| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
As partest is now resolved from maven, `test/partest` uses `ant test.suite.init`
to determine the classpath (serialized to build/pack/partest.properties)
that's necessary to run `scala.tools.partest.nest.ConsoleRunner`.
Thus, partest gets exactly the same classpath, whether run from
the command line through `test/partest` or via `ant test`.
The version of partest we're using is specified by
properties defined in versions.properties (formerly `starr.number`).
Currently, we're using:
```
scala.binary.version=2.11.0-M4
partest.version.number=1.0-RC3
```
NOTES:
- The version of Scala being tested must be backwards binary compatible with
the version of Scala that was used to compile partest.
- Once 2.11 goes final, `scala.binary.version=2.11`, and `starr.version=2.11.0`.
- Need scalacheck on classpath for test/partest scalacheck tests.
- Removed atrophied ant tests (haven't been run/changed for at least two years
I checked 81d659141a as a "random" sample).
- Removed scalacheck. It's resolved as a partest dependency.
- For now, use a locally built scalap
- Kept the trace macro in the main repo (partest-extras)
- New targets for faster pr validation: test-core-opt, test-stab-opt
- Reused partest eclipse/intellij project to partest-extras
(note: the partest dependency is hard-coded)
|
|\
| |
| |
| |
| |
| | |
Conflicts:
src/compiler/scala/tools/nsc/Global.scala
src/compiler/scala/tools/nsc/transform/patmat/MatchTranslation.scala
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Fixing #ifdiff uncovered a latent bug in partest when it detects
paths to normalize in test log output. It would not normalize
file separators for test names with non-word characters.
This commit normalizes the test name to the ticket number as a
workaround until the next partest version is picked up.
It also fixes partest in this repo in case we've not yet switched
over to the new repo. A separate PR will be submitted to the new
repo as necessary.
|
|/
|
|
|
|
|
|
| |
The new method is the same as sameMethodAndFieldSignatures, but ignores
generic signatures. This allows for testing methods which receive the same
descriptor but differing generic signatures. In particular, this happens
with value classes, which get a generic signature where a method written
in terms of the underlying values does not.
|
|
|
|
|
| |
Looks like multiple, different fixes for this issue managed to break it
again.
|
|\
| |
| | |
SI-7704 Fix partest's test category selection
|
| |
| |
| |
| | |
... not Unit now.
|
| | |
|
| |
| |
| |
| |
| |
| | |
... and move scalap.scala.tools.scalap.scalax.rules to
scalap.scala.tools.scalap.rules now that all the Monad/Functor/
Applicatve/... stuff has been removed.
|
| | |
|
| |
| |
| |
| |
| | |
in various places. This includes actors, compiler (mostly some new
macro parts) continuations, partest, scaladoc, scalap.
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Not all test configurations handle an existing log file.
In particular, kicking off javac will append to the log file
and always fail the test.
Don't even ask how I know that.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Testicolor just passed file arguments through; this commit does
either toAbsolute or getAbsoluteFile (depending on the File API
impedance mismatch) at ConsoleRunner (command line) and Runner
(ant supplies absolute paths, but in case sbt runner doesn't,
and who knows). It also sprinkles toAbsolute internally, where
it may be redundant. Do I mean redundant or redolent?
I noticed that my branch from last year relativized to test root
always; so this is one of those choices one commits to.
The point of breakage in this case was, if a path is not absolute,
use test root / path, which only works (of course) if the relative
path is relative to test root, which is no longer the case.
Because all paths are now calculated with respect to the test file
path, the "out" dir would change its relative path depending on
the invocation.
|
|/
|
|
|
|
|
|
|
|
|
|
| |
Check files split into conditional blocks with partest flags
have been broken since the new diff regime.
For some reason, no one noticed.
The clever scheme to "filter the diff" instead of just filtering
the check file is abandoned as futile and unnecessary.
Fix java6 checkfile for ifdiff fix.
|
|\
| |
| | |
SI-4594 Repl settings command
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
A settings command for the rest of us.
The usual command line options are used, except that boolean flags
are enabled with +flag and disabled with -flag.
```
scala> :settings +deprecation
scala> new BigInt(java.math.BigInteger.TEN) { }
<console>:8: warning: inheritance from class BigInt in package math is deprecated: This class will me made final.
new BigInt(java.math.BigInteger.TEN) { }
^
res0: BigInt = 10
scala> :settings -deprecation
scala> new BigInt(java.math.BigInteger.TEN) { }
res1: BigInt = 10
```
Multivalue "colon" options can be reset by supplying no values
after the colon. This behavior is different from the command line.
