net-imap 0.3.7 → 0.4.5

data/lib/net/imap/response_data.rb

@@ -2,6 +2,8 @@
 
 module Net
   class IMAP < Protocol
+    autoload :FetchData,        "#{__dir__}/fetch_data"
+    autoload :SequenceSet,      "#{__dir__}/sequence_set"
 
     # Net::IMAP::ContinuationRequest represents command continuation requests.
     #
@@ -55,17 +57,71 @@ module Net
 
     # Net::IMAP::IgnoredResponse represents intentionally ignored responses.
     #
-    # This includes untagged response "NOOP" sent by eg. Zimbra to avoid some
-    # clients to close the connection.
+    # This includes untagged response "NOOP" sent by eg. Zimbra to avoid
+    # some clients to close the connection.
     #
     # It matches no IMAP standard.
+    class IgnoredResponse < UntaggedResponse
+    end
+
+    # **Note:** This represents an intentionally _unstable_ API.  Where
+    # instances of this class are returned, future releases may return a
+    # different (incompatible) object <em>without deprecation or warning</em>.
+    #
+    # Net::IMAP::UnparsedData represents data for unknown response types or
+    # unknown extensions to response types without a well-defined extension
+    # grammar.
     #
-    class IgnoredResponse < Struct.new(:raw_data)
+    # See also: UnparsedNumericResponseData, ExtensionData, IgnoredResponse
+    class UnparsedData < Struct.new(:unparsed_data)
       ##
-      # method: raw_data
-      # :call-seq: raw_data -> string
+      # method: unparsed_data
+      # :call-seq: unparsed_data -> string
       #
-      # The raw response data.
+      # The unparsed data
+    end
+
+    # **Note:** This represents an intentionally _unstable_ API.  Where
+    # instances of this class are returned, future releases may return a
+    # different (incompatible) object <em>without deprecation or warning</em>.
+    #
+    # Net::IMAP::UnparsedNumericResponseData represents data for unhandled
+    # response types with a numeric prefix.  See the documentation for #number.
+    #
+    # See also: UnparsedData, ExtensionData, IgnoredResponse
+    class UnparsedNumericResponseData < Struct.new(:number, :unparsed_data)
+      ##
+      # method: number
+      # :call-seq: number -> integer
+      #
+      # Returns a numeric response data prefix, when available.
+      #
+      # Many response types are prefixed with a non-negative #number.  For
+      # message data, #number may represent a sequence number or a UID.  For
+      # mailbox data, #number may represent a message count.
+
+      ##
+      # method: unparsed_data
+      # :call-seq: unparsed_data -> string
+      #
+      # The unparsed data, not including #number or UntaggedResponse#name.
+    end
+
+    # **Note:** This represents an intentionally _unstable_ API.  Where
+    # instances of this class are returned, future releases may return a
+    # different (incompatible) object <em>without deprecation or warning</em>.
+    #
+    # Net::IMAP::ExtensionData represents data that is parsable according to the
+    # forward-compatible extension syntax in RFC3501, RFC4466, or RFC9051, but
+    # isn't directly known or understood by Net::IMAP yet.
+    #
+    # See also: UnparsedData, UnparsedNumericResponseData, IgnoredResponse
+    class ExtensionData < Struct.new(:data)
+      ##
+      # method: data
+      # :call-seq: data -> string
+      #
+      # The parsed extension data.
     end
 
     # Net::IMAP::TaggedResponse represents tagged responses.
@@ -108,6 +164,9 @@ module Net
     # UntaggedResponse#data when the response type is a "condition" ("OK", "NO",
     # "BAD", "PREAUTH", or "BYE").
     class ResponseText < Struct.new(:code, :text)
+      # Used to avoid an allocation when ResponseText is empty
+      EMPTY = new(nil, "").freeze
+
       ##
       # method: code
       # :call-seq: code -> ResponseCode or nil
@@ -197,6 +256,10 @@ module Net
     # * +ALREADYEXISTS+
     # * +NONEXISTENT+
     #
+    # Other supported \IMAP extension response codes:
+    # * +OBJECTID+ {[RFC8474]}[https://www.rfc-editor.org/rfc/rfc8474.html#section-7]
+    #   * +MAILBOXID+, #data will be a string
+    #
     class ResponseCode < Struct.new(:name, :data)
       ##
       # method: name
@@ -448,210 +511,6 @@ module Net
       # "UIDVALIDITY", "UNSEEN". Each value is a number.
     end
 
