| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
TermName and TypeName are exposed throughout the compiler based on what
kind of name a given abstraction ought to have. (There remain places
where one needs to create a name without knowing yet what it will be,
and those will always be Names.)
The nme object in the compiler now holds only term names. To reference a
known type name, use tpnme:
nme.List == ("List": TermName)
tpnme.List == ("List": TypeName)
The contents of nme and tpname are defined in traits, many of which
are shared, so if a name should exist only as a Type and not a
Term, it should be defined in CompilerTypeNames, but otherwise in
CompilerTermNames or CompilerCommonNames. This is partially complete but
I'm sure there are still many shared which should pick a side.
Usage of .toTermName and .toTypeName is strongly discouraged. After the
dust has settled, there will be very few places where it will make sense
to hop between namespaces like that.
There are some implicits to smooth everything out, most of which should
be removable eventually.
// these two are in no hurry to go anywhere
String => TermName
String => TypeName
// but not String => Name: def view in the compiler is no longer implicit
// these two are temporary, and can log when they kick off to help us
flush // out remaining issues of "name migration" Name => TermName
Name => TypeName
There is more work to be done before we're properly protected from
naming errors, but I will not allow another eight hour tragedy to befall
lukas or anyone else!
Review by rytz. (Formality.)
|
| |
|
|
|
|
|
|
|
|
| |
In r23627 Lukas fixed specialization bug that seemed to be the reason
for #3993 and other specialization and resident compiler bugs. So
reverting quick fix r23576 as it is not necessary anymore. No review.
Closes #3993.
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
Some cruft which finally got on my nerves enough to decruft. No more
getClass2 and getModule2 uses in Definitions, since that transition was
completed in the time of my grandpappy. (The implementations remain,
quiescent.) No review.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
A patch authored by martin to fix one of the big outstanding issues with
-Ycheck. The typer had a vision of the world which fell to pieces if
presented with a tree from after phase constructors. Now it is wiser:
perhaps a bit sadder as well, but every typer must face this day sooner
or later.
Also removed the various passing tests from test/checker-tests, but
still plenty of fascinating failures in there for anyone interested in
helping push to the finish line! No review.
|
|
|
|
|
|
|
|
|
|
| |
warnings. Fixed some bugs revealed by said warnings, and made some minor
changes to avoid warnings. (Technically it's not a bug to have unrelated
classes compare as equal, but it so often is a bug that it behooves us
not to do it intentionally so the warnings stand out.) Disabled the most
useful warning for the moment since it'd be wrong with some frequency.
No review.
|
|
|
|
|
|
| |
Taking retronym's advice on increasing UniversalFn transparency. No
review.
|
|
|
|
|
|
|
| |
Parser relaxation so that one can name the type variables in a
constructor pattern match: this is a prerequisite to realizing the full
potential of gadts. (Nothing new works here however.) No review.
|
|
|
|
|
|
|
| |
Found several minor thigns wrong with checkSensible, which tries to
issue warnings for comparisons which will always be true or false. No
review.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Deprecated the @serializable annotation, introduce a new trait
"scala.Serializable" which has to be extended instead (cross-platform).
Known issues:
- Companion objects of serializable classes (including case classes) are automatically made serializable. However, they don't extend "Serializable" statically because of the known difficulty (should be done before typing, but hard).
- Writing "case class C() extends Serializable" gives "error: trait Serializable is inherited twice"
- Functions are serializable, but don't extend Serializable dynamically (could be fixed by making FunctionN Serializable - shouldn't we?)
Note that @SerialVersionUID continues to be an annotation; it generates
a static field, which is not possible otherwise in scala.
Review by dragos, extempore.
Question to dragos: in JavaPlatform.isMaybeBoxed, why is there a test
for "JavaSerializableClass"? Is that correct?
|
|
|
|
|
|
|
| |
It looks like the fix for #2318 was causing problems, so I gave it some
more graceful failure. Leaving #2318 open because the implementation
could still work harder to find the right method. No review.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
"I invite everyone to change this rule, and observe what breaks!"
This much beloved comment from #1208 has been on my radar for two years.
The worm has turned. Closes #1208. The inferred type of an object is now
Foo.type instead of "object Foo". What once was this:
scala> val x: Map[Int, Singleton] = Map(1 -> None)
<console>:5: error: type mismatch;
found : (Int, object None)
required: (Int, Singleton)
Now exudes a pleasing aura of workingness:
scala> val x: Map[Int, Singleton] = Map(1 -> None)
x: Map[Int,Singleton] = Map(1 -> None)
No review.
|
| |
|
|
|
|
|
|
| |
Did some more polishing on the infamous partial function error message.
No review.
|
|
|
|
|
|
| |
Moved sigint handler to InterpreterLoop and made it discriminate based
on contents of current line. No review.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Fleshed out the mock code a little further so it's easy for closures to
become SAMs.
