[jruby-joni] annotated tag joni-2.1.0 created (now f50dd28)

Hideki Yamane henrich at moszumanska.debian.org
Mon Nov 16 11:30:07 UTC 2015


This is an automated email from the git hooks/post-receive script.

henrich pushed a change to annotated tag joni-2.1.0
in repository jruby-joni.

        at  f50dd28   (tag)
   tagging  a88fa88fa1e3c89a639efb74d86956056b49dd93 (commit)
 tagged by  Charles Oliver Nutter
        on  Thu Oct 31 11:17:31 2013 -0500

- Log -----------------------------------------------------------------
[maven-release-plugin]  copy for tag joni-2.1.0

Ben Browning (8):
      Add ALLOW_NESTED_REPEAT syntax option
      Add IGNORE_BACKREF_PROC_READ_NOT syntax option
      Move the JavaScript syntax from DynJS to Joni
      Modify Analyser to respect the STRICT_CHECK_BACKREF syntax option
      Add OP2_ESC_IDENTITY syntax option
      Add OP2_UNICODE_WHITESPACE syntax option
      Add ALLOW_EMPTY_CC syntax option and enabled for JavaScript
      Squash all the JS options to OP2_OPTION_ECMASCRIPT

Charles Oliver Nutter (5):
      Fix for JRUBY-4382. The fix is to ensure that a region never walks off the end of the buffer by adding a range check in opCClassNot and opCClass, similar to MRI's regex.c line 3914 on branch ruby_1_8_7 at 26113.
      Add some convenience constructors for byte[], CharSequence, and String
      Bump version to 1.1.7.
      Switch to Sonatype release process.
      [maven-release-plugin] prepare release joni-2.1.0

Douglas Campos (6):
      don't skip non-ANYCHARs when .* fails to match.
      generate sources during build
      generate javadocs during build
      add missing maven plugin versions
      bump to development version
      pin source encoding to UTF-8

Joel Hockey (2):
      whitespace change only
      Moved compile before setting regex.stackPopLevel to fix http://jira.codehaus.org/browse/JRUBY-5340.  This now matches code in oniguruma.

Jordan Sissel (2):
      Make the regexp engine respect via Thread.interrupt().
      Add InterruptedException awareness to the tests.

Marcin Mielzynski (67):
      Bump up asm for joni
      Update groupId for asm
      Fix unused option
      commented code cleanup
      fix for JRUBY-6414 - YAML.load cause errors in particular string
      Sync joni with oniguruma 5.9.2
      regression test for JRUBY-6414
      Update for character class node
      Bump dependency version for jcodings and support nonunicode \s \d \w
      Turn off failing tests for non unicode character class mode for \s \d \w
      Remove blank lines, doh
      org.joni.ast.ConsAltNode cannot be cast to org.joni.ast.StringNode issue
      Enable non unicode SWD
      Bump version to 1.1.7.
      Beyond string check goes first.
      Add test case for opPushIfPeekNext.
      Fix \k and \g IOOBEs and add warnings
      Better warnings.
      Refactor lexer switches.
      More lexer refactoring.
      Minor changes.
      Fix escape for 8 and 9.
      Better AST debug repr.
      Fix look behind.
      Simplify.
      Simplify.
      Less verbose test output.
      Anchor in disable map.
      split noNameDisableMap.
      cleanup.
      Fix char class.
      Experimental optimization by using string templates outside of bytecode
      Make OPCode.EXACTN_IC, OPCode.EXACTN_IC_SB use string templates as well.
      Test case to cover OPCode.EXACTMBN opcode.
      cleanup.
      cleanup.
      cleanup.
      cleanup.
      Make EXACTMBN opcode use string tamplates.
      Make ExactMB3N opcode use string templates.
      Optimize node lower-casing for single byte and do not copy template if
      Add a test for utf-8 as well.
      Lower indirection when compiling.
      Remove duplication.
      fix wrong templateNum
      Add a test for expandCaseFoldString.
      Fix EXACTMBN opcode selection and add a test.
      Fix StringNode only casing and add a test case.
      Fix all remaining cases from test_regexp.rb/test_unicode
      Be more thrifty wrt code buffer.
      Give more proper name.
      Do not create new String nodes on end of token.
      Ditto for interval token.
      Reuse code.
      Reuse.
      remove those pesky 'hidden' vars.
      Remove these pesky arrays.
      Ditto.
      Ditto.
      Rename fields before refactoring.
      Add a test.
      Make the test more explicit.
      Doh, comment out, since it's not reliable across platforms.
      Merge branch 'master' of https://lopex@github.com/jruby/joni.git
      Fix for JRUBY-6682 (jgem search -r ak ' causes exceptions.)
      Fix for JRUBY-6923 - UTF-8 string is not matching //i properly
      Tests for JRUBY-6923 - UTF-8 string is not matching //i properly

