Commit Graph

614 Commits

Author SHA1 Message Date
Joe Groff
a89949183d Give opened existential archetypes globally unique identifiers.
This lets us reliably print and parse opened archetypes across different compiler invocations. Using a source-related locator would be ideal, but that's complicated by the need to manufacture, print, and parse these things during SIL passes, so cop out and burn a UUID for now.

Swift SVN r22385
2014-09-30 14:07:31 +00:00
Devin Coughlin
5a9ccc5ab2 Add ASTDumper support for AvailabilityQueryExpr
This patch also moves some static utility methods involving PlatformKind out of Attr.h and into PlatformKind.h.


Swift SVN r21896
2014-09-12 00:13:48 +00:00
Devin Coughlin
1b8ed882d3 Move AvailabilityAttr::PlatformKind into its own file
Swift SVN r21728
2014-09-04 23:34:19 +00:00
Erik Eckstein
99cc7603be Add an @inline(__always) function attribute.
This will let the performance inliner inline a function even if the costs are too high.
This attribute is only a hint to the inliner.
If the inliner has other good reasons not to inline a function,
it will ignore this attribute. For example if it is a recursive function (which is
currently not supported by the inliner).

Note that setting the inline threshold to 0 does disable performance inlining at all and in
this case also the @inline(__always) has no effect.



Swift SVN r21452
2014-08-26 00:56:34 +00:00
Doug Gregor
17716524c5 Handle CF <-> Objective-C toll-free-bridged conversions in the type checker.
Introduce an attribute that describes when a given CF type is
toll-free-bridged to an Objective-C class, and which class that
is. Use that information in the type checker to provide the CF <->
Objective-C toll-free-bridged conversions directly, rather than using
the user-defined conversion machinery.

Swift SVN r21376
2014-08-21 21:36:05 +00:00
Arnold Schwaighofer
cd799f8e46 Revert "Add an inline(late) attribute"
This reverts commit r21286.

Discussions ongoing.

Swift SVN r21289
2014-08-19 18:15:25 +00:00
Arnold Schwaighofer
7aa62ce835 Add an inline(late) attribute
This disables inlining at the SIL level. LLVM inlining is still enabled. We can
use this to expose one function at the SIL level - which can participate in
dominance based optimizations but which is implemented in terms of a cheap check
and an expensive check (function call) that benefits from LLVM's inlining.

Example:

The inline(late) in the example below prevents inlining of the two checks. We
can now perform dominance based optimizations on isClassOrObjExistential.
Without blocking inlining the optimizations would apply to the sizeof check
only and we would have multiple expensive function calls.

@inline(late)
func isClassOrObjExistential(t: Type) -> Bool{
  return sizeof(t) == sizeof(AnyObject) &&
    swift_isClassOrObjExistential(t)
}

We do want inlining of this function to happen at the LLVM level because the
first check is constant folded away - IRGen replaces sizeof by constants.

rdar://17961249

Swift SVN r21286
2014-08-19 18:05:44 +00:00
Nadav Rotem
23ab9d52bd Add a comment to the enum definition. NFC.
Swift SVN r21092
2014-08-07 16:31:13 +00:00
Nadav Rotem
2e18d8ba7f Reorder the fields in the enum to allow less-than comparison on the values.
Swift SVN r21085
2014-08-07 07:18:31 +00:00
Nadav Rotem
478cf8d5c5 Document the enum. NFC.
Swift SVN r21083
2014-08-07 06:21:27 +00:00
Nadav Rotem
8dd52d63cd Add the @effects(readonly/readnone/readwrite) attribute.
Swift SVN r21070
2014-08-06 18:48:20 +00:00
Doug Gregor
63304b57b4 Introduce AvailabilityAttr::prettyPlatformName() to improve diagnostics.
Swift SVN r20958
2014-08-02 21:06:30 +00:00
Doug Gregor
f52c6789db Track the minimum deployment target and use it for 'unavailable' computations.
Swift SVN r20955
2014-08-02 18:05:45 +00:00
Jordan Rose
9f64db085a Fill out the implementation of AvailabilityAttr.
This includes proper printing support as well as proper platform checking
when seeing if a decl is unavailable. A few other places in the code will
now use AvailabilityAttr::isUnavailable instead of just checking the
is-unavailable-always flag (and not always checking the platform).

