Commit Graph

21504 Commits

Author SHA1 Message Date
Doug Gregor
81b45019e1 When building an archetype within a protocol, capture the list of
protocols it conforms to. Use this list when substituting into a
protocol's signature, so that we get an existential type back. For
example:

  var e : Enumerable
  var r = e.getElements()
  r.getFirstAndAdvance() // okay: r is a range.





Swift SVN r2139
2012-06-04 19:36:25 +00:00
Eli Friedman
27f8a5ab62 Teach the parser's handling of scopes to handle local types correctly.
Swift SVN r2138
2012-06-04 19:14:58 +00:00
Doug Gregor
9abd2340bc Allow an 'inheritance' clause on typealiases, specifying what
protocols the underlying type of the type alias shall conform to. This
isn't super-motivating by itself (one could extend the underying type
instead), but serves as documentation, makes typealiases provide the
same syntax as other nominal types in this regard, and will also be
used to specify requirements on associated types.


Swift SVN r2133
2012-06-04 16:15:47 +00:00
Doug Gregor
d41e39a363 Sink the 'inherited' members for the various nominal types down into
TypeDecl; TypeAliasDecl will get this ability soon.


Swift SVN r2132
2012-06-04 15:56:29 +00:00
Eli Friedman
26159b59ee First draft of some code to make the general unqualified lookup able to perform lookup into local scopes. Normally, name-binding doesn't need this because the parser can resolve references to locals, but we need this sort of lookup to handle cases involving local types. It's also likely this will be generally useful for tools consuming the AST, typo-correction, etc.
We probably need to add some sort of data structure to represent this information, but as a proof of concept the current code appears to work.  I'm still working out how to make sure the parser doesn't prematurely bind names and how to make name-binding use it where appropriate (and avoid it when we don't need it, because no matter how efficient we make it, it will still be relatively expensive).



Swift SVN r2112
2012-06-02 01:26:58 +00:00
Eli Friedman
d829b3c775 Unify the complete unqualified lookup logic in the UnqualifiedLookup class.
Swift SVN r2109
2012-06-01 23:26:58 +00:00
Eli Friedman
0e7ef8cebb Start refactoring code towards a single complete implementation of unqualified lookup.
Swift SVN r2108
2012-06-01 01:37:44 +00:00
Eli Friedman
ae86d64644 Rename Decl::getLocStart() to Decl::getStartLoc(). Add Decl::getLoc(), which is essentially the location which should be used for diagnostics.
Swift SVN r2105
2012-05-31 23:56:30 +00:00
Doug Gregor
1dc9b1597a Unbreak CMake build
Swift SVN r2104
2012-05-31 23:39:07 +00:00
Eli Friedman
29a13e9ae0 Move the NameLookup class into the AST library.
Swift SVN r2101
2012-05-31 22:35:50 +00:00
Eli Friedman
ba4a76038b Make oneofs never implicitly generate an ExtensionDecl. This matters for local oneofs.
Swift SVN r2098
2012-05-31 21:20:56 +00:00
Eli Friedman
c404598fcb Parsing and semantic analysis for 'break' and 'continue'.
Swift SVN r2087
2012-05-31 00:55:33 +00:00
Doug Gregor
3c2fb97bdf Introduce TypeBase::isExistentialType(), to determine whether a given
type is either a protocol type or a protocol composition type. The
long form of this query returns the minimal set of protocol
declarations required by that existential type.

Use the new isExistentialType() everywhere that we previously checked
just for ProtocolType, implementing the appropriate rules. Among other
things, this includes:
  - Type coercion
  - Subtyping relationship
  - Checking of explicit protocol conformance
  - Member name lookup

Note the FIXME for IR generation; we need to decide how we want to
encode the witnesses for the different protocols.

This is most of <rdar://problem/11548207>.


Swift SVN r2086
2012-05-31 00:26:13 +00:00
Eli Friedman
0c24d526af Add rules to verifier for CoerceExpr.
Swift SVN r2072
2012-05-30 18:23:14 +00:00
Doug Gregor
c079874625 Implement parsing, AST, type canonicalization, and type validation for
protocol conformance types, e.g., 'protocol<P, Q>'. A few things
people *might* want to scream about, or at least scrutinize:

  - The parsing of the '<' and '>' is odd, because '<' and '>' aren't
    tokens, but are part of the operator grammar. Neither are '>>',
    '>>>', '<>', etc., which also come up and need to be parsed
    here. Rather than turning anything starting with '<' or '>' into a
    different kind of token, I instead parse the initial '<' or '>'
    from an operator token and leave the rest of the token as the
    remaining operator.
  - The canonical form of a protocol-composition type is minimized by
    removing any protocols in the list that were inherited by other
    protocols in the list, then sorting it. If a singleton list is
    left, then the canonical type is simply that protocol type.
  - It's a little unfortunate that we now have two existential types
    in the system (ProtocolType and ProtocolCompositionType), because
    many places will have to check both. Once ProtocolCompositionTypes
    are working, we should consider whether it makes sense to remove
    ProtocolType.

