Permalink
Commits on May 3, 2013
  1. Merge pull request #694 from atambo/regex_equality_with_n_option

    headius committed May 3, 2013
    Regexp #== and #eql? is true if self does not specify /n option and other does and vise versa
  2. Merge pull request #696 from atambo/regex_initialize_literal

    headius committed May 3, 2013
    Regexp#initialize should raise a SecurityError on a Regexp literal
  3. Merge pull request #693 from josedonizetti/refactoring_channelstream

    headius committed May 3, 2013
    using bytelist instead of a linkedlist for a better performance
Commits on May 2, 2013
  1. Regexp #== and #eql? is true if self does not specify /n option and o…

    atambo committed May 2, 2013
    …ther does and vise versa
  2. Merge pull request #692 from nirvdrum/reduce_class_cache_memory_usage

    headius committed May 2, 2013
    Use hash rather than full sexp for cache key to significantly reduce mem...
Commits on May 1, 2013
  1. Merge pull request #687 from atambo/time_coerce_to_int

    headius committed May 1, 2013
    Time methods should coerce the day, hour, minute, or second with #to_int
  2. Merge pull request #688 from josedonizetti/io-ungetc

    headius committed May 1, 2013
    Fixing IO.ungetc tagged specs!
  3. Raise a RangeError when codepoint is invalid for the default internal…

    tychobrailleur committed with headius May 1, 2013
    … enc.
    
    Signed-off-by: Charles Oliver Nutter <headius@headius.com>
  4. Staticize the op tables.

    headius committed May 1, 2013
  5. Merge pull request #691 from tychobrailleur/string_other

    headius committed May 1, 2013
    Implement String#to_sym special cases.
  6. Merge pull request #689 from atambo/thread_already_initialized_thread…

    headius committed May 1, 2013
    …_error
    
    Thread#initialize should raise a ThreadError when already initialized
  7. Merge pull request #674 from trejkaz/master

    headius committed May 1, 2013
    My stab at fixing JRUBY-7122
  8. Fix JRUBY-7145 by checking for drive letter in Dir.mkdir.

    headius committed May 1, 2013
    Dir.mkdir did not have any checks for the case where the path
    passed in was a drive letter, as happens when FileUtils.mkdir_p
    breaks up the path into components. It should obviously not
    attempt to make a directory for a drive letter, so I have made
    this a no-op.
  9. Merge pull request #685 from josedonizetti/supporting-more-bytes-ungetc

    headius committed May 1, 2013
    REFACTORING: supporting a list of bytes
  10. Merge pull request #683 from pfoorion/master

    headius committed May 1, 2013
    Fix unseekable stream implementation of IO.copy_stream in RubyIO.java
  11. Fixes in response to PR #684.

    headius committed May 1, 2013
    * Use code range to determine ascii rather than scanning.
    * Fix bug in EmacsMuleEncoding exposed by new 7-bit logic.
    
    We will need a jcodings release for JRuby 1.7.4.
Commits on Apr 30, 2013
  1. Merge pull request #684 from tychobrailleur/string_encode

    headius committed Apr 30, 2013
    Fix some failing rubyspecs for String#encode
  2. Fix travis failures in fileutils due to smarter tmpdir.

    headius committed Apr 30, 2013
    Travis started failing after we updated tmpdir to allow using a
    world-writable dir if sticky bit is set. The bug that was exposed
    lies in FileUtils.remove_entry_secure, where it attempts to open
    the "." entry in the target directory primarily to stat, chmod,
    and chown it. Because JVM cannot open a stream to a directory
    entry, JRuby can't do it either, and because we started allowing
    tmpdir to use sticky-bit world-writable dirs, this logic started
    getting hit for Dir.mktmpdir's cleanup.
    
    The fix was to directly stat, chown, and chmod the "." entry
    rather than opening it as a file, since these operations end up
    equivalent.
    
    Note of clarification: this issue has always existed, but most
    uses of remove_entry_secure were either already in secure dirs
    (non-world-writable) or insecure dirs with no sticky bit (in
    which case remove_entry_secure raises an error, correctly). It
    only became apparent because of the recent tmpdir change.
  3. Merge pull request #682 from josedonizetti/fixing_corrupt_char_problem

    headius committed Apr 30, 2013
    Verifying if character is corrupted, meaning that one more char is needed.
  4. Include XMLConverter.

    tychobrailleur committed Apr 30, 2013