```
scala> 1 toString
warning: there were 1 feature warning(s); re-run with -feature for details
res0: String = 1
scala> :settings -language:postfixOps
scala> 1 toString
res1: String = 1
scala> :settings
-d = .
-encoding = UTF-8
-explaintypes = false
-language = List(postfixOps)
-nowarn = false
scala> :settings -language:
scala> :settings
-d = .
-encoding = UTF-8
-explaintypes = false
-language = List()
-nowarn = false
```
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Summary:
- Remove the last vestiges of xml from Predef and Contexts.
- Change build to compile scala.xml to scala-xml.jar.
- Deploy scala-xml module to maven.
- Update partest accordingly.
Note:
An older compiler cannot use the new standard library to
compile projects that use XML. Thus, skipping locker will
break the build until we use 2.11.0-M4 for STARR.
In the future build process, where we drop locker,
we would have to release a milestone that supports the old
and the new approach to xml. As soon as we'd be using that
new milestone for starr, we could drop support for the old approach.
|
|\ \
| | |
| | | |
Merge 2.10.x into master
|
| |\ \
| | |/
| |/|
| | |
| | |
| | | |
Conflicts:
src/compiler/scala/tools/nsc/typechecker/Namers.scala
src/partest/scala/tools/partest/PartestTask.scala
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Throwing a BuildException is the polite way to fail the Ant build.
Before:
BUILD FAILED
/Users/jason/code/scala2/build.xml:1522: java.lang.RuntimeException: Test suite finished with 1 case failing:
/Users/jason/code/scala2/test/files/pos/lub-dealias-widen.scala [FAILED]
at scala.sys.package$.error(package.scala:27)
[20 lines elided]
at org.apache.tools.ant.launch.Launcher.main(Launcher.java:109)
Total time: 2 minutes 35 seconds
After:
BUILD FAILED
/Users/jason/code/scala2/build.xml:1522: Test suite finished with 1 case failing:
/Users/jason/code/scala2/test/files/pos/lub-dealias-widen.scala [FAILED]
Total time: 2 minutes 34 seconds
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Before this change s.t.n.u.CommandLine was used instead.
Remove command-line options --ansi, --buildmanager, --show-log
which don't have any effect and don't seem to be used anywhere.
|
|/ / |
|
| |
| |
| |
| |
| |
| |
| | |
tools.cmd.CommandLineParser uses a small hand-rolled parser
TODO: replace partest's usage of scala.tools.nsc.util.CommandLine
by scala.tools.cmd.CommandLine
|
|\ \
| | |
| | | |
SI-7479 Make test/files/run/tailcalls.scala pass on Avian
|
| | |
| | |
| | |
| | | |
This commit modifies the test and check file to use the new diff filter.
|
| | |
| | |
| | |
| | |
| | | |
... so that the different results of the test on Avian
can be specified in the check file.
|
|\ \ \
| | |/
| |/|
| | |
| | | |
Conflicts:
src/partest/scala/tools/partest/DirectTest.scala
|
| | |
| | |
| | |
| | |
| | |
| | | |
We had several tests designed to only run if the JDK version was at
least some specified version. This commit refactors that common logic
into DirectTest.
|
| |/
|/|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
A classfile in the wild related to Vaadin lacked the InnerClasses
attribute. As such, our class file parser treated a nested enum
class as top-level, which led to a crash when trying to find its
linked module.
More details of the investigation are available in the JIRA comments.
The test introduces a new facility to rewrite classfiles.
This commit turns this situation into a logged warning, rather
than crashing. Code by @paulp, test by yours truly.
|
|\ \
| | |
| | | |
SI-7410 REPL uses improved tools.jar locator
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The logic in partest for snooping around for tools.jar
is moved to PathResolver, and ILoop uses it from there.
If JAVA_HOME is toolless, check out java.home.
The use case was that Ubuntu installs with `java` at
version 6 and `javac` at version 7; it's easy to wind
up with JAVA_HOME pointing at the version 6 JRE, as
I discovered. It's confusing when that happens.
In future, partest might run under 7 and fork tests
under 6, but those permutations are downstream.
