Class: UrlHelper
- Inherits:
-
Object
- Object
- UrlHelper
- Defined in:
- lib/url_helper.rb
Constant Summary collapse
- MAX_URL_LENGTH =
2_000
Class Method Summary collapse
- .absolute(url, cdn = Discourse.asset_host) ⇒ Object
- .absolute_without_cdn(url) ⇒ Object
- .cook_url(url, secure: false, local: nil) ⇒ Object
- .encode(url) ⇒ Object
- .encode_and_parse(url) ⇒ Object
- .encode_component(url_component) ⇒ Object
- .is_local(url) ⇒ Object
-
.is_valid_url?(url) ⇒ Boolean
Heuristic checks to determine if the URL string is a valid absolute URL, path or anchor.
- .local_cdn_url(url) ⇒ Object
- .normalized_encode(uri) ⇒ Object
- .rails_route_from_url(url) ⇒ Object
-
.relaxed_parse(url) ⇒ Object
At the moment this handles invalid URLs that browser address bar accepts where second # is not encoded.
- .schemaless(url) ⇒ Object
- .secure_proxy_without_cdn(url) ⇒ Object
- .unencode(url) ⇒ Object
Class Method Details
.absolute(url, cdn = Discourse.asset_host) ⇒ Object
75 76 77 78 |
# File 'lib/url_helper.rb', line 75 def self.absolute(url, cdn = Discourse.asset_host) cdn = "https:#{cdn}" if cdn && cdn =~ %r{\A//} url =~ %r{\A/[^/]} ? (cdn || Discourse.base_url_no_prefix) + url : url end |
.absolute_without_cdn(url) ⇒ Object
80 81 82 |
# File 'lib/url_helper.rb', line 80 def self.absolute_without_cdn(url) self.absolute(url, nil) end |
.cook_url(url, secure: false, local: nil) ⇒ Object
126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 |
# File 'lib/url_helper.rb', line 126 def self.cook_url(url, secure: false, local: nil) is_secure = SiteSetting.secure_uploads && secure local = is_local(url) if local.nil? return url if !local url = is_secure ? secure_proxy_without_cdn(url) : absolute_without_cdn(url) # we always want secure uploads to come from # Discourse.base_url_no_prefix/secure-uploads # to avoid asset_host mixups return schemaless(url) if is_secure # PERF: avoid parsing url except for extreme conditions # this is a hot path used on home page filename = url if url.include?("?") uri = URI.parse(url) filename = File.basename(uri.path) end # this technically requires a filename, but will work with a URL as long as it end with the # extension and has no query params = !FileHelper.is_supported_media?(filename) no_cdn = SiteSetting.login_required || SiteSetting.prevent_anons_from_downloading_files unless && no_cdn url = Discourse.store.cdn_url(url) url = local_cdn_url(url) if Discourse.store.external? end schemaless(url) rescue URI::Error url end |
.encode(url) ⇒ Object
54 55 56 |
# File 'lib/url_helper.rb', line 54 def self.encode(url) Addressable::URI.encode(url) end |
.encode_and_parse(url) ⇒ Object
50 51 52 |
# File 'lib/url_helper.rb', line 50 def self.encode_and_parse(url) URI.parse(Addressable::URI.encode(url)) end |
.encode_component(url_component) ⇒ Object
62 63 64 |
# File 'lib/url_helper.rb', line 62 def self.encode_component(url_component) Addressable::URI.encode_component(url_component) end |
.is_local(url) ⇒ Object
66 67 68 69 70 71 72 73 |
# File 'lib/url_helper.rb', line 66 def self.is_local(url) url.present? && ( Discourse.store.has_been_uploaded?(url) || !!(url =~ Regexp.new("^#{Discourse.base_path}/(assets|plugins|images)/")) || url.start_with?(Discourse.asset_host || Discourse.base_url_no_prefix) ) end |
.is_valid_url?(url) ⇒ Boolean
Heuristic checks to determine if the URL string is a valid absolute URL, path or anchor