-    # Net::IMAP::FetchData represents the contents of a FETCH response.
-    #
-    # Net::IMAP#fetch and Net::IMAP#uid_fetch both return an array of
-    # FetchData objects.
-    #
-    # === Fetch attributes
-    #
-    #--
-    # TODO: merge branch with accessor methods for each type of attr.  Then
-    # move nearly all of the +attr+ documentation onto the appropriate
-    # accessor methods.
-    #++
-    #
-    # Each key of the #attr hash is the data item name for the fetched value.
-    # Each data item represents a message attribute, part of one, or an
-    # interpretation of one.  #seqno is not a message attribute.  Most message
-    # attributes are static and must never change for a given <tt>[server,
-    # account, mailbox, UIDVALIDITY, UID]</tt> tuple.  A few message attributes
-    # can be dynamically changed, e.g. using the {STORE
-    # command}[rdoc-ref:Net::IMAP#store].
-    #
-    # See {[IMAP4rev1] §7.4.2}[https://www.rfc-editor.org/rfc/rfc3501.html#section-7.4.2]
-    # and {[IMAP4rev2] §7.5.2}[https://www.rfc-editor.org/rfc/rfc9051.html#section-7.5.2]
-    # for full description of the standard fetch response data items, and
-    # Net::IMAP@Message+envelope+and+body+structure for other relevant RFCs.
-    #
-    # ==== Static fetch data items
-    #
-    # The static data items
-    # defined by [IMAP4rev1[https://www.rfc-editor.org/rfc/rfc3501.html]] are:
-    #
-    # [<tt>"UID"</tt>]
-    #   A number expressing the unique identifier of the message.
-    #
-    # [<tt>"BODY[]"</tt>, <tt>"BODY[]<#{offset}>"</tt>]
-    #   The [RFC5322[https://tools.ietf.org/html/rfc5322]] expression of the
-    #   entire message, as a string.
-    #
-    #   If +offset+ is specified, this returned string is a substring of the
-    #   entire contents, starting at that origin octet.  This means that
-    #   <tt>BODY[]<0></tt> MAY be truncated, but <tt>BODY[]</tt> is NEVER
-    #   truncated.
-    #
-    #   <em>Messages can be parsed using the "mail" gem.</em>
-    #
-    #   [Note]
-    #     When fetching <tt>BODY.PEEK[#{specifier}]</tt>, the data will be
-    #     returned in <tt>BODY[#{specifier}]</tt>, without the +PEEK+.  This is
-    #     true for all of the <tt>BODY[...]</tt> attribute forms.
-    #
-    # [<tt>"BODY[HEADER]"</tt>, <tt>"BODY[HEADER]<#{offset}>"</tt>]
-    #   The [RFC5322[https://tools.ietf.org/html/rfc5322]] header of the
-    #   message.
-    #
-    #   <em>Message headers can be parsed using the "mail" gem.</em>
-    #
-    # [<tt>"BODY[HEADER.FIELDS (#{fields.join(" ")})]"</tt>,]
-    # [<tt>"BODY[HEADER.FIELDS (#{fields.join(" ")})]<#{offset}>"</tt>]
-    #   When field names are given, the subset contains only the header fields
-    #   that matches one of the names in the list.  The field names are based
-    #   on what was requested, not on what was returned.
-    #
-    # [<tt>"BODY[HEADER.FIELDS.NOT (#{fields.join(" ")})]"</tt>,]
-    # [<tt>"BODY[HEADER.FIELDS.NOT (#{fields.join(" ")})]<#{offset}>"</tt>]
-    #   When the <tt>HEADER.FIELDS.NOT</tt> is used, the subset is all of the
-    #   fields that <em>do not</em> match any names in the list.
-    #
-    # [<tt>"BODY[TEXT]"</tt>, <tt>"BODY[TEXT]<#{offset}>"</tt>]
-    #   The text body of the message, omitting
-    #   the [RFC5322[https://tools.ietf.org/html/rfc5322]] header.
-    #
