Class: Net::IMAP::SequenceSet

Inherits:
Object
  • Object
show all
Defined in:
lib/net/imap/sequence_set.rb

Overview

An IMAP sequence set is a set of message sequence numbers or unique identifier numbers (“UIDs”). It contains numbers and ranges of numbers. The numbers are all non-zero unsigned 32-bit integers and one special value ("*") that represents the largest value in the mailbox.

Certain types of IMAP responses will contain a SequenceSet, for example the data for a "MODIFIED" ResponseCode. Some IMAP commands may receive a SequenceSet as an argument, for example IMAP#search, IMAP#fetch, and IMAP#store.

Creating sequence sets

SequenceSet.new may receive a single optional argument: a non-zero 32 bit unsigned integer, a range, a sequence-set formatted string, another SequenceSet, a Set (containing only numbers or *), or an Array containing any of these (array inputs may be nested).

set = Net::IMAP::SequenceSet.new(1)
set.valid_string  #=> "1"
set = Net::IMAP::SequenceSet.new(1..100)
set.valid_string  #=> "1:100"
set = Net::IMAP::SequenceSet.new(1...100)
set.valid_string  #=> "1:99"
set = Net::IMAP::SequenceSet.new([1, 2, 5..])
set.valid_string  #=> "1:2,5:*"
set = Net::IMAP::SequenceSet.new("1,2,3:7,5,6:10,2048,1024")
set.valid_string  #=> "1,2,3:7,5,6:10,2048,1024"
set = Net::IMAP::SequenceSet.new(1, 2, 3..7, 5, 6..10, 2048, 1024)
set.valid_string  #=> "1:10,55,1024:2048"

SequenceSet.new with no arguments creates an empty sequence set. Note that an empty sequence set is invalid in the IMAP grammar.

set = Net::IMAP::SequenceSet.new
set.empty?        #=> true
set.valid?        #=> false
set.valid_string  #!> raises DataFormatError
set << 1..10
set.empty?        #=> false
set.valid?        #=> true
set.valid_string  #=> "1:10"

Using SequenceSet.new with another SequenceSet input behaves the same as calling #dup on the other set. The input’s #string will be preserved.

input = Net::IMAP::SequenceSet.new("1,2,3:7,5,6:10,2048,1024")
copy  = Net::IMAP::SequenceSet.new(input)
input.valid_string  #=> "1,2,3:7,5,6:10,2048,1024"
copy.valid_string   #=> "1,2,3:7,5,6:10,2048,1024"
copy2 = input.dup   # same as calling new with a SequenceSet input
copy ==     input   #=> true,  same set membership
copy.eql?   input   #=> true,  same string value
copy.equal? input   #=> false, different objects

copy.normalize!
copy.valid_string   #=> "1:10,1024,2048"
copy ==   input     #=> true,  same set membership
copy.eql? input     #=> false, different string value

copy << 999
copy.valid_string   #=> "1:10,999,1024,2048"
copy ==   input     #=> false, different set membership
copy.eql? input     #=> false, different string value

Use Net::IMAP::SequenceSet() to coerce a single (optional) input. A SequenceSet input is returned without duplication, even when frozen.

set = Net::IMAP::SequenceSet()
set.string   #=> nil
set.frozen?  #=> false

# String order is preserved
set = Net::IMAP::SequenceSet("1,2,3:7,5,6:10,2048,1024")
set.valid_string  #=> "1,2,3:7,5,6:10,2048,1024"
set.frozen?       #=> false

# Other inputs are normalized
set = Net::IMAP::SequenceSet([1, 2, [3..7, 5], 6..10, 2048, 1024])
set.valid_string  #=> "1:10,1024,2048"
set.frozen?       #=> false

unfrozen = set
frozen   = set.dup.freeze
unfrozen.equal? Net::IMAP::SequenceSet(unfrozen)  #=> true
frozen.equal?   Net::IMAP::SequenceSet(frozen)    #=> true

Use ::[] to coerce one or more arguments into a valid frozen SequenceSet. A valid frozen SequenceSet is returned directly, without allocating a new object. ::[] will not create an invalid (empty) set.

Net::IMAP::SequenceSet[]     #!> raises ArgumentError
Net::IMAP::SequenceSet[nil]  #!> raises DataFormatError
Net::IMAP::SequenceSet[""]   #!> raises DataFormatError

# String order is preserved
set = Net::IMAP::SequenceSet["1,2,3:7,5,6:10,2048,1024"]
set.valid_string  #=> "1,2,3:7,5,6:10,2048,1024"
set.frozen?       #=> true

# Other inputs are normalized
set = Net::IMAP::SequenceSet[1, 2, [3..7, 5], 6..10, 2048, 1024]
set.valid_string  #=> "1:10,1024,2048"
set.frozen?       #=> true

frozen   = set
unfrozen = set.dup
frozen.equal?   Net::IMAP::SequenceSet[frozen]    #=> true
unfrozen.equal? Net::IMAP::SequenceSet[unfrozen]  #=> false

Objects which respond to to_sequence_set (such as SearchResult and ThreadMember) can be coerced to a SequenceSet with ::new, ::try_convert, ::[], or Net::IMAP::SequenceSet.

search = imap.uid_search(["SUBJECT", "hello", "NOT", "SEEN"])
seqset = Net::IMAP::SequenceSet(search) - already_fetched
fetch  = imap.uid_fetch(seqset, "FAST")

Ordered and Normalized sets

Sometimes the order of the set’s members is significant, such as with the ESORT, CONTEXT=SORT, and UIDPLUS extensions. So, when a sequence set is created from a single string (such as by the parser), that #string representation is preserved. Assigning a string with #string= or #replace will also preserve that string. Use #each_entry, #entries, or #each_ordered_number to enumerate the entries in their #string order. Hash equality (using #eql?) is based on the string representation.

Internally, SequenceSet uses a normalized uint32 set representation which sorts and de-duplicates all numbers and coalesces adjacent or overlapping entries. Many methods use this sorted set representation for O(lg n) searches. Use #each_element, #elements, #each_range, #ranges, #each_number, or #numbers to enumerate the set in sorted order. Basic object equality (using #==) is based on set membership, without regard to #entry order or #string normalization.

Most modification methods reset #string to its #normalized form, so that #entries and #elements are identical. Use #append to preserve #entries order while modifying a set.

Non-normalized sets store both representations of the set, which can more than double memory usage. Very large sequence sets should avoid denormalizing methods (such as #append) unless order is significant.

Using *

IMAP sequence sets may contain a special value "*", which represents the largest number in use. From seq-number in RFC9051 §9:

In the case of message sequence numbers, it is the number of messages in a non-empty mailbox. In the case of unique identifiers, it is the unique identifier of the last message in the mailbox or, if the mailbox is empty, the mailbox’s current UIDNEXT value.

When creating a SequenceSet, * may be input as -1, "*", :*, an endless range, or a range ending in -1. When converting to #elements, #ranges, or #numbers, it will output as either :* or an endless range. For example:

Net::IMAP::SequenceSet["1,3,*"].to_a      #=> [1, 3, :*]
Net::IMAP::SequenceSet["1,234:*"].to_a    #=> [1, 234..]
Net::IMAP::SequenceSet[1234..-1].to_a     #=> [1234..]
Net::IMAP::SequenceSet[1234..].to_a       #=> [1234..]

Net::IMAP::SequenceSet[1234..].to_s       #=> "1234:*"
Net::IMAP::SequenceSet[1234..-1].to_s     #=> "1234:*"

Use #limit to convert "*" to a maximum value. When a range includes "*", the maximum value will always be matched:

Net::IMAP::SequenceSet["9999:*"].limit(max: 25)
#=> Net::IMAP::SequenceSet["25"]

Surprising * behavior

When a set includes *, some methods may have surprising behavior.

For example, #complement treats * as its own member. This way, the #intersection of a set and its #complement will always be empty. And * is sorted as greater than any other number in the set. This is not how an IMAP server interprets the set: it will convert * to the number of messages in the mailbox, the UID of the last message in the mailbox, or UIDNEXT, as appropriate. Several methods have an argument for how * should be interpreted.

But, for example, this means that there may be overlap between a set and its complement after #limit is applied to each:

~Net::IMAP::SequenceSet["*"]  == Net::IMAP::SequenceSet[1..(2**32-1)]
~Net::IMAP::SequenceSet[1..5] == Net::IMAP::SequenceSet["6:*"]

set = Net::IMAP::SequenceSet[1..5]
(set & ~set).empty? => true

(set.limit(max: 4) & (~set).limit(max: 4)).to_a => [4]

When counting the number of numbers in a set, * will be counted as if it were equal to UINT32_MAX:

UINT32_MAX = 2**32 - 1
Net::IMAP::SequenceSet["*"].count                   => 1
Net::IMAP::SequenceSet[1..UINT32_MAX - 1, :*].count => UINT32_MAX

Net::IMAP::SequenceSet["1:*"].count                 => UINT32_MAX
Net::IMAP::SequenceSet[UINT32_MAX, :*].count        => 1
Net::IMAP::SequenceSet[UINT32_MAX..].count          => 1

Use #cardinality to count the set members wxth * counted as a distinct member:

Net::IMAP::SequenceSet[1..].cardinality            #=> UINT32_MAX + 1
Net::IMAP::SequenceSet[UINT32_MAX, :*].cardinality #=> 2
Net::IMAP::SequenceSet[UINT32_MAX..].cardinality   #=> 2

What’s here?