28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 |
# File 'lib/url_helper.rb', line 28 def self.is_valid_url?(url) uri = URI.parse(url) return true if uri.is_a?(URI::Generic) && url.starts_with?("/") || url.match?(/\A\#([^#]*)/) if uri.scheme return true if uri.is_a?(URI::MailTo) if url.match?(%r{\A#{uri.scheme}://[^/]}) && ( uri.is_a?(URI::HTTP) || uri.is_a?(URI::HTTPS) || uri.is_a?(URI::FTP) || uri.is_a?(URI::LDAP) ) return true end end false rescue URI::InvalidURIError false end |
.local_cdn_url(url) ⇒ Object
161 162 163 164 165 166 167 168 |
# File 'lib/url_helper.rb', line 161 def self.local_cdn_url(url) return url if Discourse.asset_host.blank? if url.start_with?("/#{Discourse.store.upload_path}/") "#{Discourse.asset_host}#{url}" else url.sub(Discourse.base_url_no_prefix, Discourse.asset_host) end end |
.normalized_encode(uri) ⇒ Object
92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 |
# File 'lib/url_helper.rb', line 92 def self.normalized_encode(uri) url = uri.to_s if url.length > MAX_URL_LENGTH raise ArgumentError.new("URL starting with #{url[0..100]} is too long") end # Ideally we will jump straight to `Addressable::URI.normalized_encode`. However, # that implementation has some edge-case issues like https://github.com/sporkmonger/addressable/issues/472. # To temporaily work around those issues for the majority of cases, we try parsing with `::URI`. # If that fails (e.g. due to non-ascii characters) then we will fall back to addressable. # Hopefully we can simplify this back to `Addressable::URI.normalized_encode` in the future. # edge case where we expect mailto:test%40test.com to normalize to mailto:[email protected] return normalize_with_addressable(url) if url.match(/\Amailto:/) # If it doesn't pass the regexp, it's definitely not gonna parse with URI.parse. Skip # to addressable return normalize_with_addressable(url) if !url.match?(/\A#{URI.regexp}\z/) begin normalize_with_ruby_uri(url) rescue URI::Error normalize_with_addressable(url) end end |
.rails_route_from_url(url) ⇒ Object
119 120 121 122 123 124 |
# File 'lib/url_helper.rb', line 119 def self.rails_route_from_url(url) path = URI.parse(encode(url)).path Rails.application.routes.recognize_path(path) rescue Addressable::URI::InvalidURIError, URI::InvalidComponentError nil end |
.relaxed_parse(url) ⇒ Object
At the moment this handles invalid URLs that browser address bar accepts where second # is not encoded
Longer term we can add support of simpleidn and encode unicode domains
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 |
# File 'lib/url_helper.rb', line 10 def self.relaxed_parse(url) url, fragment = url.split("#", 2) uri = URI.parse(url) if uri # Addressable::URI::CharacterClasses::UNRESERVED is used here because without it # the # in the fragment is not encoded fragment = Addressable::URI.encode_component( fragment, Addressable::URI::CharacterClasses::UNRESERVED, ) if fragment&.include?("#") uri.fragment = fragment uri end rescue URI::Error end |
.schemaless(url) ⇒ Object
84 85 86 |
# File 'lib/url_helper.rb', line 84 def self.schemaless(url) url.sub(/\Ahttp:/i, "") end |
.secure_proxy_without_cdn(url) ⇒ Object
88 89 90 |
# File 'lib/url_helper.rb', line 88 def self.secure_proxy_without_cdn(url) self.absolute(Upload.secure_uploads_url_from_upload_url(url), nil) end |
.unencode(url) ⇒ Object
58 59 60 |
# File 'lib/url_helper.rb', line 58 def self.unencode(url) Addressable::URI.unencode(url) end |