-    # [<tt>"BODY[#{part}]"</tt>, <tt>"BODY[#{part}]<#{offset}>"</tt>]
-    #   The text of a particular body section, if it was fetched.
-    #
-    #   Multiple part specifiers will be joined with <tt>"."</tt>.  Numeric
-    #   part specifiers refer to the MIME part number, counting up from +1+.
-    #   Messages that don't use MIME, or MIME messages that are not multipart
-    #   and don't hold an encapsulated message, only have a part +1+.
-    #
-    #   8-bit textual data is permitted if
-    #   a [CHARSET[https://tools.ietf.org/html/rfc2978]] identifier is part of
-    #   the body parameter parenthesized list for this section.  See
-    #   BodyTypeBasic.
-    #
-    #   MESSAGE/RFC822 or MESSAGE/GLOBAL message, or a subset of the header, if
-    #   it was fetched.
-    #
-    # [<tt>"BODY[#{part}.HEADER]"</tt>,]
-    # [<tt>"BODY[#{part}.HEADER]<#{offset}>"</tt>,]
-    # [<tt>"BODY[#{part}.HEADER.FIELDS.NOT (#{fields.join(" ")})]"</tt>,]
-    # [<tt>"BODY[#{part}.HEADER.FIELDS.NOT (#{fields.join(" ")})]<#{offset}>"</tt>,]
-    # [<tt>"BODY[#{part}.TEXT]"</tt>,]
-    # [<tt>"BODY[#{part}.TEXT]<#{offset}>"</tt>,]
-    # [<tt>"BODY[#{part}.MIME]"</tt>,]
-    # [<tt>"BODY[#{part}.MIME]<#{offset}>"</tt>]
-    #   +HEADER+, <tt>HEADER.FIELDS</tt>, <tt>HEADER.FIELDS.NOT</tt>, and
-    #   <tt>TEXT</tt> can be prefixed by numeric part specifiers, if it refers
-    #   to a part of type <tt>message/rfc822</tt> or <tt>message/global</tt>.
-    #
-    #   +MIME+ refers to the [MIME-IMB[https://tools.ietf.org/html/rfc2045]]
-    #   header for this part.
-    #
-    # [<tt>"BODY"</tt>]
-    #   A form of +BODYSTRUCTURE+, without any extension data.
-    #
-    # [<tt>"BODYSTRUCTURE"</tt>]
-    #   Returns a BodyStructure object that describes
-    #   the [MIME-IMB[https://tools.ietf.org/html/rfc2045]] body structure of
-    #   a message, if it was fetched.
-    #
-    # [<tt>"ENVELOPE"</tt>]
-    #   An Envelope object that describes the envelope structure of a message.
-    #   See the documentation for Envelope for a description of the envelope
-    #   structure attributes.
-    #
-    # [<tt>"INTERNALDATE"</tt>]
-    #   The internal date and time of the message on the server.  This is not
-    #   the date and time in
-    #   the [RFC5322[https://tools.ietf.org/html/rfc5322]] header, but rather
-    #   a date and time which reflects when the message was received.
-    #
-    # [<tt>"RFC822.SIZE"</tt>]
-    #   A number expressing the [RFC5322[https://tools.ietf.org/html/rfc5322]]
-    #   size of the message.
-    #
-    #   [Note]
-    #     \IMAP was originally developed for the older RFC-822 standard, and
-    #     as a consequence several fetch items in \IMAP incorporate "RFC822"
-    #     in their name.  With the exception of +RFC822.SIZE+, there are more
-    #     modern replacements; for example, the modern version of
-    #     +RFC822.HEADER+ is <tt>BODY.PEEK[HEADER]</tt>.  In all cases,
-    #     "RFC822" should be interpreted as a reference to the
-    #     updated [RFC5322[https://tools.ietf.org/html/rfc5322]] standard.
-    #
-    # [<tt>"RFC822"</tt>]
-    #   Semantically equivalent to <tt>BODY[]</tt>.
-    # [<tt>"RFC822.HEADER"</tt>]
-    #   Semantically equivalent to <tt>BODY[HEADER]</tt>.
-    # [<tt>"RFC822.TEXT"</tt>]
-    #   Semantically equivalent to <tt>BODY[TEXT]</tt>.
-    #
