| 1 |
From 1ad9075734a7fd9d58920c714953ec851efcae6b Mon Sep 17 00:00:00 2001 |
| 2 |
From: usa <usa@b2dd03c8-39d4-4d8f-98ff-823fe69b080e> |
| 3 |
Date: Fri, 22 Feb 2013 09:35:46 +0000 |
| 4 |
Subject: [PATCH] * lib/rexml/document.rb |
| 5 |
(REXML::Document.entity_expansion_text_limit): new |
| 6 |
attribute to read/write entity expansion text limit. the |
| 7 |
default limit is 10Kb. |
| 8 |
|
| 9 |
* lib/rexml/text.rb (REXML::Text.unnormalize): check above attribute. |
| 10 |
|
| 11 |
|
| 12 |
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@39384 b2dd03c8-39d4-4d8f-98ff-823fe69b080e |
| 13 |
--- |
| 14 |
lib/rexml/document.rb | 12 ++++++++++++ |
| 15 |
lib/rexml/text.rb | 40 +++++++++++++++++++++++++--------------- |
| 16 |
test/rexml/test_entity.rb | 18 ++++++++++++++++++ |
| 17 |
3 files changed, 55 insertions(+), 15 deletions(-) |
| 18 |
|
| 19 |
diff --git a/lib/rexml/document.rb b/lib/rexml/document.rb |
| 20 |
index f9fdbda..c528037 100644 |
| 21 |
--- a/lib/rexml/document.rb |
| 22 |
+++ b/lib/rexml/document.rb |
| 23 |
@@ -213,6 +213,17 @@ def Document::entity_expansion_limit |
| 24 |
return @@entity_expansion_limit |
| 25 |
end |
| 26 |
|
| 27 |
+ @@entity_expansion_text_limit = 10_240 |
| 28 |
+ |
| 29 |
+ # Set the entity expansion limit. By default the limit is set to 10240. |
| 30 |
+ def Document::entity_expansion_text_limit=( val ) |
| 31 |
+ @@entity_expansion_text_limit = val |
| 32 |
+ end |
| 33 |
+ |
| 34 |
+ # Get the entity expansion limit. By default the limit is set to 10240 |
| 35 |
+ def Document::entity_expansion_text_limit |
| 36 |
+ return @@entity_expansion_text_limit |
| 37 |
+ end |
| 38 |
attr_reader :entity_expansion_count |
| 39 |
|
| 40 |
def record_entity_expansion |
| 41 |
diff --git a/lib/rexml/text.rb b/lib/rexml/text.rb |
| 42 |
index 6623c0c..878d13b 100644 |
| 43 |
--- a/lib/rexml/text.rb |
| 44 |
+++ b/lib/rexml/text.rb |
| 45 |
@@ -308,37 +308,35 @@ def Text::normalize( input, doctype=nil, entity_filter=nil ) |
| 46 |
|
| 47 |
# Unescapes all possible entities |
| 48 |
def Text::unnormalize( string, doctype=nil, filter=nil, illegal=nil ) |
| 49 |
- rv = string.clone |
| 50 |
- rv.gsub!( /\r\n?/, "\n" ) |
| 51 |
- matches = rv.scan( REFERENCE ) |
| 52 |
- return rv if matches.size == 0 |
| 53 |
- rv.gsub!( NUMERICENTITY ) {|m| |
| 54 |
- m=$1 |
| 55 |
- m = "0#{m}" if m[0] == ?x |
| 56 |
- [Integer(m)].pack('U*') |
| 57 |
+ sum = 0 |
| 58 |
+ string.gsub( /\r\n?/, "\n" ).gsub( REFERENCE ) { |
| 59 |
+ s = Text.expand($&, doctype, filter) |
| 60 |
+ if sum + s.bytesize > Document.entity_expansion_text_limit |
| 61 |
+ raise "entity expansion has grown too large" |
| 62 |
