| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The previous swing at determinism, ebb01e05cbe4, made decent
contact but apparently didn't hit it out of the park. The test
wavered every hundred or so runs, as witnessed occasionally in
nightly builds or pull request validation.
I setup a test to run neg/7020.scala a few hundred times, and
could trigger the failure reliably.
I then swept through the pattern matcher in search of HashMap and
HashSet creation, and changed them all to the Linked variety.
The results of that are published in retronym#ticket/7020-3 [1].
This commit represents the careful whittling down of that patch
to the minimal change required to exhibit determinism.
[1] https://github.com/retronym/scala/compare/ticket/7020-3
|
|
|
|
| |
test/disabled, not test/files/disabled.
|
| |
|
|
|
|
| |
for macros
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
Instead of trying to track the fallback attempts we rely on the
context state to inform us which fallback is the last one.
setError cannot always be called in NoBestMethodAlternativeError
because inferMethodAlternative relies on side-effects.
Review by @paulp.
|
|
|
|
|
|
| |
See t5452 for disabled/neg/t5452 and disabled/pos/spurious-overload
I give up, review by @hubertp.
|
|
|
|
|
| |
Guess that was a bad choice of class. No review.
|
|
|
|
|
| |
Rebuild all plugin tests and moved them back to the working test cases
|
|
Organized disabled directory so it works with partest. You can run
./partest --srcpath disabled to run the tests in that location. Fixed a
few tests in disabled and pending and moved to files.
|