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Partest will also read files/filters and files/kind/filters for
filter expressions (one per line, trimmed, leading #comments)
which are taken as regexes.
A test/files/filters is provided which attempts to quell HotSpot
warnings; the test for this commit requires it.
The elided lines can be revealed using the lemon juice of verbosity:
apm@mara:~/projects/snytt/test$ ./partest --verbose --show-diff files/run/t7198.scala
[snip]
>>>>> Transcripts from failed tests >>>>>
> partest files/run/t7198.scala
% scalac t7198.scala
[snip]
% filtering t7198-run.log
--Over the moon
--Java HotSpot(TM) 64-Bit Server VM warning: Failed to reserve shared memory (errno = 28).
The filtering operation is part of the transcript, which is printed on failure.
No attempt is made to be clever about not slurping the filters file a thousand times.
Previous literal patterns had to be updated because there's parens in them thar strings.
Future feature: pattern aliases, define once globally and invoke in test filters.
|
|/ /
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Some scalac output is on stderr, and it's useful to see that
in the log file, especially for debugging.
Adds a line filter for logs, specified as "filter: pattern"
in the test source.
Backslashes are made forward only when detected as paths.
Test alignments:
Deprecations which do not pertain to the system under test
are corrected in the obvious way.
When testing deprecated API, suppress warnings by deprecating
the Test object.
Check files are updated with useful true warnings, instead of
running under -nowarn.
Language feature imports as required, instead of running under -language.
Language feature not required, such as casual use of postfix.
Heed useful warning.
Ignore broken warnings. (Rarely, -nowarn.)
Inliner warnings pop up under -optimise only, so for now, just
filter them out where they occur.
Debug output from the test required an update.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
A serious issue: partest would launch the same test multiple
times if a test path was given in multiple forms (e.g. absolute
and relative paths.) Unfortunately all those tests would share
the same logfile, output directory, etc. which would predictably
lead to explosions.
Since overwriting classfiles while being loaded can lead to
jvm core dumps, it's possible this is involved in recent jvm
crashes and other test breakdowns.
This commit also alters the default partest verbosity to only
full print test transcripts under --verbose.
|
|\ \
| | |
| | | |
SI-7494 Each plugin must only be instantiated once.
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Exploit SI-6446 fix to move old pending tests for plugins to the
active pool.
Also, nuance the partest runner to allow the compiler to crash
with a FatalError if a check file is present.
Some of the plugin tests make phase assembly throw.
One of the legacy tests was moved back to pending:
There seems to be a race for who shall be deemed truly dependent.
Back to pending for you!
apm@mara:~/projects/snytt/test$ ./partest files/neg/t7494-cyclic-dependency/
Selected 1 tests drawn from specified tests
> starting 1 test in neg
ok 1 - neg/t7494-cyclic-dependency
1/1 passed (elapsed time: 00:00:04)
Test Run PASSED
apm@mara:~/projects/snytt/test$ ./partest files/neg/t7494-cyclic-dependency/
Selected 1 tests drawn from specified tests
> starting 1 test in neg
!! 1 - neg/t7494-cyclic-dependency [output differs]
> 0/1 passed, 1 failed in neg
>>>>> Transcripts from failed tests >>>>>
> partest files/neg/t7494-cyclic-dependency
% scalac t7494-cyclic-dependency/ThePlugin.scala
% scalac t7494-cyclic-dependency/sample_2.scala
error: Cycle in compiler phase dependencies detected, phase cyclicdependency1 reacted twice!
% diff files/neg/t7494-cyclic-dependency-neg.log files/neg/t7494-cyclic-dependency.check
@@ -1 +1 @@
-error: Cycle in compiler phase dependencies detected, phase cyclicdependency1 reacted twice!
+error: Cycle in compiler phase dependencies detected, phase cyclicdependency2 reacted twice!
0/1 passed, 1 failed (elapsed time: 00:00:04)
Test Run FAILED
|
|\| |
| |/
|/|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
================================================================
Merge commit 'v2.10.1-326-g4f8c306' into merge/v2.10.1-326-g4f8c306-to-master
Conflicts:
src/compiler/scala/tools/nsc/typechecker/SuperAccessors.scala
src/reflect/scala/reflect/runtime/JavaMirrors.scala
================================================================
Merge -s ours 4e64a27 ([nomaster commit range])
================================================================
Merge commit '0ae7e55' into merge/v2.10.1-326-g4f8c306-to-master
Conflicts:
src/compiler/scala/tools/nsc/typechecker/Macros.scala
|
| |\
| | |
| | | |
Par-Test includes log in transcript of failed exec
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The previous behavior was that the Failure is generated
before the log is appended to the transcript.
