Skip to content
Permalink

Comparing changes

Choose two branches to see what’s changed or to start a new pull request. If you need to, you can also or learn more about diff comparisons.

Open a pull request

Create a new pull request by comparing changes across two branches. If you need to, you can also . Learn more about diff comparisons here.
base repository: opal/opal
Failed to load repositories. Confirm that selected base ref is valid, then try again.
Loading
base: 2dce252fa67e
Choose a base ref
...
head repository: opal/opal
Failed to load repositories. Confirm that selected head ref is valid, then try again.
Loading
compare: 0e56d9630b81
Choose a head ref
  • 2 commits
  • 6 files changed
  • 2 contributors

Commits on May 28, 2015

  1. * Add implementation+tests for union method 

    * Change Regexp initializer to support Ruby constants (not just character flags), check for false/nil/other object in options, and deal with options properly when an existing expression is passed in
* Implement Regexp#options
* Add comment explaining why match method constructs a new RE
* Filter out encoding and extended flag related specs
* Mark hex and other character escaping as bugs (filtered)
* Update to latest RubySpecs that wrap extended in non_supported and exclude escape sequences
    @wied03
    wied03 committed May 28, 2015

    Verified

    This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
    GPG key ID: 4AEE18F83AFDEB23
    Expired
    Verified
    Learn about vigilant mode
    Copy the full SHA
    0c9b220 View commit details

Commits on May 29, 2015

  1. Merge pull request #890 from wied03/bugfix/regexp_union 

    Implement Regexp#union, Regexp#options, and proper handling of options arg in Regexp#new
    @vais
    vais committed May 29, 2015

    Verified

    This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
    GPG key ID: 4AEE18F83AFDEB23
    Expired
    Verified
    Learn about vigilant mode
    Copy the full SHA
    0e56d96 View commit details
Showing with 246 additions and 6 deletions.
  1. +110 −4 opal/corelib/regexp.rb
  2. +1 −1 spec/corelib
  3. +41 −0 spec/filters/bugs/regular_expressions.rb
  4. +90 −0 spec/filters/unsupported/regular_expressions.rb
  5. +1 −1 spec/opal/core/language/regexp_spec.rb
  6. +3 −0 spec/rubyspecs
114 changes: 110 additions & 4 deletions opal/corelib/regexp.rb
View file
Original file line number Diff line number Diff line change
@@ -1,5 +1,9 @@
class RegexpError < StandardError; end
class Regexp
`def.$$is_regexp = true`
IGNORECASE = 1
MULTILINE = 4

`def.$$is_regexp = true`

class << self
def escape(string)
@@ -23,11 +27,77 @@ def last_match(n=nil)
alias quote escape

def union(*parts)
`new RegExp(parts.join(''))`
%x{
var is_first_part_array, quoted_validated, part, options, each_part_options;
if (parts.length == 0) {
return /(?!)/;
}
// cover the 2 arrays passed as arguments case
is_first_part_array = parts[0].$$is_array;
if (parts.length > 1 && is_first_part_array) {
#{raise TypeError, 'no implicit conversion of Array into String'}
}
// deal with splat issues (related to https://github.com/opal/opal/issues/858)
if (is_first_part_array) {
parts = parts[0];
}
options = undefined;
quoted_validated = [];
for (var i=0; i < parts.length; i++) {
part = parts[i];
if (part.$$is_string) {
quoted_validated.push(#{escape(`part`)});
}
else if (part.$$is_regexp) {
each_part_options = #{`part`.options};
if (options != undefined && options != each_part_options) {
#{raise TypeError, 'All expressions must use the same options'}
}
options = each_part_options;
quoted_validated.push('('+part.source+')');
}
else {
quoted_validated.push(#{escape(`part`.to_str)});
}
}
}
# Take advantage of logic that can parse options from JS Regex
new(`quoted_validated`.join('|'), `options`)
end

def new(regexp, options = undefined)
`new RegExp(regexp, options)`
def new(regexp, options = undefined)
%x{
// Play nice with IE8
if (regexp.$$is_string && regexp.substr(regexp.length-1, 1) == "\\") {
#{raise RegexpError, "too short escape sequence: /#{regexp}/"}
}
if (options == undefined || #{!options}) {
options = undefined;
}
if (options != undefined) {
if (regexp.$$is_regexp) {
// options are already in regex
options = undefined;
}
else if (options.$$is_number) {
var result = '';
if (#{IGNORECASE} & options) {
result += 'i';
}
if (#{MULTILINE} & options) {
result += 'm';
}
options = result;
}
else {
options = 'i';
}
}
return new RegExp(regexp, options);
}
end
end

