| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|\
| |
| | |
partest: Enable separate compilation
|
| | |
|
| |
| |
| |
| | |
In this case, supertype is not stable and should not be cached.
|
| |
| |
| |
| | |
Java's naming convention is different from Scala's.
|
| | |
|
| |
| |
| |
| |
| |
| |
| | |
Currently, the classfiles emitted by dotty do not contain the type
parameters information that javac relies on.
Fixing this is tracked by #1303.
|
| |
| |
| |
| | |
javac wants the public class name to match the filename.
|
| |
| |
| |
| |
| |
| |
| |
| | |
Some java tests require the scala-library to be present on the
classpath, this fixes tests/pos/java-interop/{t1186, t1235, t1254,
t1642}.
Also correctly redirect the output of javac so that it will be displayed
by partest --verbose
|
| |
| |
| |
| |
| |
| | |
This lead to inference failures when separately compiling t1751 and
t294, this did not happen under joint compilation because JavaParser
does not create the overloaded constructor
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This test failed before because strawman.collections.CollectionStrawMan5
is defined in two places:
- src/strawman/collections/CollectionStrawMan5.scala
- tests/run/colltest5/CollectionStrawMan5_1.scala
The first will be compiled by scalac (unless the tests are run through a
bootstrapped dotty) and the second will be compiled by dotty, the value
class encoding of scalac and dotty are not binary compatible.
This would not be a problem if we always used the `CollectionStrawMan5`
coming from the partest output directory and ignored the one in the
dotty sources, but which one gets picked depends on the classpath and whether compilation
is joined or separate, see #1301.
For now, it's safer and simpler to just avoid having tests which define
a class that is also defined in the sources of dotty.
Also, fix a bug in colltest4 where it was importing CollectionStrawMan5
instead of CollectionStrawMan4
|
| |
| |
| |
| |
| | |
This is necessary to unbox the value class when accessing it from
separate compilation units
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
partest can separately compile files based on their suffix (_1, _2,
...), it turns out that this feature was never enabled in the dotty
version of partest and no one noticed (it prints warnings in
./tests/partest-generated/gen.log which no one reads), tests with *.java
files should be compiled both with javac and dotty, but compiling with
javac was also disabled.
Enabling this revealed some latent bugs that will be fixed in the next
few commits.
|
|\ \
| | |
| | | |
Fix #1411: Give REPL output in correct order
|
| | | |
|
|\ \ \
| |/ /
|/| | |
Fixes #1316: Remove JLine
|
|/ /
| |
| |
| |
| | |
Dotty uses ammonite.terminal since April (53bd25f) which replaces JLine.
There is no reason to keep it anymore.
|
|\ \
| | |
| | | |
fix #1354: improve type test of union types
|
| |/ |
|
|\ \
| | |
| | | |
Fix #1353 : more friendly error message for case trait
|
|/ / |
|
|\ \
| | |
| | | |
-Xprint-diff-del: more meaningful colors
|
| | |
| | |
| | |
| | |
| | | |
When a line is modified, show the deleted part in red and the added part
in green instead of using magenta and yellow which are confusing.
|
|\ \ \
| |/ /
|/| | |
Fix #1284: Make classTag depend directly on erasure
|
| | |
| | |
| | |
| | |
| | |
| | | |
In the end, a classTag reflects the erased version of a type. The only
condition for its generation should be that the erasure is stable under
possible instantiations.
|
|\ \ \
| |_|/
|/| | |
Fix parameterized typedefs with lambdas as rhs
|
| | |
| | |
| | |
| | |
| | | |
As @smarter noted, makes no sense to use a boolean parameter here.
We got there by refactoring blindness.
|
| | |
| | |
| | |
| | |
| | | |
Adapt operations TypeApplications and TypeComparer to account
for the possibilities of curried type lambdas.
|
| | |
| | |
| | |
| | |
| | |
| | | |
Previously the compiler crashed when faced with a parameterized typedef
that has a lambda as rhs. We fix this by refining the condition when
not to abstract in typeDefsig.
|
|\ \ \
| |_|/
|/| | |
Fix #1385: Temporarily lift 22 limit for functions
|
| | | |
|
|/ / |
|
|\ \
| |/
|/| |
Fix #1401: Make sure all references are forwarded
|
| |
| |
| |
| |
| | |
Make treatment in Scala2Unpickler and Namer the same and factor
out common functionality.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Faced with recursive dependencies through self types, we might have
to apply `normalizeToClassRefs` to a class P with a parent that is not
yet initialized (witnessed by P's parents being Nil). In that case
we should still execute forwardRefs on P, but we have to
wait in a suspension until P is initialized.
This avoids the problem raised in #1401. I am still not quite sure
why forwardRefs is needed, but it seems that asSeenFrom alone is not
enough to track the dependencies in this case.
|
|/
|
|
|
| |
- increase page width
- print scopes more legibly under -verbose
|
|\
| |
| | |
Fix #1378: Propagate more knowledge of result type into applications
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Interestingly, adding
mutable.BufferLike
to the whitelist succeeds under junit but fails under partest. Unfortunately I can't see
any output in the log indicating what went wrong. I only see this:
!! 306 - pos/compileStdLib [compilation failed]
# Failed test paths (this command will update checkfiles)
test/partest --update-check \
/Users/odersky/workspace/dotty/tests/partest-generated/pos/compileStdLib
|
| |
| |
| |
| | |
This should have been done in a PR that was merged before.
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| | |
If an application has functions with implicit parameter types we need to be
more aggressive about propagating knowledge of the expected result type into
the constraint.
Fixes #1378.
|
| |
| |
| |
| | |
As observed by @smarter, makes sense to do this.
|
| |
| |
| |
| | |
Error messages were suppressed so far, now get printed.
|
|\ \
| | |
| | | |
Make -Xprint-diff an opt-in option.
|
|/ /
| |
| |
| |
| |
| | |
* -Xprint:[...] alone prints the trees without diffs.
* -Xprint-diff enables the diff on -Xprint.
* -Xprint-diff-del enables the diff with deletes on -Xprint.
|
|\ \
| | |
| | | |
Fix wildcard protos
|
| | | |
|
| | | |
|
| | | |
|
|\ \ \
| |/ /
|/| | |
HK reduction: Remove special-case for typerefs
|