| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
-explainTypes means that only type tests which *fail* should be reported in more
detail by using explainTypes. Hence, callers of explainTypes should check if
type errors are being ignored, by checking context.reportErrors. Hence, this
check is added to Inferencer, and another call site is redirected to that
method.
Moreover, explainTypes should only be called if an error exists. Enforce that in
checkSubType, and remove spurious home-made explainTypes output.
Finally, in ContextErrors, stop checking `settings.explaintypes.value` before
calling `explainTypes` which will check it again.
Note that this patch does not fix all occurrences, but only the ones which
showed up during debugging. The other ones never cause problems, maybe because
they occur when contextErrors is in fact guaranteed to be true. We might want to
fix those ones anyway.
This fixes regressions in c800d1fec5241ed8c29e5af30465856f9b583246 and
78f9ef3906c78413ff8835fdad3849bfe5516be2.
Thanks to hubertp (Hubert Plociniczak) for the first round of review.
Refs #6123
backport to _2.10.x_
|
|\
| |
| | |
SI-7228, bug in subtyping.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Diligent reviewer observed that a hot spot was possibly
being made hotter. Reviewer's suggested remedy was a
spectacular bust, but studious observation revealed the news
lash that expensive methods are expensive and we should
avoid calling them if we can.
Put short-circuit test back in front of unapply call.
Now the time spent in unapply is within a few percent.
|
| |
| |
| |
| |
| | |
Fixed bug in numeric widening related to continuations,
which enabled simplifying isNumericSubType.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Should speak for itself.
Whenever someone changed @switch from an error to a warning,
it broke all the tests which depended on the error. I added
-Xfatal-warnings to a couple which needed it. And one of those
tests was then failing, as it must now since we couldn't get away
with what was being attempted, so I moved it to pending.
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Another in the category of bugs which involve narrowing,
widening, mediuming, dealiasing, weakening, normalizing,
denormalizing, supernormalizing, subnormalizing, and
double-bounded supersubnormalizing.
This is probably not the ideal fix, but it is an improvement.
|
|\ \
| | |
| | | |
Merge 2.10.x into master
|
| |\ \
| | | |
| | | |
| | | |
| | | |
| | | | |
Conflicts:
src/compiler/scala/reflect/reify/Errors.scala
src/compiler/scala/tools/nsc/typechecker/NamesDefaults.scala
|
| | |\ \
| | | | |
| | | | | |
SI-7328 Bail out of names/defaults when args are error typed
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
To avoid a crasher later on with a null type inside a
sequence argument.
|
| | |\ \ \
| | | | | |
| | | | | | |
SI-7234 Make named args play nice with dep. method types
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Some care is needed to avoid interaction with constant
types (e.g pos/z1730.scala) and with existentials
(e.g. t3507-old.scala).
|
| | |\ \ \ \
| | | | | | |
| | | | | | | |
reifier is now aware of SI-7235
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
SI-7235 is caused by a long-standing todo in typedRefinement, which leads
to originals of compound type trees swallowing their stats.
I'm not sure how exactly to fix SI-7235, but what I am sure about is that
we shouldn't silently discard stats during reification. This patch
introduces a new implementation restrictions, which now reports that
reify of compound type trees with non-empty stats isn't going to work.
|
| | | |_|/ /
| | |/| | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
TypeVars, being mutable creatures, mustn't have structural
equality/hashing, otherwise TypeRefs that differ only by
having distinct TypeVars as components get wrongly uniqued
together.
The reported bug showed the disaterous consequences: constraints
from the `C?[Int]` in the return type applied to the `?C[?A]` in
the parameter list.
