42e827e | Devin Coughlin | 21 January 2016, 00:12:56 UTC | Revert "Revert "Sema: Always synthesize accessors for structs, unless they were imported from Clang"" This reverts commit f6dd24bb68b35292c3663c92c74508e97f247e53. | 21 January 2016, 01:22:15 UTC |
ac1f5a6 | Doug Gregor | 21 January 2016, 01:18:27 UTC | Fix diagnostic type list. | 21 January 2016, 01:18:27 UTC |
6331ef7 | Doug Gregor | 21 January 2016, 00:31:28 UTC | Add SE-0021 to the changelog | 21 January 2016, 01:12:36 UTC |
a5eeb16 | Doug Gregor | 13 January 2016, 21:40:20 UTC | Remove UnresolvedConstructorExpr. UnresolvedConstructorExpr is not providing any value here; it's essentially just UnresolvedDotExpr where the name refers to an initializer, so use that instead. NFC | 21 January 2016, 01:12:35 UTC |
ab09ef1 | Doug Gregor | 12 January 2016, 18:33:04 UTC | [SE-0021] Allow naming of specific initializers via "self.init(foo:bar:)". | 21 January 2016, 01:12:35 UTC |
4431ee0 | Doug Gregor | 12 January 2016, 07:14:49 UTC | Remove all vestiges of UnresolvedSelectorExpr. NFC | 21 January 2016, 01:12:34 UTC |
af6e196 | Doug Gregor | 12 January 2016, 07:09:27 UTC | Start parsing names with argument labels. Basic implementatation of SE-0021, naming functions with argument labels. Handle parsing of compound function names in various unqualified-identifier productions, updating the AST representation of various expressions from Identifiers to DeclNames. The result doesn't capture all of the source locations we want; more on that later. As part of this, remove the parsing code for the "selector-style" method names, since we now have a replacement. The feature was never publicized and doesn't make sense in Swift, so zap it outright. | 21 January 2016, 01:12:31 UTC |
ca33e31 | Doug Gregor | 20 January 2016, 21:41:06 UTC | [Serialization] Name lookup hack for mixed omit-needless-words/non-omit-needless-words. This egregious hack makes it so that we can properly dump APIs using "-enable-omit-needless-words" even when the compiler was built without it. | 21 January 2016, 01:01:07 UTC |
107f4ee | Doug Gregor | 21 January 2016, 00:19:44 UTC | Revert "Revert "[Clang importer] Always record module prefixes for stripping."" This reverts commit 5c4d1cff64c59a45c8585c7e9cbfd974410c50f6. | 21 January 2016, 01:01:07 UTC |
f5dd311 | Doug Gregor | 20 January 2016, 21:38:35 UTC | [Clang importer] Strip the "NS" prefix from entities in Foundation. As part of the improved import of Objective-C APIs into Swift, strip the "NS" prefix from entities defined in the Foundation framework. Addresses rdar://problem/24050011, which is part of SE-0005. Naturally, this is hidden behind -enable-omit-needless-words. | 21 January 2016, 01:01:07 UTC |
93f88fc | Doug Gregor | 20 January 2016, 21:34:22 UTC | [IRGen] Separate Swift and Objective-C class names Prefix stripping will make them differ. | 21 January 2016, 01:01:07 UTC |
94df20a | Doug Gregor | 20 January 2016, 21:31:49 UTC | [Clang importer] Write results into the Swift name lookup table at the end. This copes with rare (and hard-to-reproduce-in-the-small) recursive cases where importing a declaration for a name would then perform name lookup again, and end up not finding anything the second type around. | 21 January 2016, 01:01:06 UTC |
14ce25d | Doug Gregor | 19 January 2016, 18:32:52 UTC | [Omit needless words] Prune redundant "self" type following a verb in the base name. This allows us to prune UIViewController's "dismissViewControllerAnimated" to "dismissAnimated", eliminating unnecessary redundancy. | 21 January 2016, 01:01:06 UTC |
a101fb3 | Doug Gregor | 19 January 2016, 06:06:16 UTC | [Omit needless words] Don't strip just "Error". | 21 January 2016, 01:01:06 UTC |
dd6f7c2 | Doug Gregor | 20 January 2016, 23:17:01 UTC | Revert "[Clang importer] Strip the "NS" prefix from entities in Foundation" This reverts commit b5cb8252d0657fc91528ea46bfb73074f203e643. | 21 January 2016, 01:01:06 UTC |