@@ -70,6 +140,7 @@ def match(string, pos = undefined, &block)
}
}
// global RegExp maintains state, so not using self/this
var md, re = new RegExp(self.source, 'gm' + (self.ignoreCase ? 'i' : ''));
while (true) {
@@ -93,6 +164,41 @@ def ~
def source
`self.source`
end

def options
# https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/RegExp/flags is still experimental
# we need the flags and source does not give us that
%x{
var as_string, text_flags, result, text_flag;
as_string = self.toString();
if (as_string == "/(?:)/") {
#{raise TypeError, 'uninitialized Regexp'}
}
text_flags = as_string.replace(self.source, '').match(/\w+/);
result = 0;
// may have no flags
if (text_flags == null) {
return result;
}
// first match contains all of our flags
text_flags = text_flags[0];
for (var i=0; i < text_flags.length; i++) {
text_flag = text_flags[i];
switch(text_flag) {
case 'i':
result |= #{IGNORECASE};
break;
case 'm':
result |= #{MULTILINE};
break;
default:
#{raise "RegExp flag #{`text_flag`} does not have a match in Ruby"}
}
}
return result;
}
end

alias to_s source
end
2 changes: 1 addition & 1 deletion spec/corelib
Submodule corelib updated from 68beb9 to 9f35bb
41 changes: 41 additions & 0 deletions spec/filters/bugs/regular_expressions.rb
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,41 @@
opal_filter "regular_expressions" do
fails "Regexp.new works by default for subclasses with overridden #initialize"
fails "Regexp.new given a String raises a RegexpError when passed an incorrect regexp"
fails "Regexp.new given a String with escaped characters raises a RegexpError if \\x is not followed by any hexadecimal digits"
fails "Regexp.new given a String with escaped characters accepts '\\C-\\n'"
fails "Regexp.new given a String with escaped characters accepts '\\C-\\t'"
fails "Regexp.new given a String with escaped characters accepts '\\C-\\r'"
fails "Regexp.new given a String with escaped characters accepts '\\C-\\f'"
fails "Regexp.new given a String with escaped characters accepts '\\C-\\v'"
fails "Regexp.new given a String with escaped characters accepts '\\C-\\a'"
fails "Regexp.new given a String with escaped characters accepts '\\C-\\e'"
fails "Regexp.new given a String with escaped characters accepts '\\c\\n'"
fails "Regexp.new given a String with escaped characters accepts '\\c\\t'"
fails "Regexp.new given a String with escaped characters accepts '\\c\\r'"
fails "Regexp.new given a String with escaped characters accepts '\\c\\f'"
fails "Regexp.new given a String with escaped characters accepts '\\c\\v'"
fails "Regexp.new given a String with escaped characters accepts '\\c\\a'"
fails "Regexp.new given a String with escaped characters accepts '\\c\\e'"
fails "Regexp.new given a String with escaped characters raises a RegexpError if more than six hexadecimal digits are given"
fails "Regexp.new given a String with escaped characters accepts '\\M-\\n'"
fails "Regexp.new given a String with escaped characters accepts '\\M-\\t'"
fails "Regexp.new given a String with escaped characters accepts '\\M-\\r'"
fails "Regexp.new given a String with escaped characters accepts '\\M-\\f'"
fails "Regexp.new given a String with escaped characters accepts '\\M-\\v'"
fails "Regexp.new given a String with escaped characters accepts '\\M-\\a'"
fails "Regexp.new given a String with escaped characters accepts '\\M-\\e'"
fails "Regexp.new given a String with escaped characters accepts '\\M-\\C-\\n'"
fails "Regexp.new given a String with escaped characters accepts '\\M-\\C-\\t'"
fails "Regexp.new given a String with escaped characters accepts '\\M-\\C-\\r'"
fails "Regexp.new given a String with escaped characters accepts '\\M-\\C-\\f'"
fails "Regexp.new given a String with escaped characters accepts '\\M-\\C-\\v'"
fails "Regexp.new given a String with escaped characters accepts '\\M-\\C-\\a'"
fails "Regexp.new given a String with escaped characters accepts '\\M-\\C-\\e'"
fails "Regexp.new given a String with escaped characters accepts '\\M-\\c\\n'"
fails "Regexp.new given a String with escaped characters accepts '\\M-\\c\\t'"
fails "Regexp.new given a String with escaped characters accepts '\\M-\\c\\r'"
fails "Regexp.new given a String with escaped characters accepts '\\M-\\c\\f'"
fails "Regexp.new given a String with escaped characters accepts '\\M-\\c\\v'"
fails "Regexp.new given a String with escaped characters accepts '\\M-\\c\\a'"
fails "Regexp.new given a String with escaped characters accepts '\\M-\\c\\e'"
end
90 changes: 90 additions & 0 deletions spec/filters/unsupported/regular_expressions.rb
View file
Original file line number Diff line number Diff line change
@@ -44,4 +44,94 @@
fails "String#scan supports \\G which matches the end of the previous match / string start for first match"

