Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with
or
.
Download ZIP
Commits on May 9, 2012
  1. @hns

    Clean up infix line number commit

    hns authored
     - Remove local variables no longer used
     - Restore setting line number for unary expression
     - Make sure infix expression always uses line number of left child
  2. @hns

    Modify rhino so that infix expressions use a starte position and line…

    Nick Santos authored hns committed
    … number that point to the start of the expression rather than the position of the operator.
    
    Modify IRFactory to maintain this new position logic when it does its translation pass.
    Adjusted the test file expectations accordingly.
    
    Committed by Mike Taylor
    -------------
    Created by MOE: http://code.google.com/p/moe-java
    MOE_MIGRATED_REVID=26391660
  3. @hns

    Fixes Issue 178 : Incorrect line number reporting

    Nick Santos authored hns committed
    When there are more than one comments for a statement, the comment node does not return the correct line number if using getLineno().
    Now use getAbsolutePosition() to figure out the correct line number
    
    Committed by Yi Zhu
    -------------
    Created by MOE: http://code.google.com/p/moe-java
    MOE_MIGRATED_REVID=26372000
  4. @hns

    Grab the jsdoc for WithStatement before we descend into its body.

    Nick Santos authored hns committed
    Committed by Nick Santos
    -------------
    Created by MOE: http://code.google.com/p/moe-java
    MOE_MIGRATED_REVID=26371560
  5. @hns

    Never let JSDoc from one statement spill into the next statement.

    Nick Santos authored hns committed
    See
    http://code.google.com/p/closure-compiler/issues/detail?id=428
    for more info
    
    Committed by Nick Santos
    -------------
    Created by MOE: http://code.google.com/p/moe-java
    MOE_MIGRATED_REVID=26339541
  6. @hns

    Parses JSDoc casting for Object Literal that happens inside ().

    Nick Santos authored hns committed
    Committed by Alan Leung
    -------------
    Created by MOE: http://code.google.com/p/moe-java
    MOE_MIGRATED_REVID=26109418
  7. @hns

    Allows Autosemicolon insertion before comments.

    Nick Santos authored hns committed
    Committed by Alan Leung
    -------------
    Created by MOE: http://code.google.com/p/moe-java
    MOE_MIGRATED_REVID=26025989
Commits on May 8, 2012
  1. @hns
  2. @hns

    Remove obsolete unicode test

    hns authored
Commits on May 7, 2012
  1. @hns

    Do not ignore UTF format characters in StringLiterals.

    Nick Santos authored hns committed
    Committed by Alan Leung
    -------------
    Created by MOE: http://code.google.com/p/moe-java
    MOE_MIGRATED_REVID=26010298
  2. @hns

    Simplify JsDoc parsing for Rhino users, attach the Comment node rather

    Nick Santos authored hns committed
    than just the JsDoc string.
    
    Committed by John Lenz
    -------------
    Created by MOE: http://code.google.com/p/moe-java
    MOE_MIGRATED_REVID=26010038
  3. @hns

    Fix -x bit on not-executable files

    Nick Santos authored hns committed
Commits on May 3, 2012
  1. @hns

    Make load() function throw errors in Rhino shell.

    hns authored
    This is fairly involved because the compilation and evaluation code used
    by load() is used by by command line file/inline script/shell processing so
    all error catching had to be moved up the call tree.
Commits on May 2, 2012
  1. @hns
  2. @hns

    JavaMembers code cleanup

    hns authored
    Use for-each loops,  static imports, replace StringBuffer with StringBuilder,
    reformatting. No functional changes.
  3. @hns

    Fix bug 749111 - RegExp literals are instantiated only once per scrip…

    hns authored
    …t/function
    
    This has the beneficial side effect of reducing function instantiation overhead and
    code size. Creating wrappers around precompiled regexps is also really cheap.
    