SequenceSet provides methods for:

Methods for Creating a SequenceSet

  • ::[]: Creates a validated frozen sequence set from one or more inputs.

  • ::new: Creates a new mutable sequence set, which may be empty (invalid).

  • ::try_convert: Calls to_sequence_set on an object and verifies that the result is a SequenceSet.

  • Net::IMAP::SequenceSet(): Coerce an input using ::try_convert or ::new.

  • ::empty: Returns a frozen empty (invalid) SequenceSet.

  • ::full: Returns a frozen SequenceSet containing every possible number.

Methods for Comparing

Comparison to another SequenceSet:

  • #==: Returns whether a given set contains the same numbers as self.

  • #eql?: Returns whether a given set uses the same #string as self.

Comparison to objects which are convertible to SequenceSet:

  • #===: Returns whether a given object is fully contained within self, or nil if the object cannot be converted to a compatible type.

  • #cover?: Returns whether a given object is fully contained within self.

  • #intersect? (aliased as #overlap?): Returns whether self and a given object have any common elements.

  • #disjoint?: Returns whether self and a given object have no common elements.

Methods for Querying

These methods do not modify self.

Set membership:

  • #include? (aliased as #member?): Returns whether a given element is contained by the set.

  • #include_star?: Returns whether the set contains *.

Minimum and maximum value elements:

  • #min: Returns one or more of the lowest numbers in the set.

  • #max: Returns one or more of the highest numbers in the set.

  • #minmax: Returns the lowest and highest numbers in the set.

Accessing value by offset in sorted set:

  • #[] (aliased as #slice): Returns the number or consecutive subset at a given offset or range of offsets in the sorted set.

  • #at: Returns the number at a given offset in the sorted set.

  • #find_index: Returns the given number’s offset in the sorted set.

Accessing value by offset in ordered entries

  • #ordered_at: Returns the number at a given offset in the ordered entries.

  • #find_ordered_index: Returns the index of the given number’s first occurrence in entries.

Set cardinality:

  • #cardinality: Returns the number of distinct members in the set. * is counted as its own member, distinct from UINT32_MAX.

  • #count: Returns the count of distinct numbers in the set. * is counted as equal to UINT32_MAX.

  • #empty?: Returns whether the set has no members. IMAP syntax does not allow empty sequence sets.

  • #valid?: Returns whether the set has any members.

  • #full?: Returns whether the set contains every possible value, including *.

Denormalized properties:

  • #normalized?: Returns whether #entries are sorted, deduplicated, and coalesced, and all #string entries are in normalized form.

  • #has_duplicates?: Returns whether the ordered entries repeat any numbers.

  • #size: Returns the total size of all #entries, including repeated numbers. * is counted as its own member, distinct from UINT32_MAX.

  • #count_with_duplicates: Returns the count of numbers in the ordered #entries, including repeated numbers. * is counted as equal to UINT32_MAX.

  • #count_duplicates: Returns the count of repeated numbers in the ordered #entries. * is counted as equal to UINT32_MAX.

Methods for Iterating

Normalized (sorted and coalesced):

  • #each_element: Yields each number and range in the set, sorted and coalesced, and returns self.

  • #elements (aliased as #to_a): Returns an Array of every number and range in the set, sorted and coalesced.

  • #each_range: Yields each element in the set as a Range and returns self.

  • #ranges: Returns an Array of every element in the set, converting numbers into ranges of a single value.

  • #each_number: Yields each number in the set and returns self.

  • #numbers: Returns an Array with every number in the set, expanding ranges into all of their contained numbers.

  • #to_set: Returns a Set containing all of the #numbers in the set.

Order preserving:

  • #each_entry: Yields each number and range in the set, unsorted and without deduplicating numbers or coalescing ranges, and returns self.

  • #entries: Returns an Array of every number and range in the set, unsorted and without deduplicating numbers or coalescing ranges.

  • #each_ordered_number: Yields each number in the ordered entries and returns self.

Methods for Set Operations

These methods do not modify self.

  • #| (aliased as #union and #+): Returns a new set combining all members from self with all members from the other set.

  • #& (aliased as #intersection): Returns a new set containing all members common to self and the other set.

  • #- (aliased as #difference): Returns a copy of self with all members in the other set removed.

  • #^ (aliased as #xor): Returns a new set containing all members from self and the other set except those common to both.

  • #~ (aliased as #complement): Returns a new set containing all members that are not in self

  • #above: Return a copy of self which only contains numbers above a given number.

  • #below: Return a copy of self which only contains numbers below a given value.

  • #limit: Returns a copy of self which has replaced * with a given maximum value and removed all members over that maximum.

Methods for Assigning

These methods add or replace numbers in self.

Normalized (sorted and coalesced):

These methods always update #string to be fully sorted and coalesced.

  • #add (aliased as #<<): Adds a given element to the set; returns self.

  • #add?: If the given element is not fully included the set, adds it and returns self; otherwise, returns nil.

  • #merge: In-place set #union. Adds all members of the given sets into this set; returns self.

  • #complement!: In-place set #complement. Replaces the contents of this set with its own #complement; returns self.

  • #xor!: In-place XOR operation. Adds numbers that are unique to the other set and removes numbers that are common to both; returns self.

Order preserving:

These methods may cause #string to not be sorted or coalesced.

  • #append: Adds the given entry to the set, appending it to the existing string, and returns self.

  • #string=: Assigns a new #string value and replaces #elements to match.

  • #replace: Replaces the contents of the set with the contents of a given object.

Methods for Deleting

These methods remove numbers from self, and update #string to be fully sorted and coalesced.

  • #clear: Removes all elements in the set; returns self.

  • #delete: Removes a given element from the set; returns self.

  • #delete?: If the given element is included in the set, removes it and returns it; otherwise, returns nil.

  • #delete_at: Removes the number at a given offset.

  • #intersect!: In-place set #intersection. Removes numbers that are not in the given set; returns self.

  • #slice!: Removes the number or consecutive numbers at a given offset or range of offsets.

  • #subtract: In-place set #difference. Removes all members of the given sets from this set; returns self.

  • #limit!: Replaces * with a given maximum value and removes all members over that maximum; returns self.

Methods for IMAP String Formatting

  • #to_s: Returns the sequence-set string, or an empty string when the set is empty.

  • #string: Returns the sequence-set string, or nil when empty.

  • #valid_string: Returns the sequence-set string, or raises DataFormatError when the set is empty.

  • #normalized_string: Returns a sequence-set string with its elements sorted and coalesced, or nil when the set is empty.

  • #normalize: Returns a new set with this set’s normalized sequence-set representation.

  • #normalize!: Updates #string to its normalized sequence-set representation and returns self.

Constant Summary collapse

UINT32_MAX =

The largest possible non-zero unsigned 32-bit integer

2**32 - 1

Class Method Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(input = nil) ⇒ SequenceSet

Create a new SequenceSet object from input, which may be another SequenceSet, an IMAP formatted sequence-set string, a non-zero 32 bit unsigned integer, a range, :*, a Set of numbers or *, an object that responds to to_sequence_set (such as SearchResult) or an Array of these (array inputs may be nested).

set = Net::IMAP::SequenceSet.new(1)
set.valid_string  #=> "1"
set = Net::IMAP::SequenceSet.new(1..100)
set.valid_string  #=> "1:100"
set = Net::IMAP::SequenceSet.new(1...100)
set.valid_string  #=> "1:99"
set = Net::IMAP::SequenceSet.new([1, 2, 5..])
set.valid_string  #=> "1:2,5:*"
set = Net::IMAP::SequenceSet.new("1,2,3:7,5,6:10,2048,1024")
set.valid_string  #=> "1,2,3:7,5,6:10,2048,1024"
set = Net::IMAP::SequenceSet.new(1, 2, 3..7, 5, 6..10, 2048, 1024)
set.valid_string  #=> "1:10,1024,2048"

With no arguments (or nil) creates an empty sequence set. Note that an empty sequence set is invalid in the IMAP grammar.

set = Net::IMAP::SequenceSet.new
set.empty?        #=> true
set.valid?        #=> false
set.valid_string  #!> raises DataFormatError
set << 1..10
set.empty?        #=> false
set.valid?        #=> true
set.valid_string  #=> "1:10"

When input is a SequenceSet, ::new behaves the same as calling #dup on that other set. The input’s #string will be preserved.

input = Net::IMAP::SequenceSet.new("1,2,3:7,5,6:10,2048,1024")
copy  = Net::IMAP::SequenceSet.new(input)
input.valid_string  #=> "1,2,3:7,5,6:10,2048,1024"
copy.valid_string   #=> "1,2,3:7,5,6:10,2048,1024"
copy2 = input.dup   # same as calling new with a SequenceSet input
copy ==     input   #=> true,  same set membership
copy.eql?   input   #=> true,  same string value
copy.equal? input   #=> false, different objects

copy.normalize!
copy.valid_string   #=> "1:10,1024,2048"
copy ==   input     #=> true,  same set membership
copy.eql? input     #=> false, different string value

copy << 999
copy.valid_string   #=> "1:10,999,1024,2048"
copy ==   input     #=> false, different set membership
copy.eql? input     #=> false, different string value

Alternative set creation methods

  • ::[] returns a frozen validated (non-empty) SequenceSet, without allocating a new object when the input is already a valid frozen SequenceSet.