No new tests yet because this doesn't include /parsing/ the other fields
of AvailabilityAttr. That will come next, at which point we'll test each
of the cases that has been switched over to use
AvailabilityAttr::isUnavailable.

Part of <rdar://problem/17024498>

Swift SVN r20844
2014-07-31 18:58:17 +00:00
Jordan Rose
5257550f13 [ClangImporter] Don't mark generated 'unavailable' attributes as implicit.
Nearly all of them come from some annotation written explicitly in the
Objective-C header, and all of them should be shown in the generated
interface for an imported module.

Part of <rdar://problem/17024498>

Swift SVN r20841
2014-07-31 18:58:14 +00:00
Jordan Rose
1a65ce0c16 Print @availability and @asmname on their own line.
These attributes are too long and would mess up the flow of the line if
placed directly before the decl introducer.

Swift SVN r20840
2014-07-31 18:58:13 +00:00
Jordan Rose
e5fc17843e Allow overrides to leave out the ownership qualifier of the decl they override.
If they do provide it, however, it must match. So the following combinations
are valid:

  weak var -> override weak var
  weak var -> override var

but these are not:

  var -> override weak var
  weak var -> override unowned var

To be fully correct, we should be cloning the attribute down and adjusting
the type of the overriding property, but I'd rather not touch that right now.

<rdar://problem/17837100>

Swift SVN r20694
2014-07-29 21:03:51 +00:00
Joe Groff
f2ecef61bd Parse @sil_unmanaged correctly.
Remove it after we validate ownership in a SIL TypeRepr, so we don't incorrectly complain about it not applying to types.

Swift SVN r20580
2014-07-25 23:59:36 +00:00
Chris Lattner
6d0934d33d Introduce a new 'renamed' modifier on unavailable attributes. This allows Xcode
to emit fixit's when we rename something, e.g.:

t.swift:6:9: error: 'float' has been renamed to Float
var y : float
        ^~~~~
        Float

Adopt this in the stdlib.



Swift SVN r20549
2014-07-25 18:00:37 +00:00
Jordan Rose
a2af37a19b Split PrintOptions::printEverything into printEverything and printVerbose.
The former is for debugging, the latter is for detailed presentation to users.
swift -print-ast will continue using printEverything, as will swift-ide-test,
but all other features should use printVerbose.

Swift SVN r20432
2014-07-23 22:28:55 +00:00
Joe Groff
7feeecfd21 Sema: Bring back 'assignment' as an infix operator modifier.
We need it to decide whether to admit infix operators into an optional chain, such as 'x? = 0' or 'x?.y += 2'.

Swift SVN r20295
2014-07-22 05:16:01 +00:00
Chris Lattner
bc481f0fe1 implement <rdar://problem/16859927> remove the underscore in "auto_closure"
autoclosure is one work, not two.



Swift SVN r20253
2014-07-21 15:23:50 +00:00
Chris Lattner
62ed200368 add a flag that allows us to reject SIL-only decl attributes, and use it to reject @sil_stored in
.swift files.



Swift SVN r20233
2014-07-20 21:03:25 +00:00
Chris Lattner
6fb46992f2 move the utility functions that map a string to a type/decl attribute ID out of ParseDecl into Attr.cpp,
allowing us to give them a better name and dox comment them.  This elimiantes a duplicate copy from the IDE layer as well, ensuring consistency.


Swift SVN r20075
2014-07-17 05:28:34 +00:00
Chris Lattner
a9ed3303a6 rip out "AtLoc" in DeclAttributes, now that each individual attribute tracks the location of its associated @ (if present)
Swift SVN r20072
2014-07-17 05:02:01 +00:00
Chris Lattner
fdaa28ce16 fix a warning about a zero length array, by ripping out all of the old declattributes code. :-)
Swift SVN r20071
2014-07-17 04:48:34 +00:00
Jordan Rose
31c1c20a89 Print attributes in the order they were written.
Also, don't print accessibility attributes twice. (Chris's general
modifier-handling code subsumes my accessibility-specific special cases.)

Swift SVN r20043
2014-07-16 21:39:10 +00:00
Chris Lattner
6592366f23 switch weak/unowned/unowned(unsafe) to the new declattributes machinery.
Swift SVN r20032
2014-07-16 19:46:00 +00:00
Chris Lattner
2825b9f5f8 move @transparent and @requires_stored_property_inits to the new attribute model.
All that is left is unowned(unsafe) & weak.



