| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
| |
Not just conveniences though. One of the clearest statements made
by profiling is that collections methods of the form of the enclosed
flatCollect are materially faster than the alternatives.
|
| |\ |
|
| | | |
|
| | |
| |
| |
| |
| |
| | |
Very first time I tried "reify" in the repl and it crashed.
I changed an assert to issue a proper error, as best I could
figure out how. We probably need to do a lot more of this.
|
| |/ |
|
| |
|
|
|
|
|
| |
Moved compose/andThen off Function1, curried/tupled off Function2.
Pushed Tuple2 zipped into auxiliary class. Created implicits to
plaster over the changes. This drops several hundred classfiles and
takes (unoptimized) scala-library.jar from 7.8 Mb to 7.4 Mb.
|
| | |
|
| |\ |
|
| | | |
|
| | |
| |
| |
| |
| |
| |
| | |
typer synthesized the wrong isDefinedAt method in typedMatchAnon
because a BackQuotedIdent was copied as an Ident, so that the equality check
was performed in applyOrElse (since it operates on the original tree),
but not in isDefinedAt (since it operates on the copy, which collapsed Ident and BackQuotedIdent)
|
| |/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
val pattern: Type = expr desugared to
val x = expr match { case pattern: Type => (var_1, ..., var_N)}
but `pattern: Type` is only a valid pattern if `pattern` is an Ident
thus, we desugar to
val x = (expr: Type) match { case pattern => (var_1, ..., var_N)}
... in this case (see def makePatDef)
also, MaybeBoundType now fails on illegal patterns (to defend against similar regressions)
|
| |
|
|
| |
For those of us robots who still have a little human blood.
|
| |\ |
|
| | | |
|
| | |
| |
| |
| | |
@paulp I think this is the place we should dump the contents of AbsSymbolInternal to
|
| | | | |
| | \ | |
| |\ \ \
| | |/
| |/| |
|
| | | | |
|
| | | | |
|
| | | |
| | |
| | |
| | |
| | | |
* all usages of ClassManifest and Manifest are replaced with tags
* all manifest tests are replaced with tag tests
|
| | | |
| | |
| | |
| | |
| | | |
* introduces ArrayTag and ErasureTag
* all type tags now feature erasure
|
| | | |
| | |
| | |
| | |
| | |
| | | |
This reverts commit b47189ad06e027c310d93b071fc8bf15d979225d.
This never happened.
|
| |\ \ \
| |/ /
|/| | |
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
You are a fireman, leading a team of firemen on a rescue
in a burning building. When you have cleared a room, you place
a sticker on the door which says "Room cleared!" If the other
firemen see that sticker, they will skip that room, knowing
there are no small children or adorable puppies behind it.
As the lead fireman, here is what you should not do: run from
door to door slapping "Room cleared!" stickers on them all, so
the other firemen shrug and go home.
Translation: when recursion depends on a condition, if you
alter the condition before you recurse, you may not recurse at all.
|
| |\ \ \
| |_|/
|/| | |
|
| | | | |
|
| | | | | |
| | \ \ | |
| |\ \ \ \
| | | |/
| | |/|
| | | | |
'refs/pull/421/head'; commit 'refs/pull/422/head' into develop
|
| | | |/
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
- Consolidating many islands of name organization. Folds NameManglers
into StdNames. Brings more of the string constants together with similar
constants. Move name manipulation methods which produce TypeNames
into object tpnme rather than nme.
- Starting on MethodSymbolApi, ClassSymbolApi, etc so we can put
sensible methods on sensible entities. This pushed into Definitions,
where I pulled a whole bunch out of the api side (or at least marked
my intention to do so -- too many tests use them to make them easy to
remove) and on the compiler side, returned something more specific than
Symbol a bunch of places.
- Added a number of conveniences to Definitions to make it easier to
get properly typed symbols.
Note: one way in which you might notice having better typed Symbols is
with Sets, which have the annoying property of inferring a type based on
what they've been constructed with and then hard failing when you test
for the presence of a more general type. So this:
val mySet = Set(a, b)
println(mySet(c))
..goes from compiling to not compiling if a and b receive more specific
types (e.g. they are MethodSymbols) and c is a Symbol or ClassSymbol or
whatever. This is easily remedied on a site-by-site basis - create
Set[Symbol](...) not Set(...) - but is an interesting and unfortunate
consequence of type inference married to invariance.
The changes to DummyMirror where things became ??? were driven by the
need to lower its tax; type "Nothing" is a lot more forgiving about changes
than is any specific symbol type.
|
| | |/ |
|
| | |\ \
| | | |
| | | |
| | | | |
'refs/pull/416/head'; commit 'refs/pull/417/head'; commit 'refs/pull/418/head' into develop
|
| | | |/
| |/| |
|
| | |/ |
|
| |/ |
|
| |
|
|
|
|
| |
I also tried transforming a comment into an assertion and to
my shock and happy surprise everything still worked. Let's
express those preconditions in code when we can, mmm?
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
| |
It depended on the signature of uncheckedWarnings.
|
| |\ |
|
| | |
| |
| |
| | |
And we'll need a separate mechanism for making internal fields public.
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
workaround for inliner bug in run/t3835
see SI-5672 for the optimizer bug
tiny switches (2 cases or less) don't make sense anyway,
so this is not a big deal
the annotDepMethType.flags contained -Xexperimental,
which causes Any* to be considered as a valid return type,
before we weren't exposed to this, because the product methods
used (tiny) switches; with this change, they went back to pattern matches
in any case, the -Xexperimental was intended to turn on dependent method types,
not Any* as valid type (I can only assume), and they are now on by default
the other tests were affected by our refusal to emit tiny switches
|
| | |
| |
| |
| |
| |
| |
| |
| |
| | |
detect partialfunction in cpsannotationchecker
emit apply/isDefinedAt if PF has @cps targs (applyOrElse can't be typed)
further hacky improvements to selective anf
better try/catch support in selective cps using freshly minted anonfun match
make virtpatmat resilient to scaladoc (after uncurry, don't translate matches
TODO: factor out translation all together so presentation compiler/scaladoc can skip it)
|
| | |
| |
| |
| |
| |
| | |
Would prefer to bake a little longer, but, scala days.
More elaboration to come.
|
| | | | |
| | \ | |
| |\ \ \
| | |/
| |/|
| | | |
'refs/pull/403/head'; commit 'refs/pull/404/head'; commit 'refs/pull/405/head' into develop
|
| | | | |
|
| |/ /
| |
| |
| | |
of it.
|
| | | | |
| | \ | |
| | \ | |
| | \ | |
| | \ | |
| | \ | |
| |\ \ \ \ \
| | | | | |
| | | | | |
| | | | | | |
'refs/pull/392/head'; commit 'refs/pull/397/head' into develop
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | | |
and implicit pimped on members. Thanks to Lukas for pointing it out!
|
| | | | | |/ |
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Re-enable testing the sensibility of comparing instances
of two case classes. In particular, warn if we detect that
the two objects inherit from different case classes. In
addition, avoid emitting misleading warnings when comparing "new C".
|