// implicit not necessary, but improves fun factor
scala> implicit def mkUFn(x: AnyRef) = scala.tools.reflect.UniversalFn(x)
mkUFn: (x: AnyRef)scala.tools.reflect.UniversalFn
scala> (() => 5*5*5).as[java.util.concurrent.Callable[Int]]
res1: java.util.concurrent.Callable[Int] = <function1>
scala> res1.call
res2: Int = 125
No review.
|
|
|
|
|
|
| |
Some hardening of the repl's internals extracted from a more interesting
patch in progress. No review.
|
|
|
|
|
|
| |
Some cleanups in the *Runners and a few compiler I/O conveniences. No
review.
|
|
|
|
|
|
|
| |
Temporarily disabled varargs annotation on constructors.
No review.
|
|
|
|
|
|
| |
Some work on error messages, somewhat based on ideas in #3092. No
review.
|
|
|
|
|
|
| |
Seeing if I can unbreak things without actually removing the visible
motivation for the whole exercise. No review.
|
|
|
|
|
|
|
| |
Making the installation of the repl sigint handler take place less
eagerly. And more relevantly to recent hangs, made the repl only create
daemon threads. No review.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
A new totally safe signal manager, along with all the support code
needed for that. See the classes in scala.tools.reflect.* for
interesting new weapons. Also includes inaugural handy usage:
scala> val x = 10
x: Int = 10
scala> while (true) ()
[ctrl-C]
Execution interrupted by signal.
scala> x
res1: Int = 10
No review, but feedback welcome.
|
| |
|
|
|
|
|
|
| |
I keep forgeting which phase the browser is currently showing, so added
phase name. No review
|
|
|
|
|
| |
No review.
|
|
|
|
|
|
|
|
| |
In the absence of known reasons why some exceptions are intercepted or
have their stack trace printed and others fly unchallenged, changed
script and object runners to handle everything consistently when an
exception is thrown. Closes #3978 again, no review.
|
|
|
|
|
|
| |
Temporarily reverted sun.misc.* using code until reflective version is
complete. No review.
|
|
|
|
|
| |
Review by Rytz.
|
|
|
|
|
|
|
|
|
| |
Changing EnclosingMethod generation not to drop the trailing $ from the
name of the container class. I'm sure this was done for some reason but
proguard objects rather strenuously (and fails) because it can't find
the method, since it says it's in the class but is really in the object.
Review by dragos.
|
|
|
|
|
|
|
| |
Added varargs annotation.
Review by Rytz.
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
Make interrupts non-interruptible, otherwise the interrupt exception can
just go direct to the user (IDE). Review by odersky.
|
| |
|
|
|
|
|
|
| |
removed duplicated error message for bad compiler options made some
cleanups in detach plugin code
|
|
|
|
|
|
|
|
|
|
| |
(1) the Scala types in ch.epfl.lamp.compiler.msil.emit now use Scala
collections instead of Java's, (2) a few bug fixes regarding metadata
parsing, (3) GenMSIL emits output useful for IDE debugging, (4)
TypeParser now enters symbols for generics in case the assemblies being
linked sport them, (5) a few fixes for bytecode verif in GenMSIL.
review by rytz
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Some refinement of the error messages when the found and required types
have the same simple names. No longer must we watch people scratch their
heads at such messages as:
found : scala.collection.Set[String]
required: Set[String]
Now so clear you could enjoy a movie through it:
found : scala.collection.Set[String]
required: scala.collection.immutable.Set[String]
No review.
|
| |
|
|
|
|
|
|
|
|
|
| |
* ctrl-C will no longer kill the repl unless you hit it again
* ctrl-Z will no longer make the repl useless because of jline
In the service of the first I wrote signal handling code, which we can
put to use in other ways as well. No review.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
I avoided creating any dependency on yourkit. In addition, there was
no way to give arguments to the JVM without losing the ones defined in
ANT_OPTS, which has been a massive pain for a while. So there is now
"jvm.opts" which is simply appended to ANT_OPTS, e.g.
% ant -Djvm.opts=-verbose
[echo] Forking with JVM opts: -Xms1536M -Xmx2g -Xss1M -XX:MaxPermSize=192M -XX:+UseParallelGC -verbose
There is a minimal stub defining a profiler interface:
scala.tools.util.Profiling
Then the yourkit wrapper implements that interface. Once your locker has
been rebuilt once, you can do this:
ant yourkit.run
And it will build quick.lib/comp with profiling enabled, assuming it
can find the necessary files. See the yourkit.init target for values to
change: or ant -Dyourkit.home=/path/to/it might be enough.
Review by dragos.
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
been ignoring their painstakingly set classpaths and using whatever they
can find in the vicinity of the scala script, which works OK for pack
but not so well for quick.
I also added a build.xml feature so you can drop jars into a dir
and have them automatically used everywhere. The directory is
lib/extra. Pro tip: add lib/extra to .gitignore and put jars like
yjp-controller-api-redist.jar in there.
Closes #3706, no review.
|
| |
|
|
|
|
|
|
| |
a memory leak when running scalac in resident mode (including the
presentation compiler and build manager). no review.
|