+ else |
| 63 |
+ sum += s.bytesize |
| 64 |
+ end |
| 65 |
+ s |
| 66 |
} |
| 67 |
- matches.collect!{|x|x[0]}.compact! |
| 68 |
- if matches.size > 0 |
| 69 |
- if doctype |
| 70 |
- matches.each do |entity_reference| |
| 71 |
- unless filter and filter.include?(entity_reference) |
| 72 |
- entity_value = doctype.entity( entity_reference ) |
| 73 |
- re = /&#{entity_reference};/ |
| 74 |
- rv.gsub!( re, entity_value ) if entity_value |
| 75 |
- end |
| 76 |
- end |
| 77 |
+ end |
| 78 |
+ |
| 79 |
+ def Text.expand(ref, doctype, filter) |
| 80 |
+ if ref[1] == ?# |
| 81 |
+ if ref[2] == ?x |
| 82 |
+ [ref[3...-1].to_i(16)].pack('U*') |
| 83 |
else |
| 84 |
- matches.each do |entity_reference| |
| 85 |
- unless filter and filter.include?(entity_reference) |
| 86 |
- entity_value = DocType::DEFAULT_ENTITIES[ entity_reference ] |
| 87 |
- re = /&#{entity_reference};/ |
| 88 |
- rv.gsub!( re, entity_value.value ) if entity_value |
| 89 |
- end |
| 90 |
- end |
| 91 |
- end |
| 92 |
- rv.gsub!( /&/, '&' ) |
| 93 |
- end |
| 94 |
- rv |
| 95 |
+ [ref[2...-1].to_i].pack('U*') |
| 96 |
+ end |
| 97 |
+ elsif ref == '&' |
| 98 |
+ '&' |
| 99 |
+ elsif filter and filter.include?( ref[1...-1] ) |
| 100 |
+ ref |
| 101 |
+ elsif doctype |
| 102 |
+ doctype.entity( ref[1...-1] ) or ref |
| 103 |
+ else |
| 104 |
+ entity_value = DocType::DEFAULT_ENTITIES[ ref[1...-1] ] |
| 105 |
+ entity_value ? entity_value.value : ref |
| 106 |
+ end |
| 107 |
end |
| 108 |
end |
| 109 |
end |
| 110 |
diff --git a/test/rexml/test_entity.rb b/test/rexml/test_entity.rb |
| 111 |
index e6d6f29..5900fac 100644 |
| 112 |
--- /dev/null |
| 113 |
+++ b/test/rexml/test_entity.rb |
| 114 |
@@ -0,0 +1,23 @@ def test_replace_entities |
| 115 |
+require "test/unit/testcase" |
| 116 |
+ |
| 117 |
+require 'rexml/document' |
| 118 |
+ |
| 119 |
+class EntityTester < Test::Unit::TestCase |
| 120 |
+ def test_entity_string_limit |
| 121 |
+ template = '<!DOCTYPE bomb [ <!ENTITY a "^" > ]> <bomb>$</bomb>' |
| 122 |
+ len = 5120 # 5k per entity |
| 123 |
+ template.sub!(/\^/, "B" * len) |
| 124 |
+ |
| 125 |
+ # 10k is OK |
| 126 |
+ entities = '&a;' * 2 # 5k entity * 2 = 10k |
| 127 |
+ xmldoc = REXML::Document.new(template.sub(/\$/, entities)) |
| 128 |
+ assert_equal(len * 2, xmldoc.root.text.bytesize) |
| 129 |
+ |
| 130 |
+ # above 10k explodes |
| 131 |
+ entities = '&a;' * 3 # 5k entity * 2 = 15k |
| 132 |
+ xmldoc = REXML::Document.new(template.sub(/\$/, entities)) |
| 133 |
+ assert_raises(RuntimeError) do |
| 134 |
+ xmldoc.root.text |
| 135 |
+ end |
| 136 |
+ end |
| 137 |
+end |
| 138 |
-- |
| 139 |
1.7.10 |
| 140 |
|
Parent Directory
| 
Revision Log