  • Net::IMAP::SequenceSet() coerces an input to SequenceSet, without allocating a new object when the input is already a SequenceSet.

  • ::try_convert calls to_sequence_set on inputs that support it and returns nil for inputs that don’t.

  • ::empty and ::full both return frozen singleton sets which can be combined with set operations (#|, #&, #^, #-, etc) to make new sets.

See SequenceSet@Creating+sequence+sets.



558
559
560
561
562
# File 'lib/net/imap/sequence_set.rb', line 558

def initialize(input = nil)
  @set_data = new_set_data
  @string = nil
  replace(input) unless input.nil?
end

Class Method Details

.[](first, *rest) ⇒ Object

:call-seq:

SequenceSet[*inputs] -> valid frozen sequence set

Returns a frozen SequenceSet, constructed from inputs.

When only a single valid frozen SequenceSet is given, that same set is returned.

An empty SequenceSet is invalid and will raise a DataFormatError.

Use ::new to create a mutable or empty SequenceSet.

Related: ::new, Net::IMAP::SequenceSet(), ::try_convert



455
456
457
458
459
460
461
462
# File 'lib/net/imap/sequence_set.rb', line 455

def [](first, *rest)
  if rest.empty?
    set = try_convert(first)&.validate
    set&.frozen? ? set : (set&.dup || new(first).validate).freeze
  else
    new(first).merge(*rest).validate.freeze
  end
end

.emptyObject

Returns a frozen empty set singleton. Note that valid IMAP sequence sets cannot be empty, so this set is invalid.



485
# File 'lib/net/imap/sequence_set.rb', line 485

def empty; EMPTY end

.fullObject

Returns a frozen full set singleton: "1:*"



488
# File 'lib/net/imap/sequence_set.rb', line 488

def full;  FULL end

.try_convert(obj) ⇒ Object

:call-seq:

SequenceSet.try_convert(obj) -> sequence set or nil

If obj is a SequenceSet, returns obj. If obj responds_to to_sequence_set, calls obj.to_sequence_set and returns the result. Otherwise returns nil.

If obj.to_sequence_set doesn’t return a SequenceSet or nil, an exception is raised.

Related: Net::IMAP::SequenceSet(), ::new, ::[]

Raises:



475
476
477
478
479
480
481
# File 'lib/net/imap/sequence_set.rb', line 475

def try_convert(obj)
  return obj if obj.is_a?(SequenceSet)
  return nil unless obj.respond_to?(:to_sequence_set)
  return nil unless obj = obj.to_sequence_set
  return obj if obj.is_a?(SequenceSet)
  raise DataFormatError, "invalid object returned from to_sequence_set"
end

Instance Method Details

#&(other) ⇒ Object Also known as: intersection

:call-seq:

self & other        -> sequence set
intersection(other) -> sequence set

Returns a new sequence set containing only the numbers common to this set and other.

other may be any object that would be accepted by ::new.

Net::IMAP::SequenceSet[1..5] & [2, 4, 6]
#=> Net::IMAP::SequenceSet["2,4"]

Related: #intersect?, #|, #-, #^, #~

Set identities

lhs & rhs is equivalent to:

  • rhs & lhs (commutative)

  • ~(~lhs | ~rhs) (De Morgan’s Law)

  • lhs - ~rhs

  • lhs - (lhs - rhs)

  • lhs - (lhs ^ rhs)

  • lhs ^ (lhs - rhs)



913
# File 'lib/net/imap/sequence_set.rb', line 913

def &(other) remain_frozen dup.intersect! other end

#-(other) ⇒ Object Also known as: difference

:call-seq:

self - other      -> sequence set
difference(other) -> sequence set

Returns a new sequence set built by duplicating this set and removing every number that appears in other.

other may be any object that would be accepted by ::new.

Net::IMAP::SequenceSet[1..5] - 2 - 4 - 6
#=> Net::IMAP::SequenceSet["1,3,5"]

Related: #subtract, #|, #&, #^, #~

Set identities

lhs - rhs is equivalent to:

  • ~rhs - ~lhs

  • lhs & ~rhs

  • ~(~lhs | rhs)

  • lhs & (lhs ^ rhs)

  • lhs ^ (lhs & rhs)

  • rhs ^ (lhs | rhs)



887
# File 'lib/net/imap/sequence_set.rb', line 887

def -(other) remain_frozen dup.subtract other end

#==(other) ⇒ Object

:call-seq: self == other -> true or false

Returns true when the other SequenceSet represents the same message identifiers. Encoding difference—such as order, overlaps, or duplicates—are ignored.

Net::IMAP::SequenceSet["1:3"]   == Net::IMAP::SequenceSet["1:3"]
#=> true
Net::IMAP::SequenceSet["1,2,3"] == Net::IMAP::SequenceSet["1:3"]
#=> true
Net::IMAP::SequenceSet["1,3"]   == Net::IMAP::SequenceSet["3,1"]
#=> true
Net::IMAP::SequenceSet["9,1:*"] == Net::IMAP::SequenceSet["1:*"]
#=> true

Related: #eql?, #normalize



675
676
677
678
# File 'lib/net/imap/sequence_set.rb', line 675

def ==(other)
  self.class == other.class &&
    (to_s == other.to_s || set_data == other.set_data)
end

#===(other) ⇒ Object

:call-seq: self === other -> true | false | nil

Returns whether other is contained within the set. other may be any object that would be accepted by ::new. Returns nil if StandardError is raised while converting other to a comparable type.

Related: #cover?, #include?, #include_star?



706
707
708
709
710
# File 'lib/net/imap/sequence_set.rb', line 706

def ===(other)
  cover?(other)
rescue
  nil
end

#[](index, length = nil) ⇒ Object Also known as: slice

:call-seq:

seqset[index]         -> integer or :* or nil
slice(index)          -> integer or :* or nil
seqset[start, length] -> sequence set or nil
slice(start, length)  -> sequence set or nil
seqset[range]         -> sequence set or nil
slice(range)          -> sequence set or nil

Returns a number or a subset from the sorted set, without modifying the set.

When an Integer argument index is given, the number at offset index in the sorted set is returned:

set = Net::IMAP::SequenceSet["10:15,20:23,26"]
set[0]   #=> 10
set[5]   #=> 15
set[10]  #=> 26

If index is negative, it counts relative to the end of the sorted set:

set = Net::IMAP::SequenceSet["10:15,20:23,26"]
set[-1]  #=> 26
set[-3]  #=> 22
set[-6]  #=> 15

If index is out of range, nil is returned.

set = Net::IMAP::SequenceSet["10:15,20:23,26"]
set[11]  #=> nil
set[-12] #=> nil

The result is based on the sorted and de-duplicated set, not on the ordered #entries in #string.

set = Net::IMAP::SequenceSet["12,20:23,11:16,21"]
set[0]   #=> 11
set[-1]  #=> 23

Related: #at



1540
1541
1542
1543
1544
1545
# File 'lib/net/imap/sequence_set.rb', line 1540

def [](index, length = nil)
  if    length              then slice_length(index, length)
  elsif index.is_a?(Range)  then slice_range(index)
  else                           at(index)
  end
end

#^(other) ⇒ Object Also known as: xor

:call-seq:

self ^ other -> sequence set
xor(other)   -> sequence set

Returns a new sequence set containing numbers that are exclusive between this set and other.

other may be any object that would be accepted by ::new.

Net::IMAP::SequenceSet[1..5] ^ [2, 4, 6]
#=> Net::IMAP::SequenceSet["1,3,5:6"]

Related: #|, #&, #-, #~

Set identities

lhs ^ rhs is equivalent to:

  • rhs ^ lhs (commutative)

  • ~lhs ^ ~rhs

  • (lhs | rhs) - (lhs & rhs)

  • (lhs - rhs) | (rhs - lhs)

  • (lhs ^ other) ^ (other ^ rhs)



938
# File 'lib/net/imap/sequence_set.rb', line 938

def ^(other) remain_frozen dup.xor! other end

#above(num) ⇒ Object

Returns a copy of self which only contains the numbers above num.

