contributed by mark harrah. Review by rytz.
|
| |
|
| |
|
|
|
|
|
|
| |
Fixes infinite loop when appending a ListBuffer to itself. Review by
extempore.
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
| |
Added type alias so that Response is accessible via interactive.Global
... necessary for source compatibility with 2.8.0.final. Review by
odersky.
|
|
|
|
|
|
|
|
| |
Brought the files in scala.collection.interfaces up to date,
for whatever purpose they may someday serve. Also deprecated
Seq.findIndexOf, which is identical to indexWhere and must have been
overlooked when deprecations were being handed out. No review.
|
| |
|
|
|
|
|
|
| |
Closes #3636, crash when unexpected types arrive at specialization. no
review.
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
| |
Makes utility methods in the ArrayStack companion object private.
No review.
|
|
|
|
|
|
|
|
|
|
|
| |
actually incorrect, as trunk has significantly
diverged from 2.8.x, but it's better than
leaving the version number in trunk at "2.8.0".
We should just rename the version number in
trunk at some point, so that it just lists
"trunk" rather than "2.x.x".
|
|
|
|
|
|
|
| |
Partest changes - partest now uses scalacheck.jar from the build/pack dir.
Review by phaller.
|
| |
|
|
|
|
|
|
|
| |
Temporarily disabled collections checks to keep build from breaking.
No review
|
| |
|
|
|
|
|
|
|
| |
[scaladoc] Merges Petr's changes from the colladoc branch to only
display less-important information in class comments on demand. Also
fixes a bug with index filter. No review.
|
| |
|
|
|
|
|
| |
Review by phaller (not urgent).
|
|
|
|
|
| |
for MSIL: now based on the latest ch.epfl.lamp.compiler.msil sources.
|
|
|
|
|
| |
No review yet.
|
| |
|