Marcin Mielżyński (47):
      Fix bytecode printer to support some missing singlebyte bytecodes.
      Allocate stack frames lazily since more likely weak referenced stack will go away.
      Lower default stack size to 64 entries.
      Fix wrong method invocation (joni trunk).
      Add userOptions flags to joni Regex instance.
      Fix for JRUBY-2113: Regexp behavior differs a bit from MRIs.
      Fix debug output for some look behind operations.
      Null Regexp.warnings field so it doesn't contain hard reference to the runtime.
      Expose more API for 1.9 functionality.
      Fix for JRUBY-2089: ArrayIndexOutOfBoundsException when evaluating regex with trailing \b.
      Remove unused code.
      Add Encoding.load for lazy kcode and lazy 1.9 encoding support.
      Prevent int/byte conversion.
      Quite a bit faster OPCode.JUMP bytecode implementation.
      Remove String.isEmpty since it's since jdk6.
      Merge asmified branch into joni trunk. The asm compiler is not yet implemented (only parts of asm infrastructure exist now) though the original one is abstracted via visitor pattern and won't be affected by asm compiler work (it won't even touch asm classes). This branch also includes lots of performance improvements and memory savings already, also code ranges are now loaded in thread safe way. The main reason for the merge are 1.9 oniguruma changes that need to be incorporated int [...]
      Big step on oniguruma/1.9 changes chase. Added length sanity checks for most encodings (all joni internals synced to length(byte[]bytes, int p, int end)) and a couple of additional encodings. Lots of megamorphic callsites removal. VANILLA flag introduced which indicates pure oniguruma or the one with 1.9 modifications.
      Some character length routines micro-optz.
      Add Emacs-Mule encoding impl.
      Make joni unicode tests buildable.
      Remove a tab.
      Apply joni changes after encoding framework extraction and update the version to 1.1.0.
      Remove unused path
      1.9 doesn't use capture histories.
      Sync joni up with jcodings changes and fix opEndLine and opSemiEndLine as well.
      String 'end' aware case insensitive multibyte search routines.
      Yikes, how is it possible this haven't shown up yet...
      Fix for not working numbered backreferences with \k syntax.
      update Joni version to 1.1.1
      Fix for JRUBY-3141: Array index error in JCodings running latest String#rindex specs.
      Log also dMin and dMax.
      Make tests using invalid EUC-JP characters only run in VANILLA mode.
      Always set msaBegin and msaEnd so we won't have to check against lazy captures in consumer code.
      Allow "invalid" quantifiers in non-vanilla mode.
      new joni version.
      Remove adjustStrPosition.
      Remove two fields from Matcher.
      Reduce array lookups.
      Introduce Regex.userObject.
      Fix for JRUBY-2603: Couple of new rubyspec failures for URI library. (Bad optimization info was emitted from ast for PREC_READ anchors.
      Bump Joni version.
      joni 1.1.3
      Merge pull request #4 from qmx/fixes
      Merge pull request #5 from qmx/fixes2
      Create README.md
      Update README.md
      Update README.md

Michael S. Klishin (2):
      Add .travis.yml
      Add a note about why email notifications are disabled (for now)

Nick Sieger (13):
      Creating trunk/tags structure
      Move src to trunk/src
      Move test to trunk/test
      Move pom
      Move build.xml
      Move MANIFEST.MF
      Merge pom change from 1.0
      Change joni maven group id to org.jruby.joni for now
      JRUBY-1941: Wrap WeakReference around stuff stored in ThreadLocals
      Bump Joni version to 1.0.1 for release
      Bump Joni to 1.0.2 on trunk
      Bump Joni to 1.1.5
      Merge pull request #3 from michaelklishin/master

Thomas E Enebo (1):
      Merge pull request #6 from jordansissel/respect-thread-interrupt

Thomas E. Enebo (10):
      Bump pom to use proper version of jcodings
      Bump version for updated jcoding dependency
      Bump to 1.1.6
      Bump for new jcodings and next jnoi version
      Update for new release
      Make *Interruptible methods which can throw and have ordinary methods return a specific FAIL value.  In this new world -1 is failed match and -2 is interrupted evaluation.  Using public mnemonics to discourage C-style <0 checks
      Make interrupt check actually reset the interrupt state so the thread can work for more than one regexp.  Add some unit tests
      Add working non-interruptible test case to make sure it return Matcher::INTERRUPTED
      Update to reflect major API change (search() and match() can now return -2 if interrupted in addition to searchInterruptible and matchInterruptible
      MOAR DIGITS

Thomas Enebo (1):
      Update pom to reflect current release of jcodings

-----------------------------------------------------------------------

No new revisions were added by this update.

-- 
Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/pkg-java/jruby-joni.git



More information about the pkg-java-commits mailing list