     The method to precompile regexps in generated classes used to be synchronized,
     now it merely uses a volatile field to check if regexps regexp initialization has already
     been done. The idea behind this is that concurrency is very unlikely and not
     terribly dangerous here.
  4. @hns

    Move code out of the main interpreter loop into dedicated methods.

    hns authored
    Main interpreter loop performance has shown to suffer from its size and monolithic
    structure. In the past, small additions to the main loop code have resulted in
    dramatic declines in performance (e.g. the regexp commit that will follow this one).
     Also see  commit 97b6d16 which dealt with
     the same problem.
    
    As it is, this commit improves interpreter performance by roughly 10% in various
    benchmarks.
Commits on Apr 26, 2012
  1. @hns

    Fix bug 509036 - ArrayIndexOutOfBoundsException thrown in regexp.Nati…

    hns authored
    …veRegExp.parseTerm when regex pattern includes a group (()) and ends with brace ({)
Commits on Apr 25, 2012
  1. @hns
Commits on Apr 23, 2012
  1. @hns
  2. @hns
  3. @hns
Commits on Apr 20, 2012
  1. @hns

    Add support for pulling out object and array creation in to separate …

    hns authored
    …methods.
    
    This helps reduce the code size for large functions and makes coffeescript
    work in compiled mode. Backport of cdc4a6d
    from rhino_1_8 branch.
  2. @hns

    Fix bug 708801: Boolean converted to number when optimising

    hns authored
    This patch from André Bargull fixes a ton of bugs in optimizer type flow analysis.
  3. @hns
Commits on Apr 18, 2012
  1. @hns

    Remove quasi-duplicate constant.

    hns authored
  2. @hns

    Avoid pushing program state in RegExp disjunction

    hns authored
    The trade-off is to add two ints for current
    continuation index and op to the backtrack object.
  3. @hns

    Optimize flat character matching

    hns authored
Commits on Apr 17, 2012
  1. @hns

    Major update of Rhino RegExp implementation.

    hns authored
    The major part of this consists of a backport of the improvements made to
    Spidermonkey's jsregexp.c code over the years (before being replaced by
    YARR in 2010), with some additional optimizations strewn in.
Commits on Mar 20, 2012
  1. @hns

    Fix bug 491621 - VariableDeclaration.toSource() sometimes returns inv…

    hns authored
    …alid source code
    
    This adds a `isStatement()` flag to VariableDeclaration. The flag is set by the Parser depending on context, which is much more reliable than the previous solution.
  2. @hns

    Fix bug 737431 - Destructuring assignment doesn't work in expression …

    hns authored
    …closures parameters.
    
    This creates a body node for expression closures, allowing extra code for destructuring parameters to be handled properly.
Commits on Feb 28, 2012
  1. @anba

    Bug 689314 - Calling toSource on a parsed script containing lambda re…

    anba authored
    …turns extra return and semi-colon
    
    Expression closures should not emit the implicit "return" statement when toSource() is called.
  2. @anba

    Bug 689308 - Calling toSource on a parsed script containing array of …

    anba authored
    …undefined elements returns smaller array
    
    Elided elements in an array literal are represented by the EmptyExpression node. And the toSource() representation of EmptyExpression is the empty string. That means elided elements may not be visible when calling toSource() on ArrayLiteral if they appear at the end of the array literal [bug 689308]. This case is now handled in AstNode#printList(). Other callers of printList() do not pass EmptyExpression nodes to the method, so special-casing EmptyExpression is safe.
  3. @anba

    Bug 688021 - Calling toSource on a parsed script containing debugger …

    anba authored
    …omits semi-colon
    
    Although the production DebuggerStatement is handled internally as KeywordLiteral, it is in fact a statement and therefore should have the same toSource() representation as other statement, i.e. it should end with a semi-colon and a newline.
  4. @anba

    Bug 688018 - Calling toSource on a parsed script containing 'void' th…

    anba authored
    …rows exception
    
    add test case for Bug688018
Something went wrong with that request. Please try again.