| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Makes fields fit the field width, which is fitting.
`s/including/add` seems sufficient.
Possibly, "synthesize" is an extravagance for "add",
but "add" is used previously in that column.
Resolve, load, translate, add, synthesize, replace, erase,
move, eliminate, remove, generate.
Would love to learn a word that says what typer does, if
the word "type" is too redundant or overloaded, besides the
food metaphor. Also "meat-and-potatoes" implies basic,
simple, not fussy or fancy. There are many devices,
like the heart or a Ferrari engine, that are
fundamental without being unfussy.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Essentially, we fuse mixin and lazyvals into the fields phase.
With fields mixing in trait members into subclasses, we
have all info needed to compute bitmaps, and thus we can
synthesize the synchronisation logic as well.
By doing this before erasure we get better signatures,
and before specialized means specialized lazy vals work now.
Mixins is now almost reduced to its essence: implementing
super accessors and forwarders. It still synthesizes
accessors for param accessors and early init trait vals.
Concretely, trait lazy vals are mixed into subclasses
with the needed synchronization logic in place, as do
lazy vals in classes and methods. Similarly, modules
are initialized using double checked locking.
Since the code to initialize a module is short,
we do not emit compute methods for modules (anymore).
For simplicity, local lazy vals do not get a compute method either.
The strange corner case of constant-typed final lazy vals
is resolved in favor of laziness, by no longer assigning
a constant type to a lazy val (see widenIfNecessary in namers).
If you explicitly ask for something lazy, you get laziness;
with the constant-typedness implicit, it yields to the
conflicting `lazy` modifier because it is explicit.
Co-Authored-By: Lukas Rytz <lukas@lightbend.com>
Fixes scala/scala-dev#133
Inspired by dotc, desugar a local `lazy val x = rhs` into
```
val x$lzy = new scala.runtime.LazyInt()
def x(): Int = {
x$lzy.synchronized {
if (!x$lzy.initialized) {
x$lzy.initialized = true
x$lzy.value = rhs
}
x$lzy.value
}
}
```
Note that the 2.11 decoding (into a local variable and a bitmap) also
creates boxes for local lazy vals, in fact two for each lazy val:
```
def f = {
lazy val x = 0
x
}
```
desugars to
```
public int f() {
IntRef x$lzy = IntRef.zero();
VolatileByteRef bitmap$0 = VolatileByteRef.create((byte)0);
return this.x$1(x$lzy, bitmap$0);
}
private final int x$lzycompute$1(IntRef x$lzy$1, VolatileByteRef bitmap$0$1) {
C c = this;
synchronized (c) {
if ((byte)(bitmap$0$1.elem & 1) == 0) {
x$lzy$1.elem = 0;
bitmap$0$1.elem = (byte)(bitmap$0$1.elem | 1);
}
return x$lzy$1.elem;
}
}
private final int x$1(IntRef x$lzy$1, VolatileByteRef bitmap$0$1) {
return (byte)(bitmap$0$1.elem & 1) == 0 ?
this.x$lzycompute$1(x$lzy$1, bitmap$0$1) : x$lzy$1.elem;
}
```
An additional problem with the above encoding is that the `lzycompute`
method synchronizes on `this`. In connection with the new lambda
encoding that no longer generates anonymous classes, captured lazy vals
no longer synchronize on the lambda object.
The new encoding solves this problem (scala/scala-dev#133)
by synchronizing on the lazy holder.
Currently, we don't exploit the fact that the initialized field
is `@volatile`, because it's not clear the performance is needed
for local lazy vals (as they are not contended, and as soon as
the VM warms up, biased locking should deal with that)
Note, be very very careful when moving to double-checked locking,
as this needs a different variation than the one we use for
class-member lazy vals. A read of a volatile field of a class
does not necessarily impart any knowledge about a "subsequent" read
of another non-volatile field of the same object. A pair of
volatile reads and write can be used to implement a lock, but it's
not clear if the complexity is worth an unproven performance gain.
(Once the performance gain is proven, let's change the encoding.)
- don't explicitly init bitmap in bytecode
- must apply method to () explicitly after uncurry
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
One step towards teasing apart the mixin phase, making
each phase that adds members to traits responsible for
mixing in those members into subclasses of said traits.
Another design tenet is to not emit symbols or trees
only to later remove them. Therefore, we model a
val in a trait as its accessor. The underlying field
is an implementation detail. It must be mixed into
subclasses, but has no business in a trait (an interface).