Still to come: name lookup, coercions.



Swift SVN r2066
2012-05-30 00:39:08 +00:00
Eli Friedman
a454d62619 Implement some reasonable shadowing rules for declarations.
Swift SVN r2061
2012-05-29 23:48:40 +00:00
Chris Lattner
e65472637d update for mainline api change, much nicer interface now.
Swift SVN r2020
2012-05-27 18:29:53 +00:00
Eli Friedman
155ed201c1 Name-binding doesn't touch expressions anymore, so UNBOUND_EXPR isn't necessary.
Swift SVN r1986
2012-05-25 02:28:08 +00:00
Eli Friedman
6fe5759841 Get rid of a hack where ASTWalker was trying to walk the AST for BinaryExprs in a non-standard order. It introduces complexity with no obvious benefit.
Swift SVN r1984
2012-05-25 01:39:00 +00:00
Doug Gregor
0062c84351 Introduce ExistentialMemberRefExpr for references to the members of
values of existential type, e.g.,

  var x : Printable
  x.print()

Existential member references reify the type of the implicit object
argument (implicitly, because we have no way of expressing this in the
type system), and replace the types of any other archetypes
with existential types that (don't, but will eventually) conform to
the protocols to which the archetypes conform.



Swift SVN r1963
2012-05-23 23:06:34 +00:00
Doug Gregor
9096497f0f Rename "dependent type" to "unresolved type" universally. We've been
using the term "unresolved" in expressions for a while, and it fits
for types better than "dependent type."

The term "dependent type" will likely come back at some point to mean
"involves an archetype".



Swift SVN r1962
2012-05-23 19:03:14 +00:00
Doug Gregor
3fdbcff922 Teach TypeBse::getUnlabeledType() not to remove sugar when it doesn't need to
Swift SVN r1954
2012-05-23 15:27:32 +00:00
Doug Gregor
74436a3120 Introduce the implicit 'This' type into protocols, which refers to the
type T that conforms to the given protocol. 'This' is modeled simply
as an associated type.



Swift SVN r1953
2012-05-23 14:45:55 +00:00
Doug Gregor
130a72c611 Fix typo of GET_INTRINSIC_GENERATOR_GLOBAL
Swift SVN r1952
2012-05-23 14:32:58 +00:00
Chris Lattner
d689beecc1 avoid a warning with llvm r157312
Swift SVN r1951
2012-05-23 05:20:45 +00:00
Eli Friedman
8d985dcc38 Change extension lookup so that decls in the current module don't always shadow imports, which should be enough to unblock demo work. <rdar://problem/11492474>.
Swift SVN r1950
2012-05-23 03:16:20 +00:00
Doug Gregor
8e1cd3990b Implement substitution of types, which substitutes concrete types for
archetypes. Use this substitution when checking the
variable/function/subscript witnesses during protocol conformance.

This allows us to check the conforms-to relationship for the Range
protocol as we want to express it.


Swift SVN r1945
2012-05-23 00:39:06 +00:00
Doug Gregor
d37602629e Implement parsing, AST, and conformance checking for associated types
in protocols, e.g.,

  protocol Range {
    typealias Element
    func getAndAdvance() -> Element
  }



Swift SVN r1941
2012-05-22 21:45:58 +00:00
Eli Friedman
03a0704ca0 Add a really ugly hack to get shadowing for types working. I'm working on the right solution, but I'm not sure how long that will take.
Swift SVN r1940
2012-05-22 21:16:11 +00:00
Eli Friedman
35d7084c40 Add framework for handling decls in the verifier.
Swift SVN r1926
2012-05-22 00:42:31 +00:00
Eli Friedman
3ff2ef1efb Allow the verifier to perform partial checking if there's an error in type-checking.
Swift SVN r1925
2012-05-22 00:34:12 +00:00
Doug Gregor
c68c9da1b6 Introduce ThisApplyExpr, an ApplyExpr meant for the various forms of
application expressions that provide a 'this' parameter to a
method. There will eventually be more than just DotSyntaxCallExpr, to
capture the other syntactic (and implicit) forms.