Net::IMAP::SequenceSet["5,10:22,50"].above(10) # to_s => "11:22,50"
Net::IMAP::SequenceSet["5,10:22,50"].above(20) # to_s => "21:22,50
Net::IMAP::SequenceSet["5,10:22,50"].above(30) # to_s => "50"

This returns the same result as #intersection with ((num+1)..) or #difference with (..num).

Net::IMAP::SequenceSet["5,10:22,50"] & (11..)   # to_s => "11:22,50"
Net::IMAP::SequenceSet["5,10:22,50"] - (..10)   # to_s => "11:22,50"
Net::IMAP::SequenceSet["5,10:22,50"] & (21..)   # to_s => "21:22,50"
Net::IMAP::SequenceSet["5,10:22,50"] - (..20)   # to_s => "21:22,50"

Related: #above, #-, #&



1564
1565
1566
1567
1568
# File 'lib/net/imap/sequence_set.rb', line 1564

def above(num)
  NumValidator.valid_nz_number?(num) or
    raise ArgumentError, "not a valid sequence set number"
  difference(..num)
end

#add(element) ⇒ Object Also known as: <<

:call-seq:

add(element)   -> self
self << other -> self

Adds a range or number to the set and returns self.

#string will be regenerated. Use #merge to add many elements at once.

Use #append to append new elements to #string. See SequenceSet@Ordered+and+Normalized+sets.

Related: #add?, #merge, #union, #append



976
977
978
979
980
# File 'lib/net/imap/sequence_set.rb', line 976

def add(element)
  modifying! # short-circuit before import_run
  add_run import_run element
  normalize!
end

#add?(element) ⇒ Boolean

:call-seq: add?(element) -> self or nil

Adds a range or number to the set and returns self. Returns nil when the element is already included in the set.

#string will be regenerated. Use #merge to add many elements at once.

Related: #add, #merge, #union, #include?

Returns:

  • (Boolean)


1048
1049
1050
1051
# File 'lib/net/imap/sequence_set.rb', line 1048

def add?(element)
  modifying! # short-circuit before include?
  add element unless include? element
end

#append(entry) ⇒ Object

Adds a range or number to the set and returns self.

Unlike #add, #merge, or #union, the new value is appended to #string. This may result in a #string which has duplicates or is out-of-order.

set = Net::IMAP::SequenceSet.new
set.append(1..2)  # => Net::IMAP::SequenceSet("1:2")
set.append(5)     # => Net::IMAP::SequenceSet("1:2,5")
set.append(4)     # => Net::IMAP::SequenceSet("1:2,5,4")
set.append(3)     # => Net::IMAP::SequenceSet("1:2,5,4,3")
set.append(2)     # => Net::IMAP::SequenceSet("1:2,5,4,3,2")

If entry is a string, it will be converted into normal form.

set = Net::IMAP::SequenceSet("4:5,1:2")
set.append("6:6") # => Net::IMAP::SequenceSet("4:5,1:2,6")
set.append("9:8") # => Net::IMAP::SequenceSet("4:5,1:2,6,8:9")

If entry adjacently follows the last entry, they will coalesced:

set = Net::IMAP::SequenceSet.new("2,1,9:10")
set.append(11..12) # => Net::IMAP::SequenceSet("2,1,9:12")

Non-normalized sets store the string in addition to an internal normalized uint32 set representation. This can more than double memory usage, so large sets should avoid using #append unless preserving order is required. See SequenceSet@Ordered+and+Normalized+sets.

Related: #add, #merge, #union



1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
# File 'lib/net/imap/sequence_set.rb', line 1011

def append(entry)
  modifying! # short-circuit before import_minmax
  minmax = import_minmax entry
  adj = minmax.first - 1
  if @string.nil? && (runs.empty? || max_num <= adj)
    # append to elements or coalesce with last element
    add_minmax minmax
    return self
  elsif @string.nil?
    # generate string for out-of-order append
    head, comma = normalized_string, ","
  else
    # @string already exists... maybe coalesce with last entry
    head, comma, last_entry = @string.rpartition(",")
    last_min, last_max = import_minmax last_entry
    if last_max == adj
      # coalesce with last entry
      minmax[0] = last_min
    else
      # append to existing string
      head, comma = @string, ","
    end
  end
  entry = export_minmax minmax
  add_minmax minmax
  @string = -"#{head}#{comma}#{entry}"
  self
end

#at(index) ⇒ Object

:call-seq: at(index) -> integer or nil

Returns the number at the given index in the sorted set, without modifying the set.

index is interpreted the same as in #[], except that #at only allows a single integer argument.

Related: #[], #slice, #ordered_at



1484
1485
1486
# File 'lib/net/imap/sequence_set.rb', line 1484

def at(index)
  seek_number_in_minmaxes(minmaxes, index)
end

#below(num) ⇒ Object

Returns a copy of self which only contains numbers below num.

Net::IMAP::SequenceSet["5,10:22,50"].below(10) # to_s => "5"
Net::IMAP::SequenceSet["5,10:22,50"].below(20) # to_s => "5,10:19"
Net::IMAP::SequenceSet["5,10:22,50"].below(30) # to_s => "5,10:22"

This returns the same result as #intersection with (..(num-1)) or #difference with (num..).

Net::IMAP::SequenceSet["5,10:22,50"] & (..9)    # to_s => "5"
Net::IMAP::SequenceSet["5,10:22,50"] - (10..)   # to_s => "5"
Net::IMAP::SequenceSet["5,10:22,50"] & (..19)   # to_s => "5,10:19"
Net::IMAP::SequenceSet["5,10:22,50"] - (20..)   # to_s => "5,10:19"

When the set does not contain *, #below is identical to #limit with max: num - 1. When the set does contain *, #below always drops it from the result. Use #limit when the IMAP semantics for * must be enforced.

Net::IMAP::SequenceSet["5,10:22,50"].below(30)      # to_s => "5,10:22"
Net::IMAP::SequenceSet["5,10:22,50"].limit(max: 29) # to_s => "5,10:22"
Net::IMAP::SequenceSet["5,10:22,*"].below(30)       # to_s => "5,10:22"
Net::IMAP::SequenceSet["5,10:22,*"].limit(max: 29)  # to_s => "5,10:22,29"

Related: #above, #-, #&, #limit



1595
1596
1597
1598
1599
# File 'lib/net/imap/sequence_set.rb', line 1595

def below(num)
  NumValidator.valid_nz_number?(num) or
    raise ArgumentError, "not a valid sequence set number"
  difference(num..)
end

#cardinalityObject

Returns the number of members in the set.

Unlike #count, "*" is considered to be distinct from 2³² - 1 (the maximum 32-bit unsigned integer value).

set = Net::IMAP::SequenceSet[1..10]
set.count        #=> 10
set.cardinality  #=> 10

set = Net::IMAP::SequenceSet["4294967295,*"]
set.count        #=> 1
set.cardinality  #=> 2

set = Net::IMAP::SequenceSet[1..]
set.count        #=> 4294967295
set.cardinality  #=> 4294967296

Related: #count, #count_with_duplicates



1346
# File 'lib/net/imap/sequence_set.rb', line 1346

def cardinality = minmaxes.sum(runs.count) { _2 - _1 }

#clearObject

Removes all elements and returns self.



565
566
567
568
569
570
# File 'lib/net/imap/sequence_set.rb', line 565

def clear
  modifying! # redundant check (normalizes the error message for JRuby)
  set_data.clear
  @string = nil
  self
end

#complement!Object

:call-seq: complement! -> self

In-place set #complement. Replaces the contents of this set with its own #complement. It will contain all possible values except for those currently in the set.

Related: #complement



1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
# File 'lib/net/imap/sequence_set.rb', line 1648

def complement!
  modifying! # short-circuit before querying
  return replace(self.class.full) if empty?
  return clear                    if full?
  flat = minmaxes.flat_map { [_1 - 1, _2 + 1] }
  if flat.first < 1         then flat.shift else flat.unshift 1        end
  if STAR_INT   < flat.last then flat.pop   else flat.push    STAR_INT end
  replace_minmaxes flat.each_slice(2).to_a
  normalize!
end

#countObject

Returns the count of distinct #numbers in the set.

Unlike #cardinality, "*" is considered to be equal to 2³² - 1 (the maximum 32-bit unsigned integer value).

set = Net::IMAP::SequenceSet[1..10]
set.count        #=> 10
set.cardinality  #=> 10

set = Net::IMAP::SequenceSet["4294967295,*"]
set.count        #=> 1
set.cardinality  #=> 2

set = Net::IMAP::SequenceSet[1..]
set.count        #=> 4294967295
set.cardinality  #=> 4294967296

Related: #cardinality, #count_with_duplicates



1366
1367
1368
# File 'lib/net/imap/sequence_set.rb', line 1366

def count
  cardinality + (include_star? && include?(UINT32_MAX) ? -1 : 0)
end

#count_duplicatesObject

Returns the count of repeated numbers in the ordered #entries, the difference between #count_with_duplicates and #count.

