aboutsummaryrefslogtreecommitdiff
path: root/src/core
AgeCommit message (Collapse)Author
2014-10-24Making SCL work right is more complicated than it first seemed.Reinier Zwitserloot
Right now the rules are: * _IF_ a class is being loaded, sourced by a lombok-jar originating class, we FIRST search the lombok jar, and if we can’t find it, farm out the job to the originating equinox-side loader. * _IF_ the equinox-side loader attempts to load a class, and it does NOT start with lombok, we don’t interfere and would never serve up any content from the lombok-jar (so if we have deps, they do NOT get loaded, by design). If it DOES start with lombok, we load it, and the loading class is SCL, not the equinox-side loader. * getResource() to load classes did not work (because internally classes end in .SCL.lombok and not .class). This breaks a bunch of things. Fixed by having getResource() be aware that it should try rewriting any request for a .class to .SCL.lombok. * launchified annotationprocessor, and cleaned up the launchified agent, which now, like all other launchers, just sets up classloader stuff and then calls into the lombok loader side to finish the actual processing, instead of trying to do it itself in a handicapped environment that can’t load much.
2014-09-30first take on the shadow classloader. All seems to be in order, but we still ↵Reinier Zwitserloot
have to solve the problem with adding our shadow loader to the equinox infrastructure (solved in lombok currently by adding all of lombok to the bootclasspath), and all the public API still has to be kept as actual class files by build.xml. Currently it is all shadowed away.
2014-09-29histogram tracker now nanoTime based; lombok itself can be histogrammed now ↵Reinier Zwitserloot
with -Dlombok.histogram
2014-09-15post-release version bumpReinier Zwitserloot
2014-09-15pre-release version bumpReinier Zwitserloot
2014-09-09[config] Don't access the file system nearly every call.Roel Spilker
2014-09-09[config] trivialRoel Spilker
2014-09-09[config] Don't access the file system every call.Roel Spilker
2014-09-02post-release version bumpReinier Zwitserloot
2014-09-02pre-release version bump.Reinier Zwitserloot
2014-09-02added memoization to config lookup in eclipse. This should help make it ↵Reinier Zwitserloot
(much) faster.
2014-09-02lombok.disableConfig would cause NPEs here and there because we returned ↵Reinier Zwitserloot
null instead of an empty list for listy keys. Whoops. Spotted and fixed by github user egde, thanks!
2014-08-21[i731] javac resolution broken by 1.8.0_20; fixed with some reflection ↵Roel Spilker
voodoo. -R&R
2014-08-21[config] added histogram tracker option (-Dlombok.timeConfig) and ↵Roel Spilker
disable-entirely option (-Dlombok.disableConfig) to debug continuing speed issues with the config system, for an edge release.
2014-08-16Some workaround fix-esque work to prevent serious slowdown issues when ↵Reinier Zwitserloot
refactoring.
2014-07-01[release] post-release version bumpReinier Zwitserloot
2014-07-01[release] pre-release version bumpReinier Zwitserloot
2014-06-25[post-release-bump] belated post-release version bump.Roel Spilker
2014-06-25[issue 699] [issue 682] [issue 683] lots of fixes for eclipse/ecj based ↵Roel Spilker
issues surrounding path finding.
2014-06-24[issue 699] fileName (for finding absolute path of a source file) are ↵Roel Spilker
sometimes URIs, for example with GWT. We now do the right thing here instead of crashing out.
2014-06-10remarked out the work in progress on resolving file paths in eclipse; this ↵Reinier Zwitserloot
causes ClassNotFoundErrors in ecj mode
2014-06-10pre-release version bumpReinier Zwitserloot
2014-06-09added eclipse logging to the configsystem’s error reporter, and cleaned up ↵Reinier Zwitserloot
the changelog slightly.
2014-06-06Merge branch 'master' into configResolutionInEclipseRoel Spilker
2014-06-06more work on lombok.config resolution issues. Mostly turning outright ↵Roel Spilker
showstopping errors in workarounds and logs, because we can't reproduce any of it.
2014-06-06[i554] Another try to prevent NPEs on IBM J9Roel Spilker
2014-06-05Finished refactor of FieldAugment; there's no longer a separate variant for ↵Roel Spilker
boolean and references, and the code no longer blows up with a bunch of NPEs if you try to use the reference variant (which is now the only variant) with a primitive type. Should have zero effect on features or bugs, 100% refactor.
2014-06-04[wip][#682][#683] Research on alternate ways to find the content of ↵Roel Spilker
lombok.config given a CompilationUnitDeclaration (in eclipse).
2014-06-03bugfixed ‘print build time as part of version number if edge release’ ↵Reinier Zwitserloot
feature.
2014-06-01Fixed up and extended Tolerate with support for constructors, and added docs.Reinier Zwitserloot
2014-06-01fixed up ‘consequent boolean’ proposal with a different key name and ↵Reinier Zwitserloot
documentation.
2014-06-01Revert "Remove a wrong test for copyAnnotations returning empty array."Reinier Zwitserloot
While Maaartinus' thoughts are obviously correct (far better to use an empty array to convey 'this list-like construct has 0 items in it right now', vs. a null pointer), eclipse actually uses null-as-empty almost everywhere, and a lot of eclipse code will simply break if you don't do it right. This reverts commit 842a4759165c5cd05aae63da3921ee11a3641a4b.
2014-06-01Fix a trivial bug in generateNullCheck.Maaartin Grajcar
2014-05-31Implement Tolerate.Maaartin Grajcar
2014-05-31Create GETTER_CONSEQUENT_BOOLEAN.Maaartin Grajcar
2014-05-31Remove a wrong test for copyAnnotations returning empty array.Maaartin Grajcar
2014-05-30#682: bugfix work in progress to fix IllegalArgumentException ‘Path must ↵Reinier Zwitserloot
include project and resource name’.
2014-05-30edge releases now include the release timestamp in the ‘full version’ name.Reinier Zwitserloot
2014-05-29#683: URI not absolute errors in mvn builds due to search for lombok.config ↵Reinier Zwitserloot
being broken.
2014-05-27post-release version bumpReinier Zwitserloot
2014-05-27pre-release version bumpReinier Zwitserloot
2014-05-27Changed the lombok.nonNull.exceptionType configkey to be an enum, partly to ↵Reinier Zwitserloot
enforce only sensible exceptions and partly to make it possible to have the message be ‘x is null’ when throwing IAEs.
2014-05-23Added confkey to make @NonNull generate a different exception because of the ↵Reinier Zwitserloot
IllegalArgumentException vs. NullPointerException that we really don’t want to get into.
2014-05-22added conf keys (and implementation to make them work) for ↵Reinier Zwitserloot
lombok.getter.lazy.flagUsage and lombok.equalsAndHashCode.doNotUseGetters.
2014-05-21added a test and handler for correctly handling confkey: ↵Reinier Zwitserloot
lombok.delegate.flagUsage.
2014-05-21Added an error if @Delegate is used recursively.Reinier Zwitserloot
2014-05-20@Delegate has moved to lombok.experimental.Reinier Zwitserloot
Some work on the aliasing system to make that go smoothly.
2014-05-08bugfix: our new trick to try and disable target typing by making the LHS of ↵Reinier Zwitserloot
a 'val' statement be of an unknown type instead of j.l.Object works for Javac8, but breaks Javac6! We now detect this and pick the right strategy based on whether or not the compiler supports target typing.
2014-05-06more work on JDK8 support: val on generified lub types now works a little ↵Reinier Zwitserloot
better.
2014-05-01bugfix for errors with ecj8 and builder.Reinier Zwitserloot