Swift SVN r1924
2012-05-22 00:25:40 +00:00
Doug Gregor
b1f1e1430d Introduce NominalType as a base class of
OneOfType/StructType/ClassType/ProtocolType, an eliminate the
ridiculous chained ifs that used to check for most (or all) of these.


Swift SVN r1923
2012-05-22 00:13:59 +00:00
Doug Gregor
a9906ad38f Introduce support for implicit user-defined conversions.
A user-defined conversion function is an instance method that accepts
an empty tuple and returns a value of the type we're converting to,
has the [conversion] attribute, and is named __conversion. The last of
these restrictions is a temporary hack to work around our inability to
perform a lookup across all extensions for "every function with the
conversion attribute", and shouldn't last too long.

As in C++, we only get one user-defined conversion function. Unlike in
C++, a constructor is not (and cannot) be a conversion function.

Introduce NSString <-> String conversion functions, but leave the
runtime implementations as stubs for Dave to fill in.



Swift SVN r1921
2012-05-21 23:36:16 +00:00
John McCall
38bac7c706 Add Builtin.ObjCPointer with accompanying IR-gen support and
wrap it in an 'id' type in the standard library.

Also fix a bug noticed by inspection where initWithTake for
function types wasn't entering a cleanup for the taken value.
This probably doesn't matter for existing possibilities, but
it's potentially important under exceptions.

Swift SVN r1902
2012-05-18 23:40:17 +00:00
John McCall
71c45f5212 Lion's libc++ headers do not support shared_ptr under -fno-rtti.
This is <rdar://problem/10217868>.  Apparently I'm using Lion's
libc++ headers somehow, which I should probably fix;  but since
the use of shared_ptr is just a hack until DenseMap supports
move-only types, I don't feel bad about changing it to a different
hack that avoids shared_map altogether.

Swift SVN r1897
2012-05-18 10:01:54 +00:00
Chris Lattner
b2fb91fba6 rip out my previous hack for Builtin.trap() and put in general code that
allows access to any LLVM IR intrinsic that has types that can be mapped
to swift types.  Notably, this excludes vector stuff, but there is a lot
of other goodness that can now be poked at.



Swift SVN r1890
2012-05-17 20:31:17 +00:00
Eli Friedman
e19c64a2f5 A couple minor fixes to TypeBase::getUnlabeledType for varargs tuples.
Swift SVN r1889
2012-05-17 19:51:37 +00:00
Doug Gregor
cded58955e Make %select{} formatting recursive.
Swift SVN r1884
2012-05-17 15:42:10 +00:00
Eli Friedman
5e206bc10f Rename DistinctTypeDecl to NominalTypeDecl.
Swift SVN r1882
2012-05-17 01:54:15 +00:00
Eli Friedman
6e69e27975 Kill ScalarToTupleExpr, which is no longer in use.
Swift SVN r1881
2012-05-17 01:50:48 +00:00
Eli Friedman
d4750d9014 Finish off the semantic analysis of vararg tuples.
Swift SVN r1880
2012-05-17 01:44:27 +00:00
Eli Friedman
099a97847d Parsing and AST support for varargs tuples.
Swift SVN r1875
2012-05-16 22:02:01 +00:00
Eli Friedman
b788bad8e6 A couple missing pieces for classes. Also, a few fixes for local types.
Swift SVN r1873
2012-05-16 19:21:07 +00:00
Eli Friedman
33f20a14e0 Add NewReferenceExpr, for allocating class objects.
Swift SVN r1867
2012-05-16 01:36:03 +00:00
Eli Friedman
98933b8510 More work on ClassDecls; this is enough to get member access working.
Swift SVN r1863
2012-05-16 00:22:09 +00:00
Doug Gregor
245d2a46f2 Parse an 'inheritance' clause on extensions, oneofs, structs, and
classes, with the same syntax as we have on protocols. This
inheritance specifies explicit conformance to a protocol.

Later, we can allow a class definition to have a single class type
within this list, when we introduce class inheritance.



Swift SVN r1862
2012-05-15 23:39:19 +00:00
Eli Friedman
a5a39860cd Basic parsing plus a bit more of the AST for ClassDecls.
Swift SVN r1860
2012-05-15 22:07:31 +00:00
Eli Friedman
79c3276b07 AST representation for ClassDecls.
Swift SVN r1858
2012-05-15 21:53:38 +00:00
Chris Lattner
5163b40020 Hack in support for Builtin.trap(), resolving rdar://11442881. a more
general solution to get access to arbitrary intrinsics will follow.


Swift SVN r1857
2012-05-15 17:40:46 +00:00