When #string is normalized, this is zero.

Related: #entries, #count_with_duplicates, #has_duplicates?



1431
1432
1433
1434
# File 'lib/net/imap/sequence_set.rb', line 1431

def count_duplicates
  return 0 unless @string
  count_with_duplicates - count
end

#count_with_duplicatesObject

Returns the count of numbers in the ordered #entries, including any repeated numbers.

When #string is normalized, this returns the same as #count. Like #count, "*" is considered to be equal to 2³² - 1 (the maximum 32-bit unsigned integer value).

In a range, "*" is not considered a duplicate:

set = Net::IMAP::SequenceSet["4294967295:*"]
set.count_with_duplicates  #=> 1
set.size                   #=> 2
set.count                  #=> 1
set.cardinality            #=> 2

In a separate entry, "*" is considered a duplicate:

set = Net::IMAP::SequenceSet["4294967295,*"]
set.count_with_duplicates  #=> 2
set.size                   #=> 2
set.count                  #=> 1
set.cardinality            #=> 2

Related: #count, #cardinality, #size, #count_duplicates, #has_duplicates?, #entries



1393
1394
1395
1396
1397
1398
# File 'lib/net/imap/sequence_set.rb', line 1393

def count_with_duplicates
  return count unless @string
  each_entry_minmax.sum {|min, max|
    max - min + ((max == STAR_INT && min != STAR_INT) ? 0 : 1)
  }
end

#cover?(other) ⇒ Boolean

:call-seq: cover?(other) -> true | false | nil

Returns whether other is contained within the set. other may be any object that would be accepted by ::new.

Related: #===, #include?, #include_star?, #intersect?

Returns:

  • (Boolean)


718
# File 'lib/net/imap/sequence_set.rb', line 718

def cover?(other) import_runs(other).none? { !include_run?(_1) } end

#deconstructObject

Returns an array with #normalized_string when valid and an empty array otherwise.



616
# File 'lib/net/imap/sequence_set.rb', line 616

def deconstruct; valid? ? [normalized_string] : [] end

#delete(element) ⇒ Object

:call-seq: delete(element) -> self

Deletes the given range or number from the set and returns self.

#string will be regenerated after deletion. Use #subtract to remove many elements at once.

Related: #delete?, #delete_at, #subtract, #difference



1061
1062
1063
1064
1065
# File 'lib/net/imap/sequence_set.rb', line 1061

def delete(element)
  modifying! # short-circuit before import_run
  subtract_run import_run element
  normalize!
end

#delete?(element) ⇒ Boolean

:call-seq:

delete?(number) -> integer or nil
delete?(star)   -> :* or nil
delete?(range)  -> sequence set or nil

Removes a specified value from the set, and returns the removed value. Returns nil if nothing was removed.

Returns an integer when the specified number argument was removed:

set = Net::IMAP::SequenceSet.new [5..10, 20]
set.delete?(7)      #=> 7
set                 #=> #<Net::IMAP::SequenceSet "5:6,8:10,20">
set.delete?("20")   #=> 20
set                 #=> #<Net::IMAP::SequenceSet "5:6,8:10">
set.delete?(30)     #=> nil

Returns :* when * or -1 is specified and removed:

set = Net::IMAP::SequenceSet.new "5:9,20,35,*"
set.delete?(-1)  #=> :*
set              #=> #<Net::IMAP::SequenceSet "5:9,20,35">

And returns a new SequenceSet when a range is specified:

set = Net::IMAP::SequenceSet.new [5..10, 20]
set.delete?(9..)  #=> #<Net::IMAP::SequenceSet "9:10,20">
set               #=> #<Net::IMAP::SequenceSet "5:8">
set.delete?(21..) #=> nil

#string will be regenerated after deletion.

Related: #delete, #delete_at, #subtract, #difference, #disjoint?

Returns:

  • (Boolean)


1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
# File 'lib/net/imap/sequence_set.rb', line 1099

def delete?(element)
  modifying! # short-circuit before import_minmax
  element = input_try_convert(element)
  minmax = import_minmax element
  if number_input?(element)
    return unless include_minmax? minmax
    subtract_minmax minmax
    normalize!
    export_num minmax.first
  else
    copy = dup
    subtract_minmax minmax
    normalize!
    copy if copy.subtract(self).valid?
  end
end

#delete_at(index) ⇒ Object

:call-seq: delete_at(index) -> number or :* or nil

Deletes a number the set, indicated by the given index. Returns the number that was removed, or nil if nothing was removed.

#string will be regenerated after deletion.

Related: #delete, #delete?, #slice!, #subtract, #difference



1124
1125
1126
# File 'lib/net/imap/sequence_set.rb', line 1124

def delete_at(index)
  slice! Integer(index.to_int)
end

#disjoint?(other) ⇒ Boolean

Returns true if the set and a given object have no common elements, false otherwise.

Net::IMAP::SequenceSet["5:10"].disjoint? "7,9,11" #=> false
Net::IMAP::SequenceSet["5:10"].disjoint? "11:33"  #=> true

Related: #intersection, #intersect?

Returns:

  • (Boolean)


773
774
775
# File 'lib/net/imap/sequence_set.rb', line 773

def disjoint?(other)
  empty? || import_runs(other).none? { intersect_run? _1 }
end

#each_elementObject

Yields each number or range (or :*) in #elements to the block and returns self. Returns an enumerator when called without a block.

The returned numbers are sorted and de-duplicated, even when the input #string is not. See #normalize, SequenceSet@Ordered+and+Normalized+sets.

Related: #elements, #each_entry



1271
1272
1273
1274
1275
# File 'lib/net/imap/sequence_set.rb', line 1271

def each_element # :yields: integer or range or :*
  return to_enum(__method__) unless block_given?
  runs.each do yield export_run_entry _1 end
  self
end

#each_entry(&block) ⇒ Object

Yields each number or range in #string to the block and returns self. Returns an enumerator when called without a block.

The entries are yielded in the same order they appear in #string, with no sorting, deduplication, or coalescing. When #string is in its normalized form, this will yield the same values as #each_element.

See SequenceSet@Ordered+and+Normalized+sets.

Related: #entries, #each_element



1259
1260
1261
1262
# File 'lib/net/imap/sequence_set.rb', line 1259

def each_entry(&block) # :yields: integer or range or :*
  return to_enum(__method__) unless block_given?
  each_entry_run do yield export_run_entry _1 end
end

#each_number(&block) ⇒ Object

Yields each number in #numbers to the block and returns self. If the set contains a *, RangeError will be raised.

Returns an enumerator when called without a block (even if the set contains *).

Related: #numbers, #each_ordered_number

Raises:

  • (RangeError)


1299
1300
1301
1302
1303
1304
# File 'lib/net/imap/sequence_set.rb', line 1299

def each_number(&block) # :yields: integer
  return to_enum(__method__) unless block_given?
  raise RangeError, '%s contains "*"' % [self.class] if include_star?
  minmaxes.each do each_number_in_minmax _1, _2, &block end
  self
end

#each_ordered_number(&block) ⇒ Object

Yields each number in #entries to the block and returns self. If the set contains a *, RangeError will be raised.

Returns an enumerator when called without a block (even if the set contains *).

Related: #entries, #each_number

Raises:

  • (RangeError)


1313
1314
1315
1316
1317
# File 'lib/net/imap/sequence_set.rb', line 1313

def each_ordered_number(&block)
  return to_enum(__method__) unless block_given?
  raise RangeError, '%s contains "*"' % [self.class] if include_star?
  each_entry_minmax do each_number_in_minmax _1, _2, &block end
end

#each_rangeObject

Yields each range in #ranges to the block and returns self. Returns an enumerator when called without a block.

Related: #ranges



1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
# File 'lib/net/imap/sequence_set.rb', line 1281

def each_range # :yields: range
  return to_enum(__method__) unless block_given?
  minmaxes.each do |min, max|
    if    min == STAR_INT then yield :*..
    elsif max == STAR_INT then yield min..
    else                       yield min..max
    end
  end
  self
end

#elementsObject Also known as: to_a

Returns an array of ranges and integers and :*.

The returned elements are sorted and coalesced, even when the input #string is not. * will sort last. See #normalize, SequenceSet@Ordered+and+Normalized+sets.

By itself, * translates to :*. A range containing * translates to an endless range. Use #limit to translate both cases to a maximum value.