fails "String#sub with pattern, replacement supports \\G which matches at the beginning of the string"

fails "Regexp#options includes Regexp::FIXEDENCODING for a Regexp literal with the 'u' option"
fails "Regexp#options includes Regexp::FIXEDENCODING for a Regexp literal with the 'e' option"
fails "Regexp#options includes Regexp::FIXEDENCODING for a Regexp literal with the 's' option"
fails "Regexp#options does not include Regexp::FIXEDENCODING for a Regexp literal with the 'n' option"
fails "Regexp#options includes Regexp::NOENCODING for a Regexp literal with the 'n' option"

fails "Regexp.union returns a Regexp with the encoding of an ASCII-incompatible String argument"
fails "Regexp.union returns a Regexp with the encoding of a String containing non-ASCII-compatible characters"
fails "Regexp.union returns a Regexp with US-ASCII encoding if all arguments are ASCII-only"
fails "Regexp.union returns a Regexp with the encoding of multiple non-conflicting ASCII-incompatible String arguments"
fails "Regexp.union returns a Regexp with the encoding of multiple non-conflicting Strings containing non-ASCII-compatible characters"
fails "Regexp.union returns a Regexp with the encoding of a String containing non-ASCII-compatible characters and another ASCII-only String"
fails "Regexp.union returns a Regexp with UTF-8 if one part is UTF-8"
fails "Regexp.union raises ArgumentError if the arguments include conflicting ASCII-incompatible Regexps"
fails "Regexp.union raises ArgumentError if the arguments include conflicting fixed encoding Regexps"
fails "Regexp.union raises ArgumentError if the arguments include a fixed encoding Regexp and a String containing non-ASCII-compatible characters in a different encoding"
fails "Regexp.union raises ArgumentError if the arguments include a String containing non-ASCII-compatible characters and a fixed encoding Regexp in a different encoding"
fails "Regexp.union raises ArgumentError if the arguments include an ASCII-incompatible String and an ASCII-only String"
fails "Regexp.union raises ArgumentError if the arguments include an ASCII-incompatible Regexp and an ASCII-only String"
fails "Regexp.union raises ArgumentError if the arguments include an ASCII-incompatible String and an ASCII-only Regexp"
fails "Regexp.union raises ArgumentError if the arguments include an ASCII-incompatible Regexp and an ASCII-only Regexp"
fails "Regexp.union raises ArgumentError if the arguments include an ASCII-incompatible String and a String containing non-ASCII-compatible characters in a different encoding"
fails "Regexp.union raises ArgumentError if the arguments include an ASCII-incompatible Regexp and a String containing non-ASCII-compatible characters in a different encoding"
fails "Regexp.union raises ArgumentError if the arguments include an ASCII-incompatible String and a Regexp containing non-ASCII-compatible characters in a different encoding"
fails "Regexp.union raises ArgumentError if the arguments include an ASCII-incompatible Regexp and a Regexp containing non-ASCII-compatible characters in a different encoding"

