Implement several rules that determine when an identifier on a new
line is a continuation of a selector-style call on a previous line:
- In certain contexts, such as parentheses or square brackets, it's
always a continuation because one does not split statements in
those contexts;
- Otherwise, compare the leading whitespace on the line containing
the nearest enclosing statement or declaration to the leading
whitespace for the line containing the identifier.
The leading whitespace for a line is currently defined as all space
and tab characters from the start of the line up to the first
non-space, non-tab character. Leading whitespace is compared via a
string comparison, which eliminates any dependency on the width of a
tab. One can run into a few amusing cases where adjacent lines that
look indented (under some specific tab width) aren't actually indented
according to this rule because there are different mixes of tabs and
spaces in the two lines. See the bottom of call-suffix-indent.swift
for an example.
I had to adjust two test cases that had lines with slightly different
indentation. The diagnostics here are awful; I've made no attempt at
improving them.
Swift SVN r13843
This can be used as, for example
self.init withRed(0.5) green(0.5) blue(0.5) alpha(1.0)
which mimics the declaration. THe same goes for super.
Note that this does not yet support
NSColor.init withRed(0.5) green(0.5) blue(0.5) alpha(1.0)
due to a limitation in the type checker.
Swift SVN r13825
This allows code such as
obj.closure { return 0 } onError { println("error") }
to parse appropriately. The only other functionality change here is
that we no longer allow the use of a trailing closure within the
condition of a C-style for loop, because it did awful things to
recovery. I doubt we'll miss it.
Swift SVN r13823
We don't want to parse the expr-call-suffix as a general
postfix-expression; rather, we only want it when we're directly naming
a method. This doesn't matter so much in the current model, but it
becomes important when we move over to the newer keyword argument model.
Swift SVN r13819
With enough of the world's complexity pushed around, this is now straight-forward,
but we still need to use full parser lookahead to handle some more complex cases.
Swift SVN r13727
clients to either go through the new parseExpr (which is never "basic")
or the existing parseExprBasic entrypoint if they don't want trailing
closures.
Swift SVN r13724
from the pattern to the scope (it doesn't do other argument
specific stuff like mucking with decl contexts) rename it to
addPatternVariablesToScope, and use it in two more places
in the parser.
Swift SVN r13710
automatically reparent VarDecls in their arg/body patterns and
GenericParameters to themselves. These all have to be created
before the actual context decl is created and then reparented,
so we might as well have the reparenting be done by the decl
itself. This lets us take out some setDeclContext reparenting
loops from around the parser.
I'm sure that there are a lot more places they can be removed
from as well.
NFC.
Swift SVN r13701
function. Parse inout as a contextual keyword there, shoving it into the
TypedPattern (instead of introducing a new kind of Pattern). This enables
us to parse, sema, and irgen the new '@-less' syntax for inout.
Swift SVN r13559
Introduce a new Pattern::forEachVariable that takes a lambda and iterates
over all the variables encompassed by the pattern (the ones found by
collectVariables) and runs the lambda on it.
Use this to implement collectVariables, so we only have one copy of the code.
Convert several places to use this, removing a bunch of duplicated code for
walking the pattern structure and extracting decls.
Convert a few uses of collectVariables to forEachVariable, leading to more
clear code without a temporary smallvector.
Remove simplifyPatternTypes in CSApply.cpp, which is dead.
Swift SVN r13545
To get here, make the implicit 'self' parameter a bit more like a
parsed parameter, by giving it a trivial TypeRepr so that pattern
validation will validate the type. This latter piece is important for
DynamicSelf, because we need the interface type of the function to
refer to the implicit generic parameter while the actual type refers
to the archetype.
Swift SVN r12757
Thanks to the way we've set up our diagnostics engine, there's not actually
a reason for /everything/ to get rebuilt when /one/ diagnostic changes.
I've split them up into five categories for now: Parse, Sema, SIL, IRGen,
and Frontend, plus a set of "Common" diagnostics that are used in multiple
areas of the compiler. We can massage this later.
No functionality change, but should speed up compile times!
Swift SVN r12438
As with chaining initialization ('super.init' calls), wrapping the
delegating initialization in RebindSelfInConstructorExpr ensures that
'self' gets overwritten by the result of delegation.
Note that I'd much prefer that RebindSelfInConstructorExpr be
introduced by the type checker (not the parser). That cleanup will
follow.
Swift SVN r11932
This allows them to appear in argument lists of functions, enabling behavior
like this:
func test_arguments(a : Int, var b : Int, let c : Int) {
a = 1 // ok (for now).
b = 2 // ok.
c = 3 // expected-error {{cannot assign to the result of this expression}}
}
Swift SVN r11746
- mark closure arguments (both explicit and $0's) as immutable
- Adjust the stdlib (one place) and some tests to cope with this.
- Remove some special case logic in sema for lvalue qualifying
anonymous closure exprs, which is now the wrong thing to do.
Swift SVN r11674
...rather than a raw pointer that points to a buffer with space for N
elements. Just because we *can* get N from context doesn't mean it's
convenient/safe.
No functionality change.
Swift SVN r11488
arguments. Until we have @inout methods on structs, we can't mark anything
that can be address-exposed. This temporarily regresses a few testcases,
but they will come back later.
Swift SVN r11359
loops as 'let', silgen isn't ready for it yet.
Change silgen's handling of let variables to stop using emitLoweredCopyValue,
which is busted for aggregates that contain both trivial and nontrivial types
(15669586).
Swift SVN r11352
var decls. I was originally intending to use this for argument lists, but I
don't think that's a great direction to go anymore.
In any case, it seems uncontroversial to enforce immutability on foreach
enumation patterns, so I did that (see testcase)
Swift SVN r11124
are not settable (like get-only ones). Set the 'isLet' bit in various
places, but not the particularly interesting or useful places yet.
Swift SVN r11121
We used to produce:
(swift) { () in $0 }
<REPL Input>:1:9: error: anonymous closure argument not contained in a closure
rdar://15003789
Swift SVN r10751
This allows expressions such as ".foo" and ".foo(1)" to refer to
static variables and static methods, respectively, as well as enum
cases.
To get here, rework the parsing of delayed identifier expressions a
bit, so that the argument itself is part of the delayed argument
expression rather than a separate call expression. This simplifies
both the handling of patterns of this form and the type checker, which
can now user simpler constraints.
If we really want to support (.foo)(1), we can make that work, but it
seems unnecessary and perhaps confusing.
Swift SVN r10626
Previously, the Parser and BranchStmt typedef-ed ExprStmtOrDecl as a pointer union. Using typedef made the objects compatible, but did not allow us to extend the type with helper methods, such as getSourceRange(), which is something you can get on all of the AST objects. This patch introduces ASTNode that subclasses from PointerUnion and is used by both parser and BranchStmt.
Swift SVN r9971
wide
Currently integer literals are 64-bit. In order to allow checking for overflow
while converting an integer literal to swift.UInt/Int* types we need at least
65 bits. But floating point numbers (Float32, Float64, Float80) are
BuiltinIntegerLiteralConvertible. In order to allow spelling large floating
point constants, we allow 136-bit literals.
Rationale: 128 bits are enough to represent the absolute value of min/max IEEE
Binary32, and we need 1 bit to represent the sign. 136 is 129 rounded to the
next 8 bits.
The plan is to have builtins that do the overflow check and convert 136-bit
numbers to the required width. We need these builtins for both integers and
floating point numbers to ensure that 136-bit numbers are folded into sane
constants in SIL and don’t escape to LLVM IR.
Swift SVN r9253