-    # [Note:]
-    #   >>>
-    #     Additional static fields are defined in \IMAP extensions and
-    #     [IMAP4rev2[https://www.rfc-editor.org/rfc/rfc9051.html]], but
-    #     Net::IMAP can't parse them yet.
-    #
-    #--
-    # <tt>"BINARY[#{section_binary}]<#{offset}>"</tt>:: TODO...
-    # <tt>"BINARY.SIZE[#{sectionbinary}]"</tt>::        TODO...
-    # <tt>"EMAILID"</tt>::                              TODO...
-    # <tt>"THREADID"</tt>::                             TODO...
-    # <tt>"SAVEDATE"</tt>::                             TODO...
-    #++
-    #
-    # ==== Dynamic message attributes
-    # The only dynamic item defined
-    # by [{IMAP4rev1}[https://www.rfc-editor.org/rfc/rfc3501.html]] is:
-    # [<tt>"FLAGS"</tt>]
-    #   An array of flags that are set for this message.  System flags are
-    #   symbols that have been capitalized by String#capitalize.  Keyword
-    #   flags are strings and their case is not changed.
-    #
-    # \IMAP extensions define new dynamic fields, e.g.:
-    #
-    # [<tt>"MODSEQ"</tt>]
-    #   The modification sequence number associated with this IMAP message.
-    #
-    #   Requires the [CONDSTORE[https://tools.ietf.org/html/rfc7162]]
-    #   server {capability}[rdoc-ref:Net::IMAP#capability].
-    #
-    # [Note:]
-    #   >>>
-    #     Additional dynamic fields are defined in \IMAP extensions, but
-    #     Net::IMAP can't parse them yet.
-    #
-    #--
-    # <tt>"ANNOTATE"</tt>:: TODO...
-    # <tt>"PREVIEW"</tt>::  TODO...
-    #++
-    #
-    class FetchData < Struct.new(:seqno, :attr)
-      ##
-      # method: seqno
-      # :call-seq: seqno -> Integer
-      #
-      # The message sequence number.
-      #
-      # [Note]
-      #   This is never the unique identifier (UID), not even for the
-      #   Net::IMAP#uid_fetch result.  If it was returned, the UID is available
-      #   from <tt>attr["UID"]</tt>.
-
-      ##
-      # method: attr
-      # :call-seq: attr -> hash
-      #
-      # A hash.  Each key is specifies a message attribute, and the value is the
-      # corresponding data item.
-      #
-      # See rdoc-ref:FetchData@Fetch+attributes for descriptions of possible
-      # values.
-    end
-
     # Net::IMAP::Envelope represents envelope structures of messages.
     #
     # [Note]
@@ -665,6 +524,7 @@ module Net
     # for full description of the envelope fields, and
     # Net::IMAP@Message+envelope+and+body+structure for other relevant RFCs.
     #
+    # Returned by FetchData#envelope
     class Envelope < Struct.new(:date, :subject, :from, :sender, :reply_to,
                                 :to, :cc, :bcc, :in_reply_to, :message_id)
       ##
@@ -891,13 +751,6 @@ module Net
     #                   should use BodyTypeBasic.
     # BodyTypeMultipart:: for <tt>multipart/*</tt> parts
     #
-    # ==== Deprecated BodyStructure classes
-    # The following classes represent invalid server responses or parser bugs:
-    # BodyTypeExtension:: parser bug: used for <tt>message/*</tt> where
-    #                     BodyTypeBasic should have been used.
-    # BodyTypeAttachment:: server bug: some servers sometimes return the
-    #                      "Content-Disposition: attachment" data where the
-    #                      entire body structure for a message part is expected.
     module BodyStructure
     end
 
