Commits on Oct 29, 2013
  1. @dragos

    Merge pull request #567 from dragos/improvement/quick-fix-return-type

    'Add explicit type' quick assist.
    dragos committed Oct 29, 2013
Commits on Oct 28, 2013
  1. @dragos

    Merge pull request #578 from uqbar-project/match-error-on-watch-expre…

    …ssion-1001933
    
    Fixing MatchError on Watch Expression View by defining a default icon.
    dragos committed Oct 28, 2013
Commits on Oct 25, 2013
  1. @nscarcella
  2. @dragos

    Merge pull request #577 from uqbar-project/new-source-folder-menu-100…

    …1577
    
    [1001577] - Added source folder wizzard in Scala perspective menu
    dragos committed Oct 25, 2013
Commits on Oct 23, 2013
  1. @nscarcella

    Added source folder wizard in Scala perspective menu

    Fix #1001577
    Review by @huitseeker
    nscarcella committed Oct 19, 2013
Commits on Oct 22, 2013
  1. @dragos
Commits on Oct 18, 2013
  1. @dragos

    Merge pull request #565 from dragos/improvement/scala-osgi

    Switch to vanilla Scala OSGi bundles
    dragos committed Oct 18, 2013
  2. @dotta

    Merge pull request #575 from dotta/issue/semantic-highlight-class-cas…

    …t-exception-1001925-bis
    
    `ScalaClassFileEditor` should always hold a `ScalaClassFile`
    dotta committed Oct 18, 2013
  3. @dotta

    `ScalaClassFileEditor` should always hold a `ScalaClassFile`

    There doesn't seem to be a good reason for not creating a `ScalaClassFile` when
    source attachments cannot be located.
    
    Doing so has two nice consequences:
    
    1) When sources are manually attached to a Scala JAR, and a classfile is opened
       in an editor, the source is now correctly syntactically highlighted. This
       basically fix half of ticket Re #1001926 (the other half is triggering
       semantic highlighting).
    
    2) `ScalaClassFileEditor` instances are now ensured to be always created with
       an underlying `ScalaClassFile`. This is an important fact, as
       `ScalaClassFileEditor` assumes it (in fact, the `ClassCastException`
       reported in ticket Re #1001925 was happening because this implicit contract
       was broken, i.e., it was possible that a `ScalaClassFileEditor` could be
       created with a (Java) `ClassFile` when no source attachment was found.
    
    By the way, the `ClassCastException` issue mentioned in point 2) could be
    reproduced by copying a **Scala** classfile in the root of a project, and then
    trying to opening (by double-clicking on it from the Package Explorer view)
    would result in an error prompted to the user.
    
    Fixes #1001925
    dotta committed Oct 17, 2013
  4. @dotta

    Merge pull request #576 from dotta/transient/remove-changelog

    Removed old and no longer used CHANGELOG file
    dotta committed Oct 18, 2013
  5. @dotta
  6. @dragos

    Reviewer comments:

    - added Scala compiler sources
    - added xml and combinator versions only in the 2.11 profile
    - removed useless dependency on HasLogger
    - removed exported dependencies from sdt.core project files
    - removed superfluous version numbers in dependencies
    - fixed runtime library for build-tools to always be 2.10
    dragos committed Oct 16, 2013
  7. @dragos

    Use plain Scala OSGi bundles to build the IDE.

    * toolchain is used only to build sbt
    * all Scala-related dependencies are resolved through Maven
    * bundles not needed for execution (scala-swing, continuations, various source attachments) are added to sdt.core
    * features need an extra step, because in 2.11 we need to ship two extra bundles (xml and parser combinators). This is done through the new -Pset-features profile in build-all.sh
    
    As a bonus, all Scala bundles have correct source attachments now, so plugin
    development should be much easier.
    
    Fixed #1001889.
    dragos committed Feb 1, 2013
  8. @dotta

    Merge pull request #573 from dotta/transient/remove-scalastyle-warnings

    Removed whitespaces from test sources to make Scalastyle happy
    dotta committed Oct 18, 2013
Commits on Oct 17, 2013
  1. @dotta

    Merge pull request #568 from dotta/issue/import-completion-for-module…

    …-methods
    
    Issue/import completion for module methods
    dotta committed Oct 17, 2013
  2. @dotta
  3. @dotta

    Don't add parameter list on import completion for module's methods

    Also, added ignored test for Re #1001919
    
    Fixes #1001125
    dotta committed Oct 14, 2013
  4. @dotta

    Splitted side-effecting logic from pure computation in `ScalaCompleti…

    …ons`
    
    Created two nested method `typeCompletionsAt` and `scopeCompletionsAt` to
    clearly separate the side-effecting logic from the pure computation in
    `ScalaCompletions`.
    dotta committed Oct 14, 2013
  5. @dotta
  6. @dotta

    Merge pull request #571 from dotta/issue/do-not-return-some-of-null

    `ScalaPresentationCompilerProxy.apply` should never return `Some(null)`
    dotta committed Oct 17, 2013
  7. @dotta

    Merge pull request #569 from dotta/issue/wrong-hyperlinking-1001921

    Test demonstrating wrong hyperlinking on metod w/ default arg
    dotta committed Oct 17, 2013
  8. @dotta

    `ScalaPresentationCompilerProxy.apply` should never return `Some(null)`

    The former implementation of `ScalaPresentationCompilerProxy.apply` could
    return `Some(null)` if the passed `op` returned `null`. This was not the
    intended semantic.
    
