diff options
| author | murphy <murphy@rubychan.de> | 2009-10-20 11:12:26 +0000 |
|---|---|---|
| committer | murphy <murphy@rubychan.de> | 2009-10-20 11:12:26 +0000 |
| commit | 9f6e54e92820754d4f05e7d6ef556c11c5351411 (patch) | |
| tree | 3a612819bb121418f05fae076e7640228c36824f /lib/coderay/scanners/php.rb | |
| parent | a8163c39206fd11ef30ef8efb2c73a6f6a97e424 (diff) | |
| download | coderay-9f6e54e92820754d4f05e7d6ef556c11c5351411.tar.gz | |
Updated PHP scanner (#36)
* highlighting of class and function definitions
* improved HTML/PHP detection
* heredocs (simple)
* a new test for classes
Diffstat (limited to 'lib/coderay/scanners/php.rb')
| -rw-r--r-- | lib/coderay/scanners/php.rb | 50 |
1 files changed, 42 insertions, 8 deletions
diff --git a/lib/coderay/scanners/php.rb b/lib/coderay/scanners/php.rb index 700c7cb..bfbc642 100644 --- a/lib/coderay/scanners/php.rb +++ b/lib/coderay/scanners/php.rb @@ -184,7 +184,7 @@ module Scanners HTML_INDICATOR = /<!DOCTYPE html|<(?:html|body|div|p)[> ]/i - IDENTIFIER = /[a-z_\x80-\xFF][a-z0-9_\x80-\xFF]*/i + IDENTIFIER = /[a-z_\x7f-\xFF][a-z0-9_\x7f-\xFF]*/i VARIABLE = /\$#{IDENTIFIER}/ OPERATOR = / @@ -205,7 +205,7 @@ module Scanners states = [:initial] if match?(RE::PHP_START) || # starts with <? - (match?(/\s*<(?i:\w|\?xml)/) && exist?(RE::PHP_START)) || # starts with HTML tag and contains <? + (match?(/\s*<\S/) && exist?(RE::PHP_START)) || # starts with tag and contains <? exist?(RE::HTML_INDICATOR) # is PHP inside HTML, so start with HTML else @@ -253,7 +253,10 @@ module Scanners kind = :label elsif kind == :ident && match =~ /^[A-Z]/ kind = :constant - # TODO: function and class definitions + elsif kind == :reserved && match == 'class' + states << :class_expected + elsif kind == :reserved && match == 'function' + states << :function_expected end elsif scan(/(?:\d+\.\d*|\d*\.\d+)(?:e[-+]?\d+)?|\d+e[-+]?\d+/i) @@ -277,11 +280,17 @@ module Scanners states.push :dqstring # TODO: Heredocs - # elsif match = scan(/<<</ + IDENTIFIER) - # tokens << [:open, :string] - # heredocdelim = match[RE::IDENTIFIER] - # kind = :delimiter - # states.push :heredocstring + # See http://de2.php.net/manual/en/language.types.string.php#language.types.string.syntax.heredoc + elsif match = scan(/<<<(#{RE::IDENTIFIER})/o) + tokens << [:open, :string] + heredocdelim = Regexp.escape self[1] + tokens << [match, :delimiter] + next if eos? + tokens << [scan_until(/\n(?=#{heredocdelim};?$)|\z/), :content] + next if eos? + tokens << [scan(/#{heredocdelim}/), :delimiter] + tokens << [:close, :string] + next elsif scan RE::VARIABLE kind = :local_variable @@ -379,6 +388,31 @@ module Scanners elsif scan(/\$/) kind = :content end + + when :class_expected + if scan(/\s+/) + kind = :space + elsif match = scan(/#{RE::IDENTIFIER}/o) + kind = :class + states.pop + else + states.pop + next + end + + when :function_expected + if scan(/\s+/) + kind = :space + elsif scan(/&/) + kind = :operator + elsif match = scan(/#{RE::IDENTIFIER}/o) + kind = :function + states.pop + else + states.pop + next + end + else raise_inspect 'Unknown state!', tokens, states end |