When the Index type of a Range is Comparable, we can reject invalid
ranges at their formation.
It might be worth injecting some checkability concept into ForwardIndex
so that even in generic contexts where Comparability is unknown, we can
do this same check in the Range constructor.
Also, remove pattern matching support for Ranges of RandomAccessIndex,
as that is covered by Interval.
Fixes <rdar://problem/16254937> (#Seed 4: Ranges with negative strides
are broken) to the extent possible (it's still possible to form an
invalid Range of indices that are not Comparable)
Fixes <rdar://problem/17164391> (Swift: Using ranges in for in for
counting down, causes the loop to go inifinely)
Fixes <rdar://problem/17580871> (Swift: Closed range with negative
endpoint excludes that endpoint)
Swift SVN r19903
This allows 0.0..<10.3 and 0.1...4.4 to work properly in pattern
matching.
Fixes <rdar://problem/12016900> (#Seed 4: add a "contains" method to the
range types) by making "contains" available on Interval
Addresses <rdar://problem/16254937> (#Seed 4: Ranges with negative
strides are broken) by making the formation of an invalid Interval a
runtime error.
Fixes <rdar://problem/16304317> (Range<T> has limited awesomeness: it is
restricted to types that conform to ForwardIndex)
Fixes <rdar://problem/16736924> (#Seed 4: Enable range inclusion pattern
matching for all types that conform to Equatable and Comparable)
Addresses <rdar://problem/16846325> (#Seed 4: Introduce index range) by
distinguishing Range (which operates on indices) from Interval
Fixes <rdar://problem/17051263> (Pattern-matching a Double range with an
infinite endpoint triggers an assertion failure)
Fixes <rdar://problem/17051271> (#Seed 4: Pattern-matching Double ranges
excludes fractional values)
Addresses <rdar://problem/17171420> (Separate types for closed and
half-open ranges)
Swift SVN r19900
Mechanically add "Type" to the end of any protocol names that don't end
in "Type," "ible," or "able." Also, drop "Type" from the end of any
associated type names, except for those of the *LiteralConvertible
protocols.
There are obvious improvements to make in some of these names, which can
be handled with separate commits.
Fixes <rdar://problem/17165920> Protocols `Integer` etc should get
uglier names.
Swift SVN r19883
Specialization now recurses on only that prefix of the
current matrix which shares a specialization form
(essentially, a pattern kind) with the head row. This is
inferior to the previous algorithm in a number of ways: we
may require more switches to perform a single dispatch, and
we may introduce more redundant variables in the leaves.
However, it also means that we will have fully specialized a
row along exactly one path in the decision tree, which makes
it much easier to work with dispatches that introduce new
cleanups.
This change also changes switch-dispatch to use the new
dynamic-cast instructions.
Incidentally fixes rdar://16401831.
Swift SVN r19336
When we specialize the decision matrix on a pattern, we have to start with the *first* preceding wildcard, not the last. Oops. While we're here, clean things up a bit by using Optional instead of magic integer values to track the first wildcard row. Fixes <rdar://problem/17272985>.
Swift SVN r19061
