| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
| |
A few yards short of the goal posts attempt at making our usage of
Throwable subclasses more consistent. This patch eliminates a lot of
ad hoc Exception/Error/etc. creation and various arbitrary choices are
rendered slightly less arbitrary. From now on let's try not to use the
word "Exception" or "Error" in the names of Throwable subclasses unless
they actually derive (and make sense to derive) from Exception or Error.
Review by community.
|
| |
|
| |
|
|
|
|
|
|
|
|
|
| |
test with the string "Manifest" in the source file, you may now do:
./partest --grep Manifest
No review.
|
|
|
|
|
|
| |
Removed the symlinks between scalacheck jars to satisfy windows. Tweaked
partest to accomodate. No review.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
canEquals, and has ClassManifests compare according to erasure but full
manifests also compare type arguments. Preserving symmetry means that
some things you might expect to be equal are not:
val m1 = scala.reflect.ClassManifest.fromClass(classOf[List[String]])
val m2 = manifest[List[String]]
(m1 == m2) // false
However you can always compare the erasures.
(m1.erasure == m2.erasure) // true
Review by dpp.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
Whipped ShowPickled until it would print out private[scope] from the
signature, and infrastructure created along the way. Only now at this
late hour do I realize that this work would be a lot better aimed at
creating a fake Universe and then adapting UnPickler.Scan so you can
reuse the real logic. My advice to the next guy: do that instead. No
review.
|
|
|
|
|
| |
trailing spaces in decompiled annotations are trimmed
|
| |
|
|
|
|
|
|
|
|
|
| |
Undeprecated Function.tupled based on
the type inference issues documented at:
http://stackoverflow.com/questions/2354277/function-tupled-and-placehold
er-syntax We should revisit if anon function inference improves. Review
by community.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
Following a suggestion of jrudolph, made filterKeys and mapValues
transform abstract maps, and duplicated functionality for immutable
maps. Moved transform and filterNot from immutable to general maps.
Review by phaller.
|
| |
|
|
|
|
|
|
| |
Added one previously overlooked case for computing the right tparams of
glbs of polytypes. This is a postscript to the fix of #3082.
|
| |
|
|
|
|
|
|
|
|
| |
what must be legacy scalatest.* properties to partest.*, boldly assuming
that the fact that partest is pretty much unusable outside of scalac
means there are no users outside of scalac who might be disrupted by
eliminating old property names. Review by community.
|
|
|
|
|
| |
specialized implementations.
|
|
|
|
|
|
| |
Modification to r21009 to preserve that classic invariant, (x || !x) &&
!(x && !x). No review.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Added ## method to Any as our scala hashCode method which provides
consistent answers for primitive types. And I'm sure we're all tired of
new starrs, but it's hard to add a method to Any without one. This patch
only brings ## into existence, but nothing calls it yet.
// some true assertions
scala> assert(5.5f.## == 5.5f.hashCode)
scala> assert(5.0f.## != 5.0f.hashCode && 5.0f.## == 5L.##)
No review. (Already reviewed by odersky.)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
While working on Any.## I ran across some interesting tests being made
in TreeBuilder:
val buf = new ListBuffer[(Name, Tree, Position)]
[...]
if (buf.iterator forall (name !=)) ...
This is always true because a Name will never equal a Tuple3. Oh
universal equality, will you never tire of toying with us? Given that
this bug has existed since r12886 one might reasonably question the
necessity of the conditionals falling prey to this, but leaving that
for another day, it should at least check what it's trying to check. No
review.
|
|
|
|
|
|
| |
Expanded the check from #1392 to enclose #3123 as well so that "case Int
=> " doesn't crash. Closes #3123. Review by odersky.
|
|
|
|
|
|
| |
Special cased an error message for the common situation of calling
AnyRef methods on Any or AnyVal. Review by odersky.
|
|
|
|
|
| |
Some library reorganization I discussed with martin. No review.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Removed the partest restriction that individual files must be in the
same set. Haven't you always wanted to do this? Now you can. Review by
phaller.