This commit overrides `equals` and `hashCode` in `TypeVar`
to use reference equality. An alternative fix would be to drop
the `case`-ness of the class, as was the case before 0cde930b
when this regressed.
|
| |\| | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Conflicts:
src/compiler/scala/tools/nsc/typechecker/Typers.scala
|
| | | |/ /
| | |/| |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Don't focus positions assigned to the TypeTree of a parent of a Template
when the parent exists in the source. A focused position (instance
of `OffsetPosition`) makes it invisible to the presentation compiler
when it looks for the innermost enclosing tree of a given position
(breaking hyperlinking, for example).
|
| | | | |
| | | | |
| | | | |
| | | | | |
And tweak deprecated stub so it compiles with 0.12.2.
|
|\ \ \ \ \
| | | | | |
| | | | | | |
SI-6601 Close access loophole for value class constructors
|
| | |_|_|/
| |/| | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
ExtensionMethods marks private constructors of value classes as
notPRIVATE before pickling. When the pickler reads the flags of this
symbol, the anti-shift mechanism folds this into the regular PRIVATE
flag, so the class is pickled as though it was public all along. A
seprately compiled client can then call this constructor.
To remedy this, we must:
- pickle `rawFlags`, rather than `flags`. This is symmetric with
unpickling, which sets `rawFlags` with the value it reads.
- Add `notPRIVATE` to the flagset `PickledFlags`.
We cannot make this change in a minor version, as the pickler
and unpickler must agree on `PickledFlags`.
I believe that this won't change the size of pickled flags
for the majority of symbols (ie, those without the notPRIVATE flag)
due to the variable length encoding in `writeLongNat`.
This also improves the situation for SI-6608. Reflection and
scalap (and, by extension, IntelliJ), no longer will see
as public methods that have had their access widened in
SuperAccessors (which is done selectively to support inlining
under separate compilation.)
|
|\ \ \ \ \
| |_|/ / /
|/| | | | |
Warn about locally identifiable init order issues.
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
This warns about a subset of initialization order issues which
can easily be identified by inspection, such as
val x = y
val y = 5
The likelihood of this formulation being intentional is miniscule.
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Since we've removed scala.annotations.serializable in 2.11,
you'll need to following bandaid to use eclipse on master.
This is the easiest fix, but certainly not recommended in general.
```
--- i/src/eclipse/scala-library/.classpath
+++ w/src/eclipse/scala-library/.classpath
@@ -1,7 +1,7 @@
<?xml version="1.0" encoding="UTF-8"?>
<classpath>
<classpathentry kind="src" path="library"/>
- <classpathentry kind="con" path="org.scala-ide.sdt.launching.SCALA_CONTAINER"/>
+ <classpathentry exported="true" kind="con" path="org.scala-ide.sdt.launching.SCALA_CONTAINER"/>
```
This corresponds to checking the export check box for the scala library jar
in the library's build path.
|
|\ \ \ \ \
| |/ / / /
|/| | | | |
[rebase of #2206] l JSR-223 compliance for the interpreter
|
| | | | | |
|
| | | | | |
|
|\ \ \ \ \
| | | | | |
| | | | | | |
Topic/ide config
|
| | | | | | |
|
| | |/ / /
| |/| | | |
|
|\ \ \ \ \
| |_|/ / /
|/| | | | |
Close after slurping (fixes SI-7244)
|
| |/ / /
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Since slurp by definition is the noise made by the straw when
the cup is empty and all milk shake is consumed, we can safely
close after slurping.
Use case was deleting a test artifact on Windows.
|
|\ \ \ \
| | | | |
| | | | | |
A small typo corrected
|
| | | | | |
|
| |/ / /
|/| | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Following in the footsteps of scaladoc and interactive.
The interpreter sources move into src/repl, and are given
a separate build target. As with the others, at present
they are still packaged into scala-compiler.jar.
A summary of changes:
- repl requires use of ReplGlobal (this was already implied)
- macro code's repl-specific classloader hack pulled into overridable
method and overridden in ReplGlobal
- removed -Ygen-javap option to eliminate backend's dependency on javap
- removed -Yrepl-debug option (can still be enabled with -Dscala.repl.debug)
- pushed javap code into src/repl so javax.tools dependency can bee
weakened to the repl only
- removed some "show pickled" related code which hasn't worked right
in a while and isn't the right way to do it anymore anyway. Will
return to fix showPickled and provide it with some tests.
|
|\ \ \ \
| | | | |
| | | | | |
Carve up Types.scala
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Step one of a plan to bring some order and thread
safety to this neck of the woods.