@@ -914,6 +767,7 @@ module Net
                                      :param, :content_id,
                                      :description, :encoding, :size,
                                      :md5, :disposition, :language,
+                                     :location,
                                      :extension)
       include BodyStructure
 
@@ -1049,6 +903,7 @@ module Net
                                     :description, :encoding, :size,
                                     :lines,
                                     :md5, :disposition, :language,
+                                    :location,
                                     :extension)
       include BodyStructure
 
@@ -1088,12 +943,12 @@ module Net
     # * description[rdoc-ref:BodyTypeBasic#description]
     # * encoding[rdoc-ref:BodyTypeBasic#encoding]
     # * size[rdoc-ref:BodyTypeBasic#size]
-    #
     class BodyTypeMessage < Struct.new(:media_type, :subtype,
                                        :param, :content_id,
                                        :description, :encoding, :size,
                                        :envelope, :body, :lines,
                                        :md5, :disposition, :language,
+                                       :location,
                                        :extension)
       include BodyStructure
 
@@ -1126,36 +981,41 @@ module Net
       end
     end
 
-    # === WARNING
-    # BodyTypeAttachment represents a <tt>body-fld-dsp</tt> that is
-    # incorrectly in a position where the IMAP4rev1 grammar expects a nested
-    # +body+ structure.
+    # BodyTypeAttachment is not used and will be removed in an upcoming release.
     #
-    # >>>
-    #   \IMAP body structures are parenthesized lists and assign their fields
-    #   positionally, so missing fields change the intepretation of all
-    #   following fields.  Buggy \IMAP servers sometimes leave fields missing
-    #   rather than empty, which inevitably confuses parsers.
-    #   BodyTypeAttachment was an attempt to parse a common type of buggy body
-    #   structure without crashing.
-    #
-    #   Currently, when Net::IMAP::ResponseParser sees "attachment" as the first
-    #   entry in a <tt>body-type-1part</tt>, which is where the MIME type should
-    #   be, it uses BodyTypeAttachment to capture the rest.  "attachment" is not
-    #   a valid MIME type, but _is_ a common <tt>Content-Disposition</tt>.  What
-    #   might have happened was that buggy server could not parse the message
-    #   (which might have been incorrectly formatted) and output a
-    #   <tt>body-type-dsp</tt> where a Net::IMAP::ResponseParser expected to see
-    #   a +body+.
-    #
-    # A future release will replace this, probably with a ContentDisposition
-    # nested inside another body structure object, maybe BodyTypeBasic, or
-    # perhaps a new body structure class that represents any unparsable body
-    # structure.
+    # === Bug Analysis
+    #
+    # \IMAP body structures are parenthesized lists and assign their fields
+    # positionally, so missing fields change the intepretation of all
+    # following fields.  Additionally, different body types have a different
+    # number of required fields, followed by optional "extension" fields.
+    #
+    # BodyTypeAttachment was previously returned when a "message/rfc822" part,
+    # which should be sent as <tt>body-type-msg</tt> with ten required fields,
+    # was actually sent as a <tt>body-type-basic</tt> with _seven_ required
+    # fields.
+    #
+    #   basic => type, subtype, param, id, desc, enc, octets, md5=nil,  dsp=nil, lang=nil, loc=nil, *ext
+    #   msg   => type, subtype, param, id, desc, enc, octets, envelope, body,    lines,    md5=nil, ...
+    #
+    # Normally, +envelope+ and +md5+ are incompatible, but Net::IMAP leniently
+    # allowed buggy servers to send +NIL+ for +envelope+.  As a result, when a
+    # server sent a <tt>message/rfc822</tt> part with +NIL+ for +md5+ and a
+    # non-<tt>NIL</tt> +dsp+, Net::IMAP mis-interpreted the
+    # <tt>Content-Disposition</tt> as if it were a strange body type.  In all
+    # reported cases, the <tt>Content-Disposition</tt> was "attachment", so
+    # BodyTypeAttachment was created as the workaround.
+    #
+    # === Current behavior
+    #
+    # When interpreted strictly, +envelope+ and +md5+ are incompatible.  So the
+    # current parsing algorithm peeks ahead after it has recieved the seventh
+    # body field.  If the next token is not the start of an +envelope+, we assume
+    # the server has incorrectly sent us a <tt>body-type-basic</tt> and return
+    # BodyTypeBasic.  As a result, what was previously BodyTypeMessage#body =>
+    # BodyTypeAttachment is now BodyTypeBasic#disposition => ContentDisposition.
     #
     class BodyTypeAttachment < Struct.new(:dsp_type, :_unused_, :param)
-      include BodyStructure
-
       # *invalid for BodyTypeAttachment*
       def media_type
         warn(<<~WARN, uplevel: 1)
@@ -1190,11 +1050,14 @@ module Net
       end
     end
 
+    deprecate_constant :BodyTypeAttachment
+
     # Net::IMAP::BodyTypeMultipart represents body structures of messages and
     # message parts, when <tt>Content-Type</tt> is <tt>multipart/*</tt>.
     class BodyTypeMultipart < Struct.new(:media_type, :subtype,
                                          :parts,
                                          :param, :disposition, :language,
+                                         :location,
                                          :extension)
       include BodyStructure
 
@@ -1265,23 +1128,24 @@ module Net
       end
     end
 
-    # === WARNING:
+    # === Obsolete
+    # BodyTypeExtension is not used and will be removed in an upcoming release.
+    #
     # >>>
-    #   BodyTypeExtension is (incorrectly) used for <tt>message/*</tt> parts
+    #   BodyTypeExtension was (incorrectly) used for <tt>message/*</tt> parts
     #   (besides <tt>message/rfc822</tt>, which correctly uses BodyTypeMessage).
     #
-    # A future release will replace this class with:
-    # * BodyTypeMessage for <tt>message/rfc822</tt> and <tt>message/global</tt>
-    # * BodyTypeBasic for any other <tt>message/*</tt>
+    #   Net::IMAP now (correctly) parses all message types (other than
+    #   <tt>message/rfc822</tt> or <tt>message/global</tt>) as BodyTypeBasic.
     class BodyTypeExtension < Struct.new(:media_type, :subtype,
                                          :params, :content_id,
                                          :description, :encoding, :size)
-      include BodyStructure
-
       def multipart?
         return false
       end
     end
 
+    deprecate_constant :BodyTypeExtension
+
   end
 end