Swift SVN r20012
2014-07-16 06:14:19 +00:00
Chris Lattner
1907ad9c91 Move 'mutating' and 'nonmutating' onto the new attribute processing machinery.
Also:
 - merge the test/decl/func/functions_new.swift testcase into test/Sema/immutability.swift,
   where the bulk of similar tests are.
 - Move the type checking logic for 'dynamic' out of ValidateAttrs into TypeCheckAttrs
 - Change the encoding for 'override' to 49 so that stuff vbr's so much more densely :-)




Swift SVN r20006
2014-07-16 05:32:17 +00:00
Chris Lattner
fe30b66a93 move "override" onto the new fangle decl modifier code, instead of being a virtual
attribute.  As part of this, introduce a new "NotSerialized" flag in Attr.def.
This eliminates a bunch of special case code in the parser and elsewhere for handling
this modifier.



Swift SVN r19997
2014-07-16 01:23:58 +00:00
Chris Lattner
ea8b7a471a Extend the attributes machinery to enforce the "OnFoo" flags in Attr.def at sema time, and extend
them to cover all declaration types.

This ensures that we reject attributes on declkinds where they don't make sense.  I went so far
as to make the QoI decent when an attribute can only be applied to a single kind of declaration
to make sure the error message says "@IBAction is only valid on 'func' declarations" as well.

This resolves <rdar://problem/17681151> 'dynamic' accepted by the compiler where it shouldn't be




Swift SVN r19982
2014-07-15 20:30:21 +00:00
Chris Lattner
a3c17dc166 move the @infix/@postfix/@prefix attributes onto the modern attribute infrastructure, NFC.
Swift SVN r19927
2014-07-14 14:30:26 +00:00
Chris Lattner
8078fecd18 fix <rdar://problem/17625253> Remove the 'strong' context-sensitive keyword
Swift SVN r19819
2014-07-10 19:46:22 +00:00
Doug Gregor
cb8e12368b Remove user conversions (@conversion __conversion) as a user feature.
We still have type checker support for user-defined conversions,
because the importer still synthesizes __conversion functions for CF
<-> NS classes.

Swift SVN r19813
2014-07-10 16:58:44 +00:00
Chris Lattner
fe95f81397 introduce a new 'DeclModifier' flag on attributes, which mark that the
attribute is a "modifier" of a decl, not an "attribute" and thus shouldn't
be spelt with an @ sign.  Teach the parser to parse "@foo" but reject it with
a nice diagnostic and a fixit if "foo" is a decl modifier.

Move 'dynamic' over to this (since it simplifies some code), and switch the
@optional and @required attributes to be declmodifiers (eliminating their @'s).



Swift SVN r19787
2014-07-10 05:49:10 +00:00
Joe Groff
9ddad1d718 Sema: Basic parsing and checking for 'dynamic'.
Parse 'dynamic' as a contextual keyword, and check that it's applied only to ObjC-compatible members of classes. We don't handle inheritance of the attribute yet.

Swift SVN r19781
2014-07-10 05:05:29 +00:00
Chris Lattner
35dcc3aa10 Move the @optional attribute to being a SIMPLE_DECL_ATTR, which
simplifies a bunch of code, makes them more consistent with the
other attributes, and is generally the newfangled way to do things.



Swift SVN r19779
2014-07-10 04:34:00 +00:00
Chris Lattner
02999cac51 Reinstate the @ on the @objc attribute. This is largely a revert of r19555 with a few tweaks.
Swift SVN r19706
2014-07-08 21:50:34 +00:00
Chris Lattner
7a56499d61 Start making @objc not start with an @ sign:
- Change the parser to accept "objc" without an @ sign as a contextual
   keyword, including the dance to handle the general parenthesized case.
 - Update all comments to refer to "objc" instead of "@objc".
 - Update all diagnostics accordingly.
 - Update all tests that fail due to the diagnostics change.
 - Switch the stdlib to use the new syntax.

This does not switch all tests to use the new syntax, nor does it warn about
the old syntax yet.  That will be forthcoming.  Also, this needs a bit of 
refactoring, which will be coming up.