More info:
https://gist.github.com/retronym/5081754
https://groups.google.com/forum/?fromgroups=#!topic/scala-internals/MOvmcnbyb_g
Note that the sub package is named 'tpe' and not 'Types' to avoid
potential problems on case insensitive file systems, given the existing
trait 'Types.'
|
| | | | | |
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
This is the commit which brings it all together. The booleans
forInteractive and forScaladoc are now deprecated and are not
inspected for any purpose. All behavioral changes formerly
accomplished via tests of those flags are embodied in the globals
built specifically for those tasks.
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
As with scaladoc, pushes presentation compiler specific
code into its separate source area.
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
This leverages the preceding several commits to push scaladoc
specific code into src/scaladoc. It also renders some scanner
code more comprehensible.
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Some new partest abilities which were necessary to deal with
additional standard jars. These new abilities aren't yet put
to the test because scaladoc and interactive are still going
into the compiler jar.
No longer need each jar or classes directory be named
directly for partest to find them.
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
As with the preceding commit, this has build-internal
effects only.
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
This change is not externally visible. It moves the scaladoc
sources into src/scaladoc and adds an ant target for building
them. The compilation products are still packaged into
scala-compiler.jar as before, but with a small change to
build.xml a separate jar can be created instead.
|
| |/ / /
|/| | |
| | | |
| | | |
| | | |
| | | | |
This is a non-behaviorally-changing setup commit which
re-routes bits of code through avenues which can more easily
be influenced by subclasses of Global.
|
|\ \ \ \
| | | | |
| | | | | |
SI-7231 Fix assertion when adapting Null type to Array type
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
GenICode was doing a sanity check when adapting an expression of type
Null to something else. It was just doing the wrong one. Instead of
checking whether the result expression type was a reference type it
was checking to see if it was an class reference type. This commit fixes
that and adds a test to make sure both forms of adaptation work as
expected.
|
|\ \ \ \ \
| |/ / / /
|/| | | | |
SI-7006 Prevent unreachable blocks in GenICode
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Minor cleanup from review of https://github.com/scala/scala/pull/2185
* Changed several instances of |= to ||= for better clarity (and bonus
savings!)
* Documented the return of two methods that compute the reachability
of follow-on blocks.
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
This commit makes GenICode prevent the generation of
most unreachable blocks. The new unreachable block prevention code can
be disabled with a compiler flag.
Because full unreachable analysis is no longer necessary for
normal code it makes the unreachable block analysis run only under
-optimise.
A test is included to make sure unreachable code doesn't cause issues
in code gen.
A concrete example will help.
def foo(): X = {
try
return something()
catch {
case e: Throwable =>
println(e)
throw e
}
unreachableCode()
]
Here unreachableCode() is unreachable but GenICode would create ICode
for it and then ASM would turn it into a pile of NOPS.
A previous commit added a reachability analysis step to eliminate
that unreachable code but that added a bit of time to the
compilation process even when optimization was turned off.
This commit avoids generating most unreachable
ICode in the first place so that full reachability analysis is
only needed after doing other optimization work.
The new code works by extending a mechanism that was already in place.
When GenICode encountered a THROW or RETURN it would put the
current block into "ignore" mode so that no further instructions
would be written into the block. However, that ignore mode flag
was itself ignored when it came to figuring out if follow on blocks
should be written. So this commit goes through places like try/catch
and if/else and uses the ignore mode of the current block to decide
whether to create follow on blocks, or if it already has, to kill by
putting them into ignore mode and closing them where they'll be
removed from the method's list of active blocks.
It's not quite as good as full reachability analysis. In particular
because a label def can be emitted before anything that jumps to it,
this simple logic is forced to leave label defs alone and that means
some of them may be unreachable without being removed. However, in
practice it gets close the the benefit of reachability analysis at
very nearly no cost.
|
|\ \ \ \ \
| |/ / / /
|/| | | | |
Merge 2.10.x into master
|