fails "Regexp.new given a Regexp uses US_ASCII encoding if third argument is 'n' or 'none' (case insensitive) and only ascii characters"
fails "Regexp.new given a Regexp uses ASCII_8BIT encoding if third argument is 'n' or 'none' (case insensitive) and non-ascii characters"
fails "Regexp.new given a Regexp returns a Regexp with US-ASCII encoding if only 7-bit ASCII characters are present regardless of the input String's encoding"
fails "Regexp.new given a Regexp returns a Regexp with source String having US-ASCII encoding if only 7-bit ASCII characters are present regardless of the input String's encoding"
fails "Regexp.new given a Regexp returns a Regexp with US-ASCII encoding if UTF-8 escape sequences using only 7-bit ASCII are present"
fails "Regexp.new given a Regexp returns a Regexp with source String having US-ASCII encoding if UTF-8 escape sequences using only 7-bit ASCII are present"
fails "Regexp.new given a Regexp returns a Regexp with UTF-8 encoding if any UTF-8 escape sequences outside 7-bit ASCII are present"
fails "Regexp.new given a Regexp returns a Regexp with source String having UTF-8 encoding if any UTF-8 escape sequences outside 7-bit ASCII are present"
fails "Regexp.new given a Regexp returns a Regexp with the input String's encoding"
fails "Regexp.new given a Regexp returns a Regexp with source String having the input String's encoding"
fails "Regexp.new given a Regexp sets the encoding to UTF-8 if the Regexp literal has the 'u' option"
fails "Regexp.new given a Regexp sets the encoding to EUC-JP if the Regexp literal has the 'e' option"
fails "Regexp.new given a Regexp sets the encoding to Windows-31J if the Regexp literal has the 's' option"
fails "Regexp.new given a Regexp sets the encoding to US-ASCII if the Regexp literal has the 'n' option and the source String is ASCII only"
fails "Regexp.new given a Regexp sets the encoding to source String's encoding if the Regexp literal has the 'n' option and the source String is not ASCII only"
fails "Regexp.new given a String uses US_ASCII encoding if third argument is 'n' or 'none' (case insensitive) and only ascii characters"
fails "Regexp.new given a String uses ASCII_8BIT encoding if third argument is 'n' or 'none' (case insensitive) and non-ascii characters"
fails "Regexp.new given a String returns a Regexp with US-ASCII encoding if only 7-bit ASCII characters are present regardless of the input String's encoding"
fails "Regexp.new given a String returns a Regexp with source String having US-ASCII encoding if only 7-bit ASCII characters are present regardless of the input String's encoding"
fails "Regexp.new given a String returns a Regexp with US-ASCII encoding if UTF-8 escape sequences using only 7-bit ASCII are present"
fails "Regexp.new given a String returns a Regexp with source String having US-ASCII encoding if UTF-8 escape sequences using only 7-bit ASCII are present"
fails "Regexp.new given a String returns a Regexp with UTF-8 encoding if any UTF-8 escape sequences outside 7-bit ASCII are present"
fails "Regexp.new given a String returns a Regexp with source String having UTF-8 encoding if any UTF-8 escape sequences outside 7-bit ASCII are present"
fails "Regexp.new given a String returns a Regexp with the input String's encoding"
fails "Regexp.new given a String returns a Regexp with source String having the input String's encoding"
fails "Regexp.new given a String sets the encoding to UTF-8 if the Regexp literal has the 'u' option"
fails "Regexp.new given a String sets the encoding to EUC-JP if the Regexp literal has the 'e' option"
fails "Regexp.new given a String sets the encoding to Windows-31J if the Regexp literal has the 's' option"
fails "Regexp.new given a String sets the encoding to US-ASCII if the Regexp literal has the 'n' option and the source String is ASCII only"
fails "Regexp.new given a String sets the encoding to source String's encoding if the Regexp literal has the 'n' option and the source String is not ASCII only"
fails "Regexp.new given a String with escaped characters uses US_ASCII encoding if third argument is 'n' or 'none' (case insensitive) and only ascii characters"
fails "Regexp.new given a String with escaped characters uses ASCII_8BIT encoding if third argument is 'n' or 'none' (case insensitive) and non-ascii characters"
fails "Regexp.new given a String with escaped characters returns a Regexp with US-ASCII encoding if only 7-bit ASCII characters are present regardless of the input String's encoding"