    A different issue I'm thinking about is how I can provide an overload of
    `ScalaPresentationCompiler.apply` to allow to pass either an `op` of type
    `ScalaPresentationCompiler => U` or `ScalaPresentationCompiler => Option[U]`.
    The simplest way to do so would be to have two different methods, but I want to
    think about this a bit more (in fact, I'm still not convinced we shouldn't
    simply keep the current implementation).
    dotta committed Oct 17, 2013
Commits on Oct 16, 2013
  1. @dotta

    Merge pull request #570 from dotta/wip/enable-retry-flaky-test

    `retryFlakyTest` now enabled by default
    dotta committed Oct 16, 2013
  2. @dotta

    `retryFlakyTest` now enabled by default

    As the last failures of the PR job validator demonstrate, it's quite easy to
    get several flaky builds in a row. This is painful to a degree that cannot be
    tolerated anymore, which is why I'm enabling the `retryFlakyTest` flag by
    default, and you should explicitly pass `-DretryFlakyTests=false` to the build
    if you don't want known flaky test to be re-run.
    
    By the way, I'm not removing the `retryFlakyTests` profile because the scala
    validation script uses it, and this way we can easily change again the default
    of `retryFlakyTest` without having to go and update the validation script.
    dotta committed Oct 16, 2013
  3. @dotta

    Test demonstrating wrong hyperlinking on metod w/ default arg

    This test demonstrates that the Tree returned by `askTypeAt` is wrong when the
    request is executed *after* fully-typechecking the sources used for the test.
    
    I've debugged the problem inside the Scala IDE, and the reason why this happen
    is that `askTypeAt` returns the enclosing `foo` DefDef method symbol, instead
    of the expected Select `Bar.bar` node. The enclosing method is returned becuase
    all Trees inside the method block have a *TransparentPosition*, all the way
    down. The impelmentation of `askTypeAt` relies on
    `RangePositions.Locator.traverse` to return the deepest, non-transparent, tree
    that includes the passed position. And, because all positions are transparent
    inside the `foo` method's body, then the `foo` symbol is returned.
    
    An additional interesting fact is that it looks like one of the transformation
    happening during typechecking is to blame for updating the Tree with the wrong
    kind of positions. In fact, if the source file isn't typechecked, the correct
    symbol is returned.
    
    Enable the test once https://issues.scala-lang.org/browse/SI-7915 is fixed
    
    Re #1001921
    dotta committed Oct 16, 2013
Commits on Oct 12, 2013
  1. @dragos

    Merge pull request #496 from danchia/fix-autocompletion-inside-method…

    …-t1001218
    
    Fix autocompletion inside method t1001218
    dragos committed Oct 12, 2013
Commits on Oct 11, 2013
  1. @dotta

    Merge pull request #566 from dotta/wip/add-cleanup-to-flakytest-retry

    Add a `cleanUp` argument to the `FlakyTest.retry` combinator
    dotta committed Oct 11, 2013
  2. @dotta

    Add a `cleanUp` argument to the `FlakyTest.retry` combinator

    As evinced by the recent failures of the `pr-scala-integrate-ide` job, before
    re-running a test we need to make sure that we clean-up any side-effect created
    by the previous test run.
    dotta committed Oct 11, 2013
  3. @danchia

    add scaladoc

    danchia committed Oct 11, 2013
  4. @danchia

    remove closure

    danchia committed Oct 11, 2013
  5. @danchia

    incorporating some of @dragos suggestions

    avoid name duplication in contextType
    remove side-effecting
    try and use scope
    danchia committed Oct 11, 2013
  6. @dragos

    Merge pull request #564 from dragos/issue/remove-scala-dbc

    Removed references to scala-dbc.
    dragos committed Oct 11, 2013
Commits on Oct 10, 2013
  1. @danchia

    fix bug with Apply type completion

    dotta found this, where we're still using prefix matching for function
    applies, where we really shouldn't since we already know the function
    name.
    danchia committed Oct 10, 2013
  2. @danchia

    add additional method to test file

    to make sure we don't regress by providing a red herring for completion
    danchia committed Oct 10, 2013
  3. @dragos

    Removed references to scala-dbc.

    Scala-dbc was not released since 2.9.
    dragos committed Oct 10, 2013