Also trying to reduce tree creation by changing less in subtrees
during tree transforms.
A lot of nice fixes fall out from this rework:
- Correct bridges and more precise generic signatures for
mixed in accessors, since they are now created before erasure.
- Correct enclosing method attribute for classes nested in trait fields.
Trait fields are now created as MethodSymbol (no longer TermSymbol).
This symbol shows up in the `originalOwner` chain of a class declared
within the field initializer. This promoted the field getter to
being the enclosing method of the nested class, which it is not
(the EnclosingMethod attribute is a source-level property).
- Signature inference is now more similar between vals and defs
- No more field for constant-typed vals, or mixed in accessors
for subclasses. A constant val can be fully implemented in a trait.
TODO:
- give same treatment to trait lazy vals (only accessors, no fields)
- remove support for presuper vals in traits
(they don't have the right init semantics in traits anyway)
- lambdalift should emit accessors for captured vals in traits,
not a field
Assorted notes from the full git history before squashing below.
Unit-typed vals: don't suppress field
it affects the memory model -- even a write of unit to a field is relevant...
unit-typed lazy vals should never receive a field
this need was unmasked by test/files/run/t7843-jsr223-service.scala,
which no longer printed the output expected from the `0 to 10 foreach`
Use getter.referenced to track traitsetter
reify's toolbox compiler changes the name of the trait
that owns the accessor between fields and constructors (`$` suffix),
so that the trait setter cannot be found when doing mkAssign in constructors
this could be solved by creating the mkAssign tree immediately during fields
anyway, first experiment: use `referenced` now that fields runs closer
to the constructors phase (I tried this before and something broke)
Infer result type for `val`s, like we do for `def`s
The lack of result type inference caused pos/t6780 to fail
in the new field encoding for traits, as there is no separate accessor,
and method synthesis computes the type signature based on the ValDef tree.
This caused a cyclic error in implicit search, because now the
implicit val's result type was not inferred from the super member,
and inferring it from the RHS would cause implicit search to consider
the member in question, so that a cycle is detected and type checking fails...
Regardless of the new encoding, we should consistently infer result types
for `def`s and `val`s.
Removed test/files/run/t4287inferredMethodTypes.scala and test/files/presentation/t4287c,
since they were relying on inferring argument types from "overridden" constructors
in a test for range positions of default arguments. Constructors don't override,
so that was a mis-feature of -Yinfer-argument-types.
Had to slightly refactor test/files/presentation/doc, as it was relying
on scalac inferring a big intersection type to approximate the anonymous
class that's instantiated for `override lazy val analyzer`.
Now that we infer `Global` as the expected type based on the overridden val,
we make `getComment` private in navigating between good old Skylla and Charybdis.
I'm not sure why we need this restriction for anonymous classes though;
only structural calls are restricted in the way that we're trying to avoid.
The old behavior is maintained nder -Xsource:2.11.
Tests:
- test/files/{pos,neg}/val_infer.scala
- test/files/neg/val_sig_infer_match.scala
- test/files/neg/val_sig_infer_struct.scala
need NMT when inferring sig for accessor
Q: why are we calling valDefSig and not methodSig?
A: traits use defs for vals, but still use valDefSig...
keep accessor and field info in synch
|
| |
|
|
|
|
|
|
| |
This commit adds a do-nothing phase called "Delambdafy" that will
eventually be responsible for doing the final translation of lambdas
into classes.
|
|
|
|
| |
Patern translation now happens earlier.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
One last flurry with the broom before I leave you slobs to code
in your own filth. Eliminated all the trailing whitespace I
could manage, with special prejudice reserved for the test cases
which depended on the preservation of trailing whitespace.
Was reminded I cannot figure out how to eliminate the trailing
space on the "scala> " prompt in repl transcripts. At least
reduced the number of such empty prompts by trimming transcript
code on the way in.
Routed ConsoleReporter's "printMessage" through a trailing
whitespace stripping method which might help futureproof
against the future of whitespace diseases. Deleted the up-to-40
lines of trailing whitespace found in various library files.
It seems like only yesterday we performed whitespace surgery
on the whole repo. Clearly it doesn't stick very well. I suggest
it would work better to enforce a few requirements on the way in.
|
|
|
|
|
|
|
|
| |
Let optimiser components and continuations plugin opt-out
when required flags are not set.
Wasted time on a whitespace error in check file, so let
--debug dump the processed check file and its diff.
|
|
|
|
|
|
|
|
|
|
|
| |
Refactor the calculation of the "phase chain" a bit.
