Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP

Loading…

please see http://jira.codehaus.org/browse/JRUBY-5848 for details #279

Closed
wants to merge 17 commits into from

7 participants

@mkristian
Collaborator
  • removed all these provided deps since they are not needed for building the jar and not needed anywhere else
  • let jruby-core depend on jruby-stdlib - just easier to setup maven with it
  • fixed the deps of jruby-core
  • finally cleanup the poms and its hierachy - the only pom with deps is no jruby-core
  • added simple integraton-test to jruby-core to ensure this feature to stay alive ;)
mkristian added some commits
@mkristian mkristian * moved version to preview3
        * removed all these provided deps since they are not needed for building the jar and not needed anywhere else
        * let jruby-core depend on jruby-stdlib - just easier to setup maven with it
        * fixed the deps of jruby-core
        * finally cleanup the poms and its hierachy - the only pom with deps is no jruby-core
        * added simple integraton-test to jruby-core to ensure this feature to stay alive ;)
1b6eaab
@mkristian mkristian use the snapshots until released to maven-central 403db78
@travisbot

This pull request passes (merged 1b6eaab into b37a0bd).

@travisbot

This pull request passes (merged 403db78 into b37a0bd).

@tychobrailleur
Collaborator

Awesome cleanup. Hopefully the jnr-* jars will be released too, that will save a lot of the hassle of having to recompile most of them locally.

@mkristian
Collaborator

yes - I am really waiting to have a working jruby-core artifact, it can be used in jbundler to resolve jar dependencies more consistently and/or warn when there are duplicated jars in the classloader.

maybe, I should mention that leaving the provided away is totally save:
http://maven.apache.org/guides/introduction/introduction-to-dependency-mechanism.html#Dependency_Scope

   provided . . . .This scope is only available on the compilation and test classpath, and is not transitive.

and jruby does not use the compilation and test classpath instead jruby uses ant to do the compiling the code.

headius and others added some commits
@headius headius Fix JRUBY-6877
cext object.cpp and struct.cpp require including alloca.h on Solaris

Patch by Clayton O'Neill.
c51313b
@headius headius Make values final and eliminate some unnecessary allocations. fea744f
@headius headius Fix JRUBY-6578: File.readlink with chdir
Patch by Benoit Daloze <eregontp@gmail.com>
af8a0ab
@headius headius Fix JRUBY-6568
A Proc on a symbol like (&:first) works in 1.8, but not in 1.9 mode

In 1.9 mode, we do a better job of yielding the *actual* argument
given ([0] in this case) rather than always doing the extraneous
boxing that 1.8 mode does. However, we did not implement those
yieldSpecific paths in the Symbol#to_proc proc, so they fell back
on the old 1.8 logic that did arg list unwrapping. Implementing
the specific paths appears to make 1.9 mode yield to procified
Symbols properly.
7098c77
@headius headius Fix JRUBY-5674: Difference in include behaviour in --1.9 mode
The problem was not actually a difference in include behavior;
the issue was that we were always defining a top-level Generator
class that interfered with the lookup of the user's library's
Generator class. Moved Generator to JRuby::Generator to avoid the
namespace pollution.
5e461a9
@donv donv * Updated Eclipse settings to Eclipse 4.2
* Ignore missing serializable id since it is mostly missing and we don't care.
87bed60
@donv donv * Remove unused import to remove compiler warning in Eclipse 8adac7f
@headius headius Fix #274 (second part)
Our Java-based #inject block did not report it wanted any args,
and it seems we were not giving it any. Modified to be "OPTIONAL"
and all cases seem to work.
713f911
@kares kares no need to report memory/stack as properties
as they are resolved from RuntimeMXBean (see 5d86783e2db0879a13feb6037af0891eb24785e5)
c92f837
@headius headius Update to working i386 Windows jffi. c0cea77
@enebo enebo toString should work under normal conditions 9ca32ad
@headius headius Modifications to bring our RubyGems more in line with original.
* Removed Maven support.
* Have 'gem' just load 'jgem' in the same dir, since that's the
one that RubyGems updates.
8f8fa33
@headius headius Fix shebang. 2e34036
@mkristian mkristian get openssl working and exclude original of repacked classes b03b534
@mkristian mkristian added the last missing bits to jruby-core 108b6fc
@headius
Owner

I think we took care of all this in other commits.

@headius headius closed this
@mkristian
Collaborator
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Commits on Sep 3, 2012
  1. @mkristian

    * moved version to preview3

    mkristian authored
            * removed all these provided deps since they are not needed for building the jar and not needed anywhere else
            * let jruby-core depend on jruby-stdlib - just easier to setup maven with it
            * fixed the deps of jruby-core
            * finally cleanup the poms and its hierachy - the only pom with deps is no jruby-core
            * added simple integraton-test to jruby-core to ensure this feature to stay alive ;)
  2. @mkristian
Commits on Sep 14, 2012
  1. @headius @mkristian

    Fix JRUBY-6877

    headius authored mkristian committed
    cext object.cpp and struct.cpp require including alloca.h on Solaris
    
    Patch by Clayton O'Neill.
  2. @headius @mkristian
  3. @headius @mkristian

    Fix JRUBY-6578: File.readlink with chdir

    headius authored mkristian committed
    Patch by Benoit Daloze <eregontp@gmail.com>
  4. @headius @mkristian

    Fix JRUBY-6568

    headius authored mkristian committed
    A Proc on a symbol like (&:first) works in 1.8, but not in 1.9 mode
    
    In 1.9 mode, we do a better job of yielding the *actual* argument
    given ([0] in this case) rather than always doing the extraneous
    boxing that 1.8 mode does. However, we did not implement those
    yieldSpecific paths in the Symbol#to_proc proc, so they fell back
    on the old 1.8 logic that did arg list unwrapping. Implementing
    the specific paths appears to make 1.9 mode yield to procified
    Symbols properly.
  5. @headius @mkristian

    Fix JRUBY-5674: Difference in include behaviour in --1.9 mode

    headius authored mkristian committed
    The problem was not actually a difference in include behavior;
    the issue was that we were always defining a top-level Generator
    class that interfered with the lookup of the user's library's
    Generator class. Moved Generator to JRuby::Generator to avoid the
    namespace pollution.
  6. @donv @mkristian

    * Updated Eclipse settings to Eclipse 4.2

    donv authored mkristian committed
    * Ignore missing serializable id since it is mostly missing and we don't care.
  7. @donv @mkristian
  8. @headius @mkristian

    Fix #274 (second part)

    headius authored mkristian committed
    Our Java-based #inject block did not report it wanted any args,
    and it seems we were not giving it any. Modified to be "OPTIONAL"
    and all cases seem to work.
  9. @kares @mkristian

    no need to report memory/stack as properties

    kares authored mkristian committed
    as they are resolved from RuntimeMXBean (see 5d86783e2db0879a13feb6037af0891eb24785e5)
  10. @headius @mkristian

    Update to working i386 Windows jffi.

    headius authored mkristian committed
  11. @enebo @mkristian

    toString should work under normal conditions

    enebo authored mkristian committed
  12. @headius @mkristian

    Modifications to bring our RubyGems more in line with original.

    headius authored mkristian committed
    * Removed Maven support.
    * Have 'gem' just load 'jgem' in the same dir, since that's the
    one that RubyGems updates.
  13. @headius @mkristian

    Fix shebang.

    headius authored mkristian committed
  14. @mkristian
  15. @mkristian
Something went wrong with that request. Please try again.