| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
|
|
| |
made unsafeTypeParams transform infos less eagerly. Fixed private[this]
variance hole.
|
|
|
|
|
| |
accidental commit by drmaciver.
|
|
|
|
|
|
| |
Changed linkedSym to use rawInfo rather than info so the linked* methods
in Symbol consistently refrain from forcing the type.
|
| |
|
|
|
|
|
| |
Fixed failing stability test by restoring removed code.
|
|
|
|
|
|
| |
Reverted a small parser change which had the unfortunate side effect of
making it much harder to enter code into the interpreter.
|
| |
|
|
|
|
|
|
|
|
|
| |
Another big Interpreter patch, this one mostly attacking
InterpreterLoop. This adds opening cuts of a number of features: tab
completion for repl identifiers (requires using -Ycompletion for now),
a :power repl command that enables power user commands, and more. I'll
document it properly once it's a bit less experimental.
|
|
|
|
|
|
|
|
| |
Removed most of the assertions from NewScanners and replaced them with
parse errors so as to limit the number of gratuitous compiler crashes.
See at least tickets #822, #1674, #1682, and #1765. The error cascading
still needs work.
|
|
|
|
|
|
|
|
| |
readability of generated code in preparation for accomplishing more
advanced things on that front, but along the way I was also able to
refactor out significant amounts of duplication and remove some dead
code. This also fixes bug #1546.
|
| |
|
| |
|
|
|
|
|
| |
Removed call to normalize, which fixes crashes in the optimiser
|
| |
|
|
|
|
|
|
|
| |
Fix and test case for #1107, many long days in the making. If you fully
understand why this change causes #1107 to work I implore you to add a
comment or two to Symbols.scala.
|
| |
|
|
|
|
|
| |
Iterable no-arg mkString changed to call mkString() to avoid NPEs.
|
|
|
|
|
| |
Fix for #1365, '_' should be a standard identifier char.
|
|
|
|
|
|
| |
Fix for #1745 - only emit generic signature if method is not
deferred.
|
| |
|
|
|
|
|
|
| |
Created separate -Xprint-icode setting so using -Xprint:all doesn't fill
the current directory with usually unwanted .icode files.
|
|
|
|
|
| |
Fix for #1749; removed some dead code related to regexp patterns.
|
| |
|
|
|
|
|
| |
Reverting last commit as it seems to have caused breakage.
|
| |
|
| |
|
|
|
|
|
|
| |
Altered 2-argument versions of Predef's assert, assume, and require to
be call-by-name.
|
|
|
|
|
| |
Removed $tag method from ScalaObject and rebuilt starr.
|
|
|
|
|
|
| |
Temporarily suppressing generic signatures on static forwarders until I
can figure out the right way to fix #1745.
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
fixed several problems with cyclic references uncovered by experimenting
with collections. Added early type definitions.
|
| |
|
|
|
|
|
| |
added a hook for annotation checkers in Typers.adapt
|
|
|
|
|
|
| |
Continuation of r17177 to also check type annotations; test case
broadened accordingly.
|
| |
|
|
|
|
|
|
| |
Fix and test case for #935; added test in RefChecks to confirm that
parameterized annotations conform to type bounds.
|
|
|
|
|
|
| |
When inferring a type is scala.Whatever, have to take care to _root_
qualify it so a package named scala doesn't subvert it. Fixes #430.
|
|
|
|
|
| |
scala.util.Properties: release InputStream after properties loading
|
|
|
|
|
|
| |
scala.util.Properties.versionString: fix default value that is used if
library.properties file cannot be loaded
|
|
|
|
|
|
| |
Proposed fix for #1735; may need refinement, but I think this patch is
less dangerous than the already commited patch for #363.
|
|
|
|
|
|
| |
Unfortunately I can't figure out a good way to test this. But it doesn't
seem to break anything existing.
|
| |
|
|
|
|
|
|
| |
Look for main method in companion object and add a static forwarder to
companion class if possible + test case; fixes bug #363.
|
|
|
|
|
| |
supplemented lacking documentation for sameElements; addresses #888.
|
|
|
|
|
|
| |
Improved error messages for certain illegal overriding situations, plus
test cases. Resolves bug #276.
|
| |
|
|
|
|
|
| |
Fix for #1465 - java annotations aren't parameterized.
|