In particular, initial and terminal phases are not special
except that they must be head and last.
When done, filter for enabled phases. At this commit,
nobody claims to be disabled.
Additional sanity support of phases settings.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This commit adds analysis and optimization of constants to remove
unnecessary branches. It uses abstract interpretation to determine
what constant(s) a particular stack slot or variable might or might not
hold at a given spot and uses that knowledge to eliminate branches that
cannot be taken. Its primary goal is null check removal, but it also
works for other constants.
Several tests are modified to include the new optimization phase.
Two new tests are added. One verifies that branching still works as
expected. The other verifies that branches are removed.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In Global, SubComponent is called a phase descriptor, but it doesn't
actually have a description. (Phase itself does.) This fix adds
a description to PluginComponent so that plugins can describe what
they do in -Xshow-phases.
Elliptical descriptions
Exploded archives
Plugged-in partest
Roundup at the Little h!
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
sort field accessors, necessary after typers -- apparently...
don't throw TypeError, use issueTypeError
don't run patmat phase when -Xoldpatmat
only virtualize matches when -Xexperimental
recycle cps type of match for re-typechecking:
when one of the internal cps-type-state annotations is present, strip all CPS annotations
a cps-type-state-annotated type makes no sense as an expected type (matchX.tpe is used as pt in translateMatch)
don't synth FunctionN impls during typer, only do this for PartialFunction
updated check now function synth for match is deferred until uncurry
patmat-transform try/catch with match in cps
cleanup in selective anf
remove TODO: can there be cases that are not CaseDefs -- nope
|
| |
|
|
|
|
| |
enforced. Super calls and specialized still missing.
|
|\
| |
| |
| |
| |
| | |
Conflicts:
src/compiler/scala/tools/nsc/Global.scala
test/files/run/programmatic-main.check
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Major cleanup of reification:
* LiftCode phase has been removed
* Code has been deprecated and will be removed as we roll a new starr
* Logic related to type-directed lifting has been purged
scala.reflect.macro.Context#reify now provides the same services
as LiftCode provided (except that it returns Tree, not Code).
For testing purposes, I've retained the oh-so-convenient automagic lift.
test/files/codelib/code.jar now hosts Code.lift reimplemented in a macro,
so that the tests can continue working as if nothing has happened.
|
| | |
|
|/
|
|
|
|
|
| |
Temporarily removed getClass from AnyVal to get build going.
Disabled anyval-childen test.
Fixed some other build problems.
Implemented step 1 + 2 of inline classes proposal.
|
|
|
|
|
|
| |
[recommit] Backend optimization: Inline exception handlers. Review by
dragos.
|
| |
|
|
|
|
|
|
|
|
| |
Cleaned up and brought up to date the help text for -X and -Y options.
Made some enhancements to PhasesSetting, which are documented if you
run scalac -X or -Y. (Try it!) Disabled some dead settings and renamed
others to reflect their current purpose. No review.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
consistently, and remove things which are not being used anywhere
in the visible universe. Beyond general polish here are some of the
feature-like additions I can remember:
* -Xshow-phases now includes descriptions of the phases.
* -Xshow-class and -Xshow-object did not work as far as I could tell:
if they didn't, now they do. If they did, now they work better.
And you don't have to give it a fully qualified name anymore.
* -Xprint-icode will generate *.icode files (don't also have to say -Xprint:icode)
* counts of deprecation and unchcked warnings are given
* More documentation of what global is doing.
I tried not to break anything which might be using Global, but let me
know if I overshot somewhere. No review.
|
|
|
|
|
|
|
|
| |
changes necessary to plug it back in while preserving everything which
has happened since then in tests and such, but we should be the lookout
for overreversion. Review by phaller (but as a formality, I don't think
it requires direct review.)
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
If I work on this patch any longer without checking in I will go
stark raving mad. It is broken up into a couple pieces. This one is
the changes to test/. It includes fixing a bunch of tests, removing
deprecated constructs, moving jars used by tests to the most specific
plausible location rather than having all jars on the classpath of all
tests, and some filesystem layout change (continuations get their whole
own srcpath.) This would be the world's most tedious review, so let's
say no review.
[Note: after this commit, I doubt things will build very smoothly until
the rest of the partest changes follow. Which should only be seconds,
but just in case.]
|
|
More classpath work, and cleanups in the vicinities of everything
manipulating classpaths. Review by anyone willing to slog through the
approximately dozen different ways the classpath can be influenced.
|