Net::IMAP::SequenceSet["2,5:9,6,*,12:11"].elements
#=> [2, 5..9, 11..12, :*]

Related: #each_element, #ranges, #numbers



1200
# File 'lib/net/imap/sequence_set.rb', line 1200

def elements; each_element.to_a end

#empty?Boolean

Returns true if the set contains no elements

Returns:

  • (Boolean)


834
# File 'lib/net/imap/sequence_set.rb', line 834

def empty?; runs.empty? end

#encode_with(coder) ⇒ Object

For YAML serialization



1847
1848
1849
1850
# File 'lib/net/imap/sequence_set.rb', line 1847

def encode_with(coder) # :nodoc:
  # we can perfectly reconstruct from the string
  coder['string'] = to_s
end

#entriesObject

Returns an array of ranges and integers and :*.

The entries are in the same order they appear in #string, with no sorting, deduplication, or coalescing. When #string is in its normalized form, this will return the same result as #elements. This is useful when the given order is significant, for example in a ESEARCH response to IMAP#sort.

See SequenceSet@Ordered+and+Normalized+sets.

Related: #each_entry, #elements



1184
# File 'lib/net/imap/sequence_set.rb', line 1184

def entries; each_entry.to_a end

#eql?(other) ⇒ Boolean

:call-seq: eql?(other) -> true or false

Hash equality requires the same encoded #string representation.

Net::IMAP::SequenceSet["1:3"]  .eql? Net::IMAP::SequenceSet["1:3"]
#=> true
Net::IMAP::SequenceSet["1,2,3"].eql? Net::IMAP::SequenceSet["1:3"]
#=> false
Net::IMAP::SequenceSet["1,3"]  .eql? Net::IMAP::SequenceSet["3,1"]
#=> false
Net::IMAP::SequenceSet["9,1:*"].eql? Net::IMAP::SequenceSet["1:*"]
#=> false

Related: #==, #normalize

Returns:

  • (Boolean)


694
# File 'lib/net/imap/sequence_set.rb', line 694

def eql?(other) self.class == other.class && string == other.string end

#find_index(number) ⇒ Object

Returns the (sorted and deduplicated) index of number in the set, or nil if number isn’t in the set.

Related: #[], #at, #find_ordered_index



1452
1453
1454
1455
1456
1457
1458
1459
# File 'lib/net/imap/sequence_set.rb', line 1452

def find_index(number)
  number = import_num number
  each_minmax_with_index(minmaxes) do |min, max, idx_min|
    number <  min and return nil
    number <= max and return export_num(idx_min + (number - min))
  end
  nil
end

#find_ordered_index(number) ⇒ Object

Returns the first index of number in the ordered #entries, or nil if number isn’t in the set.

Related: #find_index



1465
1466
1467
1468
1469
1470
1471
1472
1473
# File 'lib/net/imap/sequence_set.rb', line 1465

def find_ordered_index(number)
  number = import_num number
  each_minmax_with_index(each_entry_minmax) do |min, max, idx_min|
    if min <= number && number <= max
      return export_num(idx_min + (number - min))
    end
  end
  nil
end

#freezeObject

Freezes and returns the set. A frozen SequenceSet is Ractor-safe.



653
654
655
656
657
# File 'lib/net/imap/sequence_set.rb', line 653

def freeze
  return self if frozen?
  freeze_set_data
  super
end

#full?Boolean

Returns true if the set contains every possible element.

Returns:

  • (Boolean)


837
# File 'lib/net/imap/sequence_set.rb', line 837

def full?; set_data == FULL_SET_DATA end

#has_duplicates?Boolean

:call-seq: has_duplicates? -> true | false

Returns whether or not the ordered #entries repeat any numbers.

Always returns false when #string is normalized.

Related: #entries, #count_with_duplicates, #count_duplicates

Returns:

  • (Boolean)


1443
1444
1445
1446
# File 'lib/net/imap/sequence_set.rb', line 1443

def has_duplicates?
  return false unless @string
  count_with_duplicates != count
end

#hashObject

See #eql?



697
# File 'lib/net/imap/sequence_set.rb', line 697

def hash; [self.class, string].hash end

#include?(element) ⇒ Boolean Also known as: member?

Returns true when a given number or range is in self, and false otherwise. Returns nil when number isn’t a valid SequenceSet element (Integer, Range, *, sequence-set string).

set = Net::IMAP::SequenceSet["5:10,100,111:115"]
set.include? 1      #=> false
set.include? 5..10  #=> true
set.include? 11..20 #=> false
set.include? 100    #=> true
set.include? 6      #=> true, covered by "5:10"
set.include? 6..9   #=> true, covered by "5:10"
set.include? "6:9"  #=> true, strings are parsed
set.include? 4..9   #=> false, intersection is not sufficient
set.include? "*"    #=> false, use #limit to re-interpret "*"
set.include? -1     #=> false, -1 is interpreted as "*"

set = Net::IMAP::SequenceSet["5:10,100,111:*"]
set.include? :*     #=> true
set.include? "*"    #=> true
set.include? -1     #=> true
set.include?(200..) #=> true
set.include?(100..) #=> false

Related: #include_star?, #cover?, #===, #intersect?

Returns:

  • (Boolean)


744
745
746
747
# File 'lib/net/imap/sequence_set.rb', line 744

def include?(element)
  run = import_run element rescue nil
  !!include_run?(run) if run
end

#include_star?Boolean

Returns true when the set contains *.

Returns:

  • (Boolean)


752
# File 'lib/net/imap/sequence_set.rb', line 752

def include_star?; max_num == STAR_INT end

#init_with(coder) ⇒ Object

For YAML deserialization



1853
1854
1855
1856
# File 'lib/net/imap/sequence_set.rb', line 1853

def init_with(coder) # :nodoc:
  @set_data = new_set_data
  self.string = coder['string']
end

#inspectObject

Returns an inspection string for the SequenceSet.

Net::IMAP::SequenceSet.new.inspect
#=> "Net::IMAP::SequenceSet()"

Net::IMAP::SequenceSet(1..5, 1024, 15, 2000).inspect
#=> 'Net::IMAP::SequenceSet("1:5,15,1024,2000")'

Frozen sets have slightly different output:

Net::IMAP::SequenceSet.empty.inspect
#=> "Net::IMAP::SequenceSet.empty"

Net::IMAP::SequenceSet[1..5, 1024, 15, 2000].inspect
#=> 'Net::IMAP::SequenceSet["1:5,15,1024,2000"]'

Large sets (by number of #entries) have abridged output, with only the first and last entries:

Net::IMAP::SequenceSet(((1..5000) % 2).to_a).inspect
#=> #<Net::IMAP::SequenceSet 2500 entries "1,3,5,7,9,11,13,15,17,19,21,23,25,27,29,31,...(2468 entries omitted)...,4969,4971,4973,4975,4977,4979,4981,4983,4985,4987,4989,4991,4993,4995,4997,4999">

Related: #to_s, #string



1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
# File 'lib/net/imap/sequence_set.rb', line 1802

def inspect
  case (count = count_entries)
  when 0
    (frozen? ? "%s.empty" : "%s()") % [self.class]
  when ..INSPECT_MAX_LEN
    (frozen? ? "%s[%p]" : "%s(%p)") % [self.class, to_s]
  else
    if @string
      head = @string[INSPECT_ABRIDGED_HEAD_RE]
      tail = @string[INSPECT_ABRIDGED_TAIL_RE]
    else
      head = export_runs(runs.first(INSPECT_TRUNCATE_LEN)) + ","
      tail = "," + export_runs(runs.last(INSPECT_TRUNCATE_LEN))
    end
    '#<%s %d entries "%s...(%d entries omitted)...%s"%s>' % [
      self.class, count,
      head, count - INSPECT_TRUNCATE_LEN * 2, tail,
      frozen? ? " (frozen)" : "",
    ]
  end
end

#intersect!(other) ⇒ Object

In-place set #intersection. Removes any elements that are missing from other from this set, keeping only the #intersection, and returns self.

other can be any object that would be accepted by ::new.

set = Net::IMAP::SequenceSet.new(1..5)
set.intersect! [2, 4, 6]
set #=> Net::IMAP::SequenceSet("2,4")

Related: #intersection, #intersect?



1670
1671
1672
1673
# File 'lib/net/imap/sequence_set.rb', line 1670

def intersect!(other)
  modifying! # short-circuit before processing input
  subtract SequenceSet.new(other).complement!
end

#intersect?(other) ⇒ Boolean Also known as: overlap?

Returns true if the set and a given object have any common elements, false otherwise.

Net::IMAP::SequenceSet["5:10"].intersect? "7,9,11" #=> true
Net::IMAP::SequenceSet["5:10"].intersect? "11:33"  #=> false

Related: #intersection, #disjoint?, #cover?, #include?

Returns:

  • (Boolean)


761
762
763
# File 'lib/net/imap/sequence_set.rb', line 761

def intersect?(other)
  valid? && import_runs(other).any? { intersect_run? _1 }
end

#limit(max:) ⇒ Object

Returns a frozen SequenceSet with * converted to max, numbers and ranges over max removed, and ranges containing max converted to end at max.