fails "Regexp.new given a String with escaped characters returns a Regexp with source String having US-ASCII encoding if only 7-bit ASCII characters are present regardless of the input String's encoding"
fails "Regexp.new given a String with escaped characters returns a Regexp with US-ASCII encoding if UTF-8 escape sequences using only 7-bit ASCII are present"
fails "Regexp.new given a String with escaped characters returns a Regexp with source String having US-ASCII encoding if UTF-8 escape sequences using only 7-bit ASCII are present"
fails "Regexp.new given a String with escaped characters returns a Regexp with UTF-8 encoding if any UTF-8 escape sequences outside 7-bit ASCII are present"
fails "Regexp.new given a String with escaped characters returns a Regexp with source String having UTF-8 encoding if any UTF-8 escape sequences outside 7-bit ASCII are present"
fails "Regexp.new given a String with escaped characters returns a Regexp with the input String's encoding"
fails "Regexp.new given a String with escaped characters returns a Regexp with source String having the input String's encoding"
fails "Regexp.new given a String with escaped characters sets the encoding to UTF-8 if the Regexp literal has the 'u' option"
fails "Regexp.new given a String with escaped characters sets the encoding to EUC-JP if the Regexp literal has the 'e' option"
fails "Regexp.new given a String with escaped characters sets the encoding to Windows-31J if the Regexp literal has the 's' option"
fails "Regexp.new given a String with escaped characters sets the encoding to US-ASCII if the Regexp literal has the 'n' option and the source String is ASCII only"
fails "Regexp.new given a String with escaped characters sets the encoding to source String's encoding if the Regexp literal has the 'n' option and the source String is not ASCII only"
fails "Regexp.new given a String ignores the third argument if it is 'e' or 'euc' (case-insensitive)"
fails "Regexp.new given a String ignores the third argument if it is 's' or 'sjis' (case-insensitive)"
fails "Regexp.new given a String ignores the third argument if it is 'u' or 'utf8' (case-insensitive)"
fails "Regexp.new given a String with escaped characters accepts \\u{H} for a single Unicode codepoint"
fails "Regexp.new given a String with escaped characters accepts \\u{HH} for a single Unicode codepoint"
fails "Regexp.new given a String with escaped characters accepts \\u{HHH} for a single Unicode codepoint"
fails "Regexp.new given a String with escaped characters accepts \\u{HHHH} for a single Unicode codepoint"
fails "Regexp.new given a String with escaped characters accepts \\u{HHHHH} for a single Unicode codepoint"
fails "Regexp.new given a String with escaped characters accepts \\u{HHHHHH} for a single Unicode codepoint"
fails "Regexp.new given a String with escaped characters accepts characters followed by \\u{HHHH}"
fails "Regexp.new given a String with escaped characters accepts \\u{HHHH} followed by characters"
fails "Regexp.new given a String with escaped characters accepts escaped hexadecimal digits followed by \\u{HHHH}"
fails "Regexp.new given a String with escaped characters accepts escaped octal digits followed by \\u{HHHH}"
fails "Regexp.new given a String with escaped characters accepts a combination of escaped octal and hexadecimal digits and \\u{HHHH}"
fails "Regexp.new given a String with escaped characters raises a RegexpError if less than four digits are given for \\uHHHH"
fails "Regexp.new given a String with escaped characters raises a RegexpError if the \\u{} escape is empty"
# regexp extended related
fails "Regexp.new given a String accepts a Fixnum of two or more options ORed together as the second argument"
end
2 changes: 1 addition & 1 deletion spec/opal/core/language/regexp_spec.rb
View file
Original file line number Diff line number Diff line change
@@ -17,4 +17,4 @@
str = "a"
/[#{str}-z]/.should == /[a-z]/
end
end
end
3 changes: 3 additions & 0 deletions spec/rubyspecs
View file
Original file line number Diff line number Diff line change
@@ -130,6 +130,9 @@ corelib/core/regexp/escape_spec
corelib/core/regexp/last_match_spec
corelib/core/regexp/match_spec
corelib/core/regexp/quote_spec
corelib/core/regexp/options_spec
corelib/core/regexp/union_spec
corelib/core/regexp/new_spec

corelib/language/BEGIN_spec
corelib/language/alias_spec