| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
|
|
|
| |
revert commit that probably broke the build due to a bug in closure
elimination revert commit that reverted the commit that probably didn't
break the build
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Revert r21790: "closes #3373, #3177: validity check of an implicit "
value should consider the value as well as its accessor Revert r21792: "
"sorry, last-minute edit broke build: re-added parens "
will investigate how it broke and why my local pre-commit tests did not
catch the failure...
no review
|
|
|
|
|
|
| |
closes #3373, #3177: validity check of an implicit value should consider
the value as well as its accessor review by odersky
|
| |
|
|
|
|
|
|
| |
Added test for several tickets that were fixed earlier this week. no
review.
|
|
|
|
|
|
|
|
|
|
| |
fixed #3349 : method symbol cached too aggressively
cooking raw types changes a symbol's info, but the change was masked by
caching in MethodSymbol
review by odersky
|
| |
|
|
|
|
|
|
| |
Fixed implicits problem in specs; Manifest[Class[_]] can now be
generated. review by dubochet.
|
|
|
|
|
|
|
| |
Fixed bug in update method visibility which caused x += y not to find
x's update method when defined generically. Closes #3278. Already
reviewed by odersky.
|
|
|
|
|
|
|
| |
Improved @tailrec error messages to specify the reason. In the process
fixed old bug involving tail call transformation. Closes #3275, #2018.
Review by dragos.
|
|
|
|
|
|
|
| |
back in the form of a compiler plugin. Also promoted -Yfatal-warnings to
-Xfatal-warnings: this is key to writing tests which involve warnings
and should be at least semi-supported. Closes #3300, no review.
|
| |
|
|
|
|
|
|
|
|
|
| |
As a brief diversion from real work, implemented Damerau–Levenshtein
and ran it on trunk to elicit obvious misspellings. Unfortunately
they're mostly in places like compiler comments which real people never
see, but I fixed them anyway. All those English Lit majors who peruse
our sources are sure to be pleased. No review.
|
|
|
|
|
|
|
|
| |
Fix for the partest task to fail the build when a test fails, and fixes
for 2/3 of the quietly failing tests. I'm not quite sure what to do
about the view ones, it doesn't look like a simple rename is going to
cut it, so: review by odersky.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
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.]
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
| |
different from the other errors, so no second try was done for them.)
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
closes #3152: refactored adjustTypeArgs and methTypeArgs so that tparams
are correctly split into ones that were inferred successfully, and
that thus have a corresponding type argument, and those that weren't
determined
I didn't investigate the exact cause of the final error message in the
bug report, but Jason Zaugg's observations seems correct and I never
liked that uninstantiated buffer in the first place.
review by odersky
|
|
|
|
|
|
|
|
|
| |
make normalize slightly more aggressive in loading symbol info, while )
tolerating the righteous cycle (use sym.info.typeParameters instead )
of unsafeParams this is needed to make sure higher-kinded types have )
their type parameters (otherwise we'd get a PolyType with NoSymbol for )
typeParams )
|
| |
|
|
|
|
|
|
|
|
|
| |
Fix for #3136 by reverting the line in r18184 which caused this and
other regressions. The downside is that the #1697 test case no longer
passes, but protracted shrug because it wasn't entirely fixed anyway.
Review by moors. (Can you triangulate your way to a patch where both
work simultaneously? It's today's bonus challenge!)
|
| |
|
| |
|
|
|
|
|
| |
specialized implementations.
|
| |
|
|
|
|
|
| |
worksforme
|
|
|
|
|
|
| |
closes #2421 -- now also deals with chained implicits
no review
|
|
|
|
|
|
|
| |
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
|
|
|
|
|
|
|
| |
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.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
is final and does not conform to the pattern, it will no longer compile.
See all the exciting things you can no longer do:
"bob".reverse match { case Seq('b', 'o', 'b') => true } // denied!
"bob".toArray match { case Seq('b', 'o', 'b') => true } // rejected!
final class Dunk
def f3(x: Dunk) = x match { case Seq('b', 'o', 'b') => true } // uh-uh!
And so forth. Review by odersky.
|
|
|
|
|
|
|
|
|
|
|
|
| |
The non-intrusive bits of my hopefully pending "use the static type of
the scrutinee to rule out some type/extractor patterns" patch. Includes
a cleanup of the (still inadequate) type-parameter-ignoring match test
which had been interfering with martin's digestion.
Also: implicit search is disabled when typing a pattern, because the
matcher never invokes implicits to satisfy a pattern. At worst maybe
we'll get a performance bump. No review.
|
| |
|
| |
|
|
|
|
|
| |
suppresses generation of manifests for abstract type members.
|
|
|
|
|
|
|
|
| |
Moved some test cases from pending to files since the bugs they were
watching for seem to be fixed. Moved some other test cases from pending
to disabled because they deceptively claim to pass while investigation
reveals the ticket needs to remain open. Closes #1996, #2660.
|
|
|
|
|
|
|
|
|
| |
One of those "$.05 for the bolt, $50,000 for knowing where to put it"
commits. Closes #425, #816, #2310, #2691. All credit for this patch goes
to me for having the genius to know when new eyes were needed (although
if you're feeling generous some could also go to walter korman for the
actual debugging and code writing part.)
|
| |
|