| Age | Commit message (Collapse) | Author |
|---|
|
content' signal; now you need an explicit directive to allow this behaviour.
[configuration] added tests for @Accessors configuration key implementations.
|
|
cases, and added first tests to test implementations of configuration-based stuff (FlagUsages, specifically). Error reporter has been refactored.
|
|
now can start with directives (configuration keys, 'ignore', and java language version limits), and the test directory runner now has a framework to handle these. All existing tests pass at this point.
|
|
which have become more important as input files now may need to set up some lombok configuration stuff.
|
|
|
|
@Value's final-by-default and private-by-default with it; now
appropriate warnings are emitted)
* Builder now errors out on presence of most lombok annotations on an
explicit builder class.
* Builder now takes @FieldDefaults/@Value into account.
* Builder on type now generates the constructor as package private
instead of private to avoid synthetic accessor constructors.
* added a bunch of test cases.
* added a test case feature: If the expected file is omitted entirely
but there are expected messages, the differences in the output itself
are ignored.
* streamlined checking for boolean-ness (removed some duplicate code)
* added 'fluent' and 'chain' to @Builder.
|
|
Eclipse breaks the javadoc out of the raw source and we can't modify that without breaking a billion things. To solve this issue we'd have to write some very complicated patches to intercept this process and somehow propagate the node that the javadoc is attached to AND translate from the ast model to dom or whatever is being used there. Not gonna happen for this low priority feature.
|
|
and updated ECJ version detection.
|
|
AST/LombokNode. Tests updates to honour these with //version X at the top
of any test file (now also in eclipse, which until now always said it was v6)
|
|
platform and implemented it for javac BUT NOT FOR ECJ!
|
|
obviously doesn't work on windows; we force unix line ending now.
|
|
|
|
to be organization-name.jar instead of just name.jar, in order to account for
the ever lovely and wonderful apache's crazy decision to call the entirely
separate log4j v2.0 also 'log4j'. This does mean you'll have to 'ant clean'.
|
|
* re-introduction of onMethod/onConstructor/onParam
* tests checking error/warnings rewritten to be more heuristic, in order to accomodate difference in messaging between java6 and java 7
* Ability to eliminate java's own output of erroneous error messages (heh); i.e. those messages that are invalidated by lombok's actions. This mechanism is used for onMethod/onConstructor/onParam
* First steps to unifying a billion setGeneratedBy calls into a single visitor traversal for eclipse' HandleGetter/Setter/Constructor/Wither
* To simplify 'zooming in' the tests on just a few files, added an 'accept' mechanism.
* Updated copyright headers of website to 2013.
|
|
|
|
eclipse for fields with @Deprecated on them.
|
|
more text (such as the explanation for why it is being ignored).
|
|
produced errors and warnings (i.e. we intentionally compile code with errors in them to verify that the appropriate error or warning message is emitted when lombok is active during a compilation run of either javac or ecj) - instead of string comparisons, it's a little more complex. This to enable testing of both javac6 and javac7, even if they produce (slightly) different error output.
Updated all message files in the 'expected output' directories to represent this change.
|
|
in our tests, the tests are in the unfortunate scenario where we always compile against a given javac (lib/build/javac6.jar), and always run the tests against a given javac, but that javac tries to use the bootclasspath of the host JRE, and if that is JRE7, you get all sorts of errors.
I fixed it by still compiling against a given javac (we can only ship one lombok.jar after all), but having the test task run with a given bootclasspath and a given javac.jar.
There are 2 tasks that download both rt.jar and javac.jar for either OpenJDK6 or OpenJDK7, and it writes a properties file with those locations. The test task will use this property file, and explain what you need to do if it is not there.
Incidentally, this brought to light issue 422: Delombok in java7 produces VerifyErrors.
|
|
on things to a directory (use -Dlombok.tests.dump_actual_files=/path/to/dir).
Useful if you KNOW lombok is working but something changed in output, i.e. order of generated methods.
|
|
|
|
|
|
|
|
the ecj tests are written in a slightly more close-to-reality way, and do work for 'val'. eclipse test aspect removed.
|
|
through the complete compile process. This means val and @Delegate testing is now enabled.
TODO: Update -eclipse tests.
|
|
|
|
|
|
|
|
Fixes issue #241.
|
|
launch config that adds the appropriate -javaagent VM param to ensure the runWithEclipse tests work.
|
|
LoggerSlf4jWithPackage).
|
|
type-binding and so forth. Similar to eclipse, the test needs to be executed with lombok as vm-agent. @Delegate runs with this setup, sadly val does not, at least for now.
And as expected the tests shows that @Synchronized and @SneakyThrows are currently broken for eclipse but not for ecj.
|
|
runner.
|
|
new way delombok works.
|
|
classpath are relevant, and the flow needs to be parse all, then enter all, then transform all, instead of parse->enter->transform sequentially for each file in isolation.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
file to test
|
|
|
|
well as code to process entire directories at a time.
Also removed duplication from the testcases.
|