Net::IMAP::SequenceSet["5,10:22,50"].limit(max: 20).to_s
#=> "5,10:20"

* is always interpreted as the maximum value. When the set contains *, it will be set equal to the limit.

Net::IMAP::SequenceSet["*"].limit(max: 37)
#=> Net::IMAP::SequenceSet["37"]
Net::IMAP::SequenceSet["5:*"].limit(max: 37)
#=> Net::IMAP::SequenceSet["5:37"]
Net::IMAP::SequenceSet["500:*"].limit(max: 37)
#=> Net::IMAP::SequenceSet["37"]

Related: #limit!



1619
1620
1621
1622
1623
1624
1625
1626
# File 'lib/net/imap/sequence_set.rb', line 1619

def limit(max:)
  max = import_num(max)
  if    empty?                      then self.class.empty
  elsif !include_star? && max < min then self.class.empty
  elsif max(star: STAR_INT) <= max  then frozen? ? self : dup.freeze
  else                                   dup.limit!(max: max).freeze
  end
end

#limit!(max:) ⇒ Object

Removes all members over max and returns self. If * is a member, it will be converted to max.

Related: #limit



1632
1633
1634
1635
1636
1637
1638
1639
# File 'lib/net/imap/sequence_set.rb', line 1632

def limit!(max:)
  modifying! # short-circuit before querying
  star = include_star?
  max  = import_num(max)
  subtract_minmax [max + 1, STAR_INT]
  add_minmax      [max,     max     ] if star
  normalize!
end

#max(count = nil, star: :*) ⇒ Object

:call-seq:

max(star: :*) => integer or star or nil
max(count) => SequenceSet

Returns the maximum value in self, star when the set includes *, or nil when the set is empty.

When count is given, a new SequenceSet is returned, containing only the last count numbers. An empty SequenceSet is returned when self is empty. (star is ignored when count is given.)

Related: #min, #minmax, #slice



789
790
791
792
793
794
795
796
797
798
799
# File 'lib/net/imap/sequence_set.rb', line 789

def max(count = nil, star: :*)
  if count
    if cardinality <= count
      frozen? ? self : dup
    else
      slice(-count..) || remain_frozen_empty
    end
  elsif (val = max_num)
    val == STAR_INT ? star : val
  end
end

#merge(*sets) ⇒ Object

In-place set #union. Merges all of the elements that appear in any of the sets into this set, and returns self.

The sets may be any objects that would be accepted by ::new.

#string will be regenerated after all sets have been merged.

Related: #add, #add?, #union



1155
1156
1157
1158
1159
# File 'lib/net/imap/sequence_set.rb', line 1155

def merge(*sets)
  modifying! # short-circuit before import_runs
  add_runs import_runs sets
  normalize!
end

#min(count = nil, star: :*) ⇒ Object

:call-seq:

min(star: :*) => integer or star or nil
min(count) => SequenceSet

Returns the minimum value in self, star when the only value in the set is *, or nil when the set is empty.

When count is given, a new SequenceSet is returned, containing only the first count numbers. An empty SequenceSet is returned when self is empty. (star is ignored when count is given.)

Related: #max, #minmax, #slice



813
814
815
816
817
818
819
# File 'lib/net/imap/sequence_set.rb', line 813

def min(count = nil, star: :*)
  if count
    slice(0...count) || remain_frozen_empty
  elsif (val = min_num)
    val != STAR_INT ? val : star
  end
end

#minmax(star: :*) ⇒ Object

:call-seq: minmax(star: :*) => [min, max] or nil

Returns a 2-element array containing the minimum and maximum numbers in self, or nil when the set is empty. star is handled the same way as by #min and #max.

Related: #min, #max



828
# File 'lib/net/imap/sequence_set.rb', line 828

def minmax(star: :*); [min(star: star), max(star: star)] unless empty? end

#normalizeObject

Returns a SequenceSet with a normalized string representation: entries have been sorted, deduplicated, and coalesced, and all entries are in normal form. Returns self for frozen normalized sets, and a normalized duplicate otherwise.

See SequenceSet@Ordered+and+Normalized+sets.

Net::IMAP::SequenceSet["1:5,3:7,10:9,10:11"].normalize
#=> Net::IMAP::SequenceSet["1:7,9:11"]

Related: #normalize!, #normalized_string, #normalized?



1751
1752
1753
# File 'lib/net/imap/sequence_set.rb', line 1751

def normalize
  frozen? && normalized? ? self : remain_frozen(dup.normalize!)
end

#normalize!Object

Resets #string to be sorted, deduplicated, and coalesced. Returns self. See SequenceSet@Ordered+and+Normalized+sets.

Related: #normalize, #normalized_string, #normalized?



1759
1760
1761
1762
1763
# File 'lib/net/imap/sequence_set.rb', line 1759

def normalize!
  modifying! # redundant check (normalizes the error message for JRuby)
  @string = nil
  self
end

#normalized?Boolean

Returns whether #string is fully normalized: entries have been sorted, deduplicated, and coalesced, and all entries are in normal form. See SequenceSet@Ordered+and+Normalized+sets.

Net::IMAP::SequenceSet["1,3,5"].normalized?  #=> true
Net::IMAP::SequenceSet["20:30"].normalized?  #=> true

Net::IMAP::SequenceSet["3,5,1"].normalized?  #=> false, not sorted
Net::IMAP::SequenceSet["1,2,3"].normalized?  #=> false, not coalesced
Net::IMAP::SequenceSet["1:5,2"].normalized?  #=> false, repeated number

Net::IMAP::SequenceSet["1:1"].normalized?    #=> false, number as range
Net::IMAP::SequenceSet["5:1"].normalized?    #=> false, backwards range

Returns true if (and only if) #string is equal to #normalized_string:

seqset = Net::IMAP::SequenceSet["1:3,5"]
seqset.string             #=> "1:3,5"
seqset.normalized_string  #=> "1:3,5"
seqset.entries            #=> [1..3, 5]
seqset.elements           #=> [1..3, 5]
seqset.normalized?        #=> true

seqset = Net::IMAP::SequenceSet["3,1,2"]
seqset.string             #=> "3,1,2"
seqset.normalized_string  #=> "1:3"
seqset.entries            #=> [3, 1, 2]
seqset.elements           #=> [1..3]
seqset.normalized?        #=> false

Can return false even when #entries and #elements are the same:

seqset = Net::IMAP::SequenceSet["5:1"]
seqset.string             #=> "5:1"
seqset.normalized_string  #=> "1:5"
seqset.entries            #=> [1..5]
seqset.elements           #=> [1..5]
seqset.normalized?        #=> false

Note that empty sets are normalized, even though they are not #valid?:

seqset = Net::IMAP::SequenceSet.empty
seqset.normalized?        #=> true
seqset.valid?             #=> false

Related: #normalize, #normalize!, #normalized_string

Returns:

  • (Boolean)


1736
1737
1738
# File 'lib/net/imap/sequence_set.rb', line 1736

def normalized?
  @string.nil? || normal_string?(@string)
end

#normalized_stringObject

Returns a normalized sequence-set string representation, sorted and deduplicated. Adjacent or overlapping elements will be merged into a single larger range. See SequenceSet@Ordered+and+Normalized+sets.

Net::IMAP::SequenceSet["1:5,3:7,10:9,10:11"].normalized_string
#=> "1:7,9:11"

Returns nil when the set is empty.

Related: #normalize!, #normalize, #string, #to_s, #normalized?



1775
1776
1777
# File 'lib/net/imap/sequence_set.rb', line 1775

def normalized_string
  export_runs(runs) unless runs.empty?
end

#numbersObject

Returns a sorted array of all of the number values in the sequence set.

The returned numbers are sorted and de-duplicated, even when the input #string is not. See #normalize, SequenceSet@Ordered+and+Normalized+sets.

Net::IMAP::SequenceSet["2,5:9,6,12:11"].numbers
#=> [2, 5, 6, 7, 8, 9, 11, 12]

If the set contains a *, RangeError is raised. See #limit.

Net::IMAP::SequenceSet["10000:*"].numbers
#!> RangeError

WARNING: Even excluding sets with *, an enormous result can easily be created. An array with over 4 billion integers could be returned, requiring up to 32GiB of memory on a 64-bit architecture.

Net::IMAP::SequenceSet[10000..2**32-1].numbers
# ...probably freezes the process for a while...
#!> NoMemoryError (probably)

For safety, consider using #limit or #intersection to set an upper bound. Alternatively, use #each_element, #each_range, or even #each_number to avoid allocation of a result array.

Related: #elements, #ranges, #to_set



1247
# File 'lib/net/imap/sequence_set.rb', line 1247

def numbers; each_number.to_a end

#ordered_at(index) ⇒ Object

:call-seq: ordered_at(index) -> integer or nil

Returns the number at the given index in the ordered #entries, without modifying the set.

index is interpreted the same as in #at (and #[]), except that #ordered_at applies to the ordered #entries, not the sorted set.