Swift SVN r19555
2014-07-04 05:57:57 +00:00
Jordan Rose
b3addba09e Split "private(set)" attributes out into a separate class.
No functionality change, i.e. these still don't do anything.

Swift SVN r19547
2014-07-04 01:53:47 +00:00
Ted Kremenek
04095fb5aa *Start* importing more availability information from the SDK.
This is a WIP.  This patch includes:

- Adds version tuple information for 'introduced', 'deprecated',
  and 'obsoleted' to the 'availability' attribute.

- Add Clang importer support to import __attribute__((availability))
  version tuples into Swift as pieces of the 'availability'
  attribute.

- Add serialization support for the 'availability' attribute with
  this extra information.  This is not tested other than the
  tests currently passing.  This is not expected to be
  really exercised (with interesting versions) until
  parsing support is added for the version tuples.  However,
  existing @availability attributes in the test suite are being
  serialized, which should just include "empty" version information.

What's not in this patch:

- Parsing support in Swift for 'deprecated', 'introduced', or
  'obsoleted'.  All of this information is currently being pulled
  in from the Clang Importer.

- Warning support for using deprecated declarations based on the
  availability information and the minimum deployment target.

- Some harmony reconciling the 'IsUnavailable' field in
  AvailabilityAttr, which attempts to eagerly compute if something
  is unavailable so we don't have to replicate the checking logic
  elsewhere.  The idea is that when we either import availability
  information or lazily deserialize it we can compute whether or
  not something is conditionally unavailable or deprecated right
  there, and not have to have all clients within the frontend
  of the availability information need to pass the minimum
  deployment target.  Right now 'IsUnavailable' is also used
  to encode if the attribute represents unconditional unavailability,
  e.g. @availability(*, unavailable).

This patch, however, should contain enough information to start
looking at implementing weak linking support.

NOTE: the serialization of the attribute is a bit ugly.  I wasn't
certain if Jordan's serialization meta-programming supported
serializing values that decomposed into multiple values in a record,
so this ugly macro-based implementation is in place which compacts
all the version tuple information for an availability attribute
into a single record.

Swift SVN r19487
2014-07-02 23:13:14 +00:00
Manman Ren
b5a78c04ce [attribute] remove InlineKind::Always.
Swift SVN r19450
2014-07-02 05:57:20 +00:00
Manman Ren
2b2330bf5b [attribute] rename @noinline to @inline(never).
Add support for parsing inline(never), it can be easily expanded to
handle inline(always).

rdar://17527111


Swift SVN r19447
2014-07-02 01:27:05 +00:00
Jordan Rose
748efed913 [Accessibility] Protocol witnesses must be as accessible as the protocol.
...unless the type has less accessibility than the protocol, in which case
they must be as accessible as the type.

This restriction applies even with access control checking disabled, but
shouldn't affect any decls not already marked with access control modifiers.

Swift SVN r19382
2014-06-30 23:38:35 +00:00
Nadav Rotem
92d8688983 Add the @semantics attribute.
The @semantics attribute allows the stdlib to mark some functions as
having a specific semantics. The optimizer can use this information
to optimize the code.



Swift SVN r19328
2014-06-30 05:52:46 +00:00
Manman Ren
c777ae70b1 Use hasAttribute<NoinlineAttr>() instead of has(DAK_Noinline).
has(DAK_Noinline) will return true for attributes that have been marked invalid.

Thanks Dmitri for the feedback.


Swift SVN r19241
2014-06-26 18:08:48 +00:00
Ted Kremenek
a3b704ec42 Mark all imported APIs deprecated in iOS 7 (or earlier) or OS 10.9 (or earlier) as unavailable.
This patch takes care that synthesized initializers are created
when overriding unavailable initializers (thanks Doug).

Swift SVN r19202
2014-06-26 00:50:46 +00:00
Manman Ren
863138ead7 [noinline diagnostic] allow noinline on Func Decl only.
Also move noinline attribute to use the new infrastructure i.e SIMPLE_DECL_ATTR,
per feedback from Dmitri.

rdar://15882816


Swift SVN r19188
2014-06-25 23:17:43 +00:00
Manman Ren
7667b829bb [noinline attribute] add noinline attribute.
Propagate it to SILFunction and use it in PerformanceInliner. We also serialize
and parse the attribute.

rdar://15882816


Swift SVN r19150
2014-06-24 23:07:45 +00:00