| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
| |
|
|
|
|
|
| |
tentative solution of typeParams problem discovered by Iulian.
|
|
|
|
|
|
|
| |
Hacky but no-risk fix for #3189, which was caused by the ghost of
regular expression patterns rising from the grave to haunt the parser.
No review.
|
|
|
|
|
|
| |
closes #3429
review by odersky
|
|
|
|
|
|
| |
closes #3404
review by odersky
|
|
|
|
|
|
| |
Changed strategy to allow specialization on type params at @specialized
positions. no review.
|
| |
|
| |
|
|
|
|
|
|
|
| |
[scaladoc] Closes #3428 (HTML markup supported in Scaladoc). Improves
wiki-syntax parsing of lists. Improves stylesheet for headings and code
blocks in comments. Review by malayeri.
|
|
|
|
|
|
|
|
|
|
|
| |
Swapped the order of the arguments to returning to make it consistent
with the way I've been using it for a long time. No review, how about
a question for nobody in particular instead: can we put this in the
library somewhere so I can stop looking around for it everywhere I go?
It's really handy.
def returning[T](x: T)(f: T => Unit): T = { f(x) ; x }
|
|
|
|
|
|
|
| |
Starting to look like the pattern matcher and I can meet in the middle.
More distributing my acquired bottom-up knowledge among the current
code. No review.
|
|
|
|
|
|
|
|
|
|
|
| |
Fixed a missed opportunity for code rewriting in specialize: a method
that was specialized both on enclosing class types and its own was
routed to the generic one. For instance, a map[@specialized B] would
exhibit this bug.
Restricted warnings to non-synthetic methods, removing warnings for
Tuple default argument methods. No review.
|
|
|
|
|
|
| |
Took advantage of my new wealth of bottom-up pattern matching knowledge
to simplify MixTypes a bunch. No review.
|
|
|
|
|
|
|
| |
Realized that somehow the synthetic flag on pattern matcher temp vars
was lost somewhere, and classfile debugging info was being generated for
every one of them. Fixed. No review.
|
| |
|
| |
|
|
|
|
|
|
|
| |
crashing-under-optimise test into pending since that looks like
a genuine bug - I'll open it a ticket. No review.
|
|
|
|
|
|
| |
Removed an assertion which was asserting an untrue statement thus
causing chaos and strife. Closes #3411, no review.
|
|
|
|
|
|
|
|
| |
Fixed a long standing issue with interpreter commands containing more
than a single statement. It was the classic "reversing a list and then
flattening it is not the same as flattening and then reversing it."
Closes #3116, no review.
|
|
|
|
|
|
|
| |
Notice and fail when a file given with -i is incomplete (which under
other conditions in the repl is quite different from not compiling, thus
the bug.) Closes #3011, no review.
|
|
|
|
|
|
|
|
| |
Overhauled sequence length logic in the pattern matcher. Removes
unnecessary boxing and a few varieties of wrongness.
Closes #3395, #3150, #2958, #2945, #2187. No review.
|
|
|
|
|
|
| |
Closes #3424, double definition when overriding a generic method with
specialized types, inherited from a specialized class. no review.
|
|
|
|
|
|
| |
[scaladoc] Closes #3351 (white space before first period in comment). No
review.
|
|
|
|
|
| |
Removed more than 3400 svn '$Id' keywords and related junk.
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
Remove the sourcepath prefix from the logs and only then do dos2unix
paths changing (otherwise #3054 fails). It is a bit ugly but it is only
done for buildmanager (paul was doing it anyway in his reverted partes).
That should really fix the windows build. 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.)
|
|
|
|
|
|
| |
Removed some logging and prune CollectMethodBodies to not traverse more
than necessary. No review.
|
|
|
|
|
|
|
|
| |
Tightened what gets specialized: only when the type parameter appears
at top level, or as a type argument to a Java array. For example T,
Array[T] cause specialization, but List[T] does not. Resurrected
spec-matrix, forgotten among the disabled tests. No review.
|
|
|
|
|
|
| |
revert the revert of r21791 (fix constructor parameter annotations).
close #3390. together with a new starr it builds. no review.
|
|
|
|
|
|
| |
Fixing the conflict between the two private-related bugs, fixing the
build. no review.
|
| |
|
| |
|
|
|
|
|
|
|
| |
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
|
|
|
|
|
| |
sorry, last-minute edit broke build: re-added parens
|
|
|
|
|
| |
fix and test where constructor parameter annotations end up. no review
|
|
|
|
|
|
| |
closes #3373, #3177: validity check of an implicit value should consider
the value as well as its accessor review by odersky
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
| |
Removed 4-year old println when getModuleOrClass fails. It polluted
debug output when -Ydebug was present because now every (successful)
compiler run hits a few MissingRequirementErrors, quietly swallowed
by getModule2/getClass2. The errors are normal during the transition
to 'scala.collection.immutable.List' and 'scala.Seq/Sequence', while
definitions are built. No review.
|
| |
|
| |
|