Related: #[], #slice, #ordered_at



1497
1498
1499
# File 'lib/net/imap/sequence_set.rb', line 1497

def ordered_at(index)
  seek_number_in_minmaxes(each_entry_minmax, index)
end

#rangesObject

Returns an array of ranges

The returned elements are sorted and coalesced, even when the input #string is not. * will sort last. See #normalize, SequenceSet@Ordered+and+Normalized+sets.

* translates to an endless range. By itself, * translates to :*... Use #limit to set * to a maximum value.

Net::IMAP::SequenceSet["2,5:9,6,*,12:11"].ranges
#=> [2..2, 5..9, 11..12, :*..]
Net::IMAP::SequenceSet["123,999:*,456:789"].ranges
#=> [123..123, 456..789, 999..]

Related: #each_range, #elements, #numbers, #to_set



1219
# File 'lib/net/imap/sequence_set.rb', line 1219

def ranges; each_range.to_a end

#replace(other) ⇒ Object

Replace the contents of the set with the contents of other and returns self.

other may be another SequenceSet or any other object that would be accepted by ::new.



577
578
579
580
581
582
583
584
585
586
587
# File 'lib/net/imap/sequence_set.rb', line 577

def replace(other)
  case other
  when SequenceSet then
    modifying! # short circuit before doing any work
    @set_data = other.dup_set_data
    @string = other.instance_variable_get(:@string)
  when String      then self.string = other
  else                  clear; merge other
  end
  self
end

#send_data(imap, tag) ⇒ Object

Unstable API: for internal use only (Net::IMAP#send_data)



1842
1843
1844
# File 'lib/net/imap/sequence_set.rb', line 1842

def send_data(imap, tag) # :nodoc:
  imap.__send__(:put_string, valid_string)
end

#sizeObject

Returns the combined size of the ordered #entries, including any repeated numbers.

When #string is normalized, this returns the same as #cardinality. Like #cardinality, "*" is considered to be be distinct from 2³² - 1 (the maximum 32-bit unsigned integer value).

set = Net::IMAP::SequenceSet["4294967295:*"]
set.size                   #=> 2
set.count_with_duplicates  #=> 1
set.count                  #=> 1
set.cardinality            #=> 2

set = Net::IMAP::SequenceSet["4294967295,*"]
set.size                   #=> 2
set.count_with_duplicates  #=> 2
set.count                  #=> 1
set.cardinality            #=> 2

Related: #cardinality, #count_with_duplicates, #count, #entries



1420
1421
1422
1423
# File 'lib/net/imap/sequence_set.rb', line 1420

def size
  return cardinality unless @string
  each_entry_minmax.sum {|min, max| max - min + 1 }
end

#slice!(index, length = nil) ⇒ Object

:call-seq:

slice!(index)          -> integer or :* or nil
slice!(start, length)  -> sequence set or nil
slice!(range)          -> sequence set or nil

Deletes a number or consecutive numbers from the set, indicated by the given index, start and length, or range of offsets. Returns the number or sequence set that was removed, or nil if nothing was removed. Arguments are interpreted the same as for #slice or #[].

#string will be regenerated after deletion.

Related: #slice, #delete_at, #delete, #delete?, #subtract, #difference



1141
1142
1143
1144
1145
# File 'lib/net/imap/sequence_set.rb', line 1141

def slice!(index, length = nil)
  modifying! # short-circuit before slice
  deleted = slice(index, length) and subtract deleted
  deleted
end

#stringObject

Returns the IMAP sequence-set string representation, or nil when the set is empty. Note that an empty set is invalid in the IMAP syntax.

Use #valid_string to raise an exception when the set is empty, or #to_s to return an empty string.

If the set was created from a single string, it is not normalized. If the set is updated the string will be normalized.

Related: #valid_string, #normalized_string, #to_s, #inspect



612
# File 'lib/net/imap/sequence_set.rb', line 612

def string; @string || normalized_string if valid? end

#string=(input) ⇒ Object

Assigns a new string to #string and resets #elements to match. Assigning nil or an empty string are equivalent to calling #clear.

Non-empty strings are validated but not normalized.

Use #add, #merge, or #append to add a string to an existing set.

Related: #replace, #clear



626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
# File 'lib/net/imap/sequence_set.rb', line 626

def string=(input)
  if input.nil?
    clear
  elsif (str = String.try_convert(input))
    modifying! # short-circuit before parsing the string
    entries = each_parsed_entry(str).to_a
    clear
    if normalized_entries?(entries)
      replace_minmaxes entries.map!(&:minmax)
    else
      add_minmaxes entries.map!(&:minmax)
      @string = -str
    end
  else
    raise ArgumentError, "expected a string or nil, got #{input.class}"
  end
  input
end

#subtract(*sets) ⇒ Object

In-place set #difference. Removes all of the elements that appear in any of the given sets from this set, and returns self.

The sets may be any objects that would be accepted by ::new.

Related: #difference



1167
1168
1169
1170
1171
# File 'lib/net/imap/sequence_set.rb', line 1167

def subtract(*sets)
  modifying! # short-circuit before import_runs
  subtract_runs import_runs sets
  normalize!
end

#to_sObject

Returns the IMAP sequence-set string representation, or an empty string when the set is empty. Note that an empty set is invalid in the IMAP syntax.

Related: #string, #valid_string, #normalized_string, #inspect



650
# File 'lib/net/imap/sequence_set.rb', line 650

def to_s; string || "" end

#to_setObject

Returns a Set with all of the #numbers in the sequence set.

If the set contains a *, RangeError will be raised.

See #numbers for the warning about very large sets.

Related: #elements, #ranges, #numbers



1326
# File 'lib/net/imap/sequence_set.rb', line 1326

def to_set; Set.new(numbers) end

#valid?Boolean

Returns false when the set is empty.

Returns:

  • (Boolean)


831
# File 'lib/net/imap/sequence_set.rb', line 831

def valid?; !empty? end

#valid_stringObject

Returns the IMAP sequence-set string representation, or raises a DataFormatError when the set is empty.

Use #string to return nil or #to_s to return an empty string without error.

Related: #string, #normalized_string, #to_s

Raises:



596
597
598
599
# File 'lib/net/imap/sequence_set.rb', line 596

def valid_string
  raise DataFormatError, "empty sequence-set" if empty?
  string
end

#validateObject

Unstable API: currently for internal use only (Net::IMAP#validate_data)



1836
1837
1838
1839
# File 'lib/net/imap/sequence_set.rb', line 1836

def validate # :nodoc:
  empty? and raise DataFormatError, "empty sequence-set is invalid"
  self
end

#xor!(other) ⇒ Object

In-place set #xor. Adds any numbers in other that are missing from this set, removes any numbers in other that are already in this set, and returns self.

other can be any object that would be accepted by ::new.

set = Net::IMAP::SequenceSet.new(1..5)
set.xor! [2, 4, 6]
set #=> Net::IMAP::SequenceSet["1,3,5:6"]

Related: #xor, #merge, #subtract



1686
1687
1688
1689
1690
1691
# File 'lib/net/imap/sequence_set.rb', line 1686

def xor!(other)
  modifying! # short-circuit before processing input
  other = SequenceSet.new(other)
  copy  = dup
  merge(other).subtract(other.subtract(copy.complement!))
end

#|(other) ⇒ Object Also known as: +, union

:call-seq:

self + other -> sequence set
self | other -> sequence set
union(other) -> sequence set

Returns a new sequence set that has every number in the other object added.

other may be any object that would be accepted by ::new.

Net::IMAP::SequenceSet["1:5"] | 2 | [4..6, 99]
#=> Net::IMAP::SequenceSet["1:6,99"]

Related: #add, #merge, #&, #-, #^, #~

Set identities

lhs | rhs is equivalent to:

  • rhs | lhs (commutative)

  • ~(~lhs & ~rhs) (De Morgan’s Law)

  • (lhs & rhs) ^ (lhs ^ rhs)



860
# File 'lib/net/imap/sequence_set.rb', line 860

def |(other) remain_frozen dup.merge other end

#~Object Also known as: complement

:call-seq:

~ self     -> sequence set
complement -> sequence set

Returns the complement of self, a SequenceSet which contains all numbers except for those in this set.

~Net::IMAP::SequenceSet.full  #=> Net::IMAP::SequenceSet.empty
~Net::IMAP::SequenceSet.empty #=> Net::IMAP::SequenceSet.full
~Net::IMAP::SequenceSet["1:5,100:222"]
#=> Net::IMAP::SequenceSet["6:99,223:*"]
~Net::IMAP::SequenceSet["6:99,223:*"]
#=> Net::IMAP::SequenceSet["1:5,100:222"]

Related: #complement!, #|, #&, #-, #^

Set identities

~set is equivalent to:

  • full - set, where “full” is Net::IMAP::SequenceSet.full



961
# File 'lib/net/imap/sequence_set.rb', line 961

def ~; remain_frozen dup.complement! end