% ./partest `ack --type=scala -l HashSet | head -6`
Testing individual files
testing: [...]/files/jvm/serialization.scala [ OK ]
testing: [...]/files/jvm/t1600.scala [ OK ]
Testing individual files
testing: [...]/files/pos/collections.scala [ OK ]
testing: [...]/files/pos/t2183.scala [ OK ]
Testing individual files
testing: [...]/files/run/bug1074.scala [ OK ]
testing: [...]/files/run/bug2512.scala [ OK ]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Quite a lot more work on XML equality than I can properly justify
spending time on, but you know how it is once you get started on
something. This commit will likely break some code out there in the
world but this is impossible to avoid if we are to achieve sane equality
in trunk.
For anyone who was relying upon the 2.7 equality behavior for
scala.xml.* classes, using "xml_==" instead of "==" for comparisons will
restore the old behavior. The standard == on xml elements now attempts
to behave in such a way that symmetry and hashCode contracts will be
preserved. It's probably not 100% there yet, but I can tell you this: it
is closer today than it was yesterday.
Review by community.
|
| |
|
|
|
|
|
|
|
| |
Undid my awful code which had broken the thread scheduler selection.
Further unbroke it beyond that unbreaking hopefully to the point where
java 1.6 on OSX is recognized as such. Review by haller.
|
| |
|
|
|
|
|
|
| |
closes #2421 -- now also deals with chained implicits
no review
|
|
|
|
|
| |
see #2634: updated docs to indicate zipped is strict
|
|
|
|
|
|
|
| |
the problem was that corresponds on Seq's does not check length of
sequences before testing the predicate, whereas in some cases that
predicate relied on this invariant (when it was doing substitution)
|
|
|
|
|
|
|
|
|
|
| |
closes #2797 -- no review (already done in ticket by Martin)
1) isHigherKindedType is now false for singletontype 2) toInstance
recurses when pre is a typevar: the problem is that pre.widen.typeSymbol
isNonBottomSubClass symclazz is true while pre.baseType(symclazz) is
NoType
|
|
|
|
|
|
| |
Added a --bare option to ShowPickled so I can diff signatures without
all the explicit indices blowing any points of similarity. No review.
|
|
|
|
|
|
| |
More return type annotation to work around my other frequent guest in
the world of #3082-connected pickler bugs. No review.
|
|
|
|
|
|
| |
What appears to be a workaround for #3082, which I am hitting literally
20 times a day. Will detail in ticket. Review by odersky.
|
|
|
|
|
|
| |
More partest cleanups, and putting back a couple lines in build.xml
which were left a little too commented out. No review.
|
|
|
|
|
|
|
| |
Looking at iulian's patch to the squeezer sent me off looking at
equalsStructure, which clearly was written in a bygone era. Rewritten
with modern tools. No review.
|
|
|
|
|
|
| |
Working on making the bootstrap process transparent and consistent. And
removed a bunch of what is now cruft in partest. No review.
|
| |
|
| |
|
|
|
|
|
|
|
|
|
| |
Made the squeezer worthy of its name: a block with an empty list of
stats and another block as the result expression is rewritten to the
inner block (recursively). This makes the output from the pattern
matcher nicer, eliminating loads of empty nested blocks. Review by
extempore.
|
| |
|
|
|
|
|
|
|
| |
Reorganizes children a little so they always come back sorted the same
way the pickler does. Taking advantage of -Yfatal-warnings in the test
case. Review by community.
|
|
|
|
|
|
| |
timings on all the individual tests and prints it sorted by glacialness.
Review by community.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Created -Yfatal-warnings, as otherwise I can't see how to make partest
fail on the presence of an unchecked warning. It'll come in handy
anyway. Now we have a real tough guy's command line for those who want
it:
% scalac -unchecked -deprecation -Yfatal-warnings `find . -name
'*.scala'`
Not for the timid. Review by community.
|