That meant that the summary transcripts wouldn't include
the log file. Luckily, the transcript would say something
like "jvm > showFail-run.log".
|
| |/
| |
| |
| |
| | |
A flag is set in NestUI to look for "diff" in the transcript
of failing tests.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
A compiler crash does not count as Does Not Compile for
purposes of a negative test. Changing the test kind from
"neg" to "dnc" is out of scope for this PR.
Even if the user asks to update the check file with the
crash result, we must prevent him from doing so. Any
further attempts to update the check file with a crash
will dispatch the Scala SWAT squad which will race to
his location and physically restrain the user. Only
Martin holds the code which will allow the squad to
stand down. So make sure his cell is on next time you
want to --update-check.
A neg test will stop trying to compile after the first
failed round, which is all that matters.
By popular request, a new test outcome is emitted when
the check file is updated. It is called "Updated". It
even has its own short status, the double-plus that is
reminiscent of diff output.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Partest --timeout "30 seconds" to time out the test run.
It will not hang on timeout ("I hang with Par-Test" t-shirts
not-withstanding).
It's beyond the scope of this commit to investigate argument
parsing:
`partest --timeout "\"30 seconds"\" --pos`
|
| |
| |
| |
| |
| |
| |
| |
| | |
Partest uses test-interface API to invoke ScalaCheck. This obviates
ad hoc output checking for result status.
The context class loader is set to a loader that the scaladoc scalacheck
tests can use.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Allow the hyphens to line up on a column within the output
for a test category or "kind".
```
$ partest files/neg/t4134.scala --run
Selected 1348 tests drawn from 1 named test categories, specified tests
& starting 1 test in neg
ok 1 - neg/t4134.scala
& starting 1347 tests in run
ok 1 - run/absoverride.scala
ok 2 - run/amp.scala
ok 3 - run/adding-growing-set.scala
```
|
| |
| |
| |
| |
| |
| | |
A failing test generates a test failure.
The failure log is added to the transcript.
The log stream is closed when done.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Since srcDir is no longer used to extract the test kind from any
given test path, it is especially benign to rely on the defness
of srcDir in order to allow redefinition during an Ant run.
The parameter is configured via the partest.srcdir property.
Maybe the keyword should be def for definite values and
red for redefinable values and ind for indefinite values.
And ded for dedefinable values optionally reset to None.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This commit lets partest locate tools.jar the way REPL does, with
the addition that java.home.parent is also tried.
The partest script will use JAVAC_CMD if set, or else JAVA_HOME, and
will try the sibling of JAVACMD if set (on the theory that if you specify
java, you are avoiding the path lookup and javac may also be in
that special place), or else query the path for javac.
The use cases are: no env vars, look around java.home; JDK or JAVA_HOME is
set; JAVACMD is set; and finally tools.jar can live in jre/lib/ext and
the fallback deep search will find it.
These cases have been tried on cygwin with Java installed under
s"Program${space}Files", which is usually the most brittle environment.
That means tested with bash.
The windows partest.bat has not been upgraded or side-graded.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This is Paul's test demonstrating that Javac errors are
correctly transcribed in the test transcript.
A gratuitous Scala class is added to a later round to
show that the test halts after the first error.
The runner must supply absolute paths to javac so that
absolute paths are reported in errors and stripped away
by partest.
The check file is differentiated for Java 6 and 7,
and partest's runner will now post-process the
`diff log check` to strip the diff which does not apply.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Paulptest includes color and simplified test scarfing.
Scalap tests are moved to the conventional name and location.
Testicolor missed out on Josh Suereth's tweak to sort the files in
a compilation round. Restore sortiness to test sources.
Testicolor is due to one of Paul's branches on a timeline
that apparently did not include the destruction of planet
Earth and its colonies by the Xindi.
Thanks also to Szabolcs Berecz for his merge effort. Merging
is thankless work, but not as thankless as merging in a
timeline that actually does terminate in the destruction of
your home world and Enterprise.
Archer had a supremely difficult choice: rescue humanity or
live out his retirement with T'Pol waiting on him hand and
foot? I'm sure I don't know how I'd choose.
|