c702b9d | Doug Gregor | 20 January 2016, 23:16:59 UTC | Revert "[Clang importer] Always record module prefixes for stripping." This reverts commit 07b2c88d8b04815fac1eaefc26cdd431980e2780. | 21 January 2016, 01:01:06 UTC |
44be782 | Jordan Rose | 16 January 2016, 02:37:43 UTC | [ClangImporter] Don't drop "Ref" from a CF types if it creates a conflict. https://bugs.swift.org/browse/SR-545 (cherry picked from commit 621edeb6640397958d80bd75bf97348265ddeaa7) | 20 January 2016, 23:31:13 UTC |
7202828 | Adrian Prantl | 20 January 2016, 23:20:39 UTC | Debug Info: Make sure to enter a new SILDebugScope before generating weak copies for capture list variables. Other wise two variables with identical names will end up in the same scope. This fixes a crash in the compiler. rdar://problem/24150188 SR-526 (cherry picked from commit f1315f8b5c2a5ab2506fed0b0760c60213942f37) | 20 January 2016, 23:21:38 UTC |
e129b6f | Slava Pestov | 20 January 2016, 22:46:46 UTC | Add dummy Optional declaration to MicroStdlib.swift test This broke when we switched on materializeForSet emission for all stored properties of structs. | 20 January 2016, 23:15:17 UTC |
ce7f06c | Slava Pestov | 20 January 2016, 22:21:35 UTC | SILGen: Fix materializeForSet emission for static computed properties of generic types The original implementation of SILGenMaterializeForSet.cpp did not support the case where SelfInterfaceType contained a generic parameter. Recent changes to switch SILGen to always open-code materializeForSet, and Sema to always emit accessors, exposed this in validation-tests. | 20 January 2016, 23:15:16 UTC |
7d4131f | Dmitri Gribenko | 20 January 2016, 23:14:42 UTC | Swift 3 migration: improve annotations on MutableSliceable | 20 January 2016, 23:14:49 UTC |
b29f84e | Dmitri Gribenko | 20 January 2016, 23:14:28 UTC | Swift 3 migration: correct annotations on IndexingGenerator | 20 January 2016, 23:14:49 UTC |
f6dd24b | Slava Pestov | 20 January 2016, 20:50:55 UTC | Revert "Sema: Always synthesize accessors for structs, unless they were imported from Clang" This reverts commit 8cbc026e8dfc6fffea120b14b8be32fb1a87469b. | 20 January 2016, 20:51:20 UTC |
01a3983 | Doug Gregor | 20 January 2016, 19:14:01 UTC | [Clang importer] Eliminate retain cycle involving the Clang importer implementation. The Clang importer implementation held on to a Clang instance, which held a list of module file extensions, which included the Clang importer implementation... break the cycle by separating out the module file extension from the Clang importer. Fixes SR-562 / rdar://problem/24225173. | 20 January 2016, 19:51:45 UTC |
ff57c68 | Xi Ge | 20 January 2016, 18:46:31 UTC | [SourceKit][CodeFormat] When indent to siblings, respect tuple elements' names. rdar://24251847 | 20 January 2016, 19:43:11 UTC |
8cbc026 | Slava Pestov | 20 January 2016, 08:40:25 UTC | Sema: Always synthesize accessors for structs, unless they were imported from Clang This fixes the issue that "SILGen: Correctly emit accessors synthesized to witness protocol requirements" was meant to solve, but in a simpler way. A better fix would be to first address the issue where @_transparent function bodies are not serialized in some cases, and then only emit synthesized accessors as needed, in the original version of this patch. To fix the duplicate symbol issues, we would emit the synthesized accessors with shared linkage, which would always work once serialized bodies were available. For resilient structs of course, we'll always need to emit accessors anyway. | 20 January 2016, 08:55:12 UTC |
fc8ddcf | Slava Pestov | 20 January 2016, 08:51:11 UTC | Sema: Fix for setter->isMutating() check before setter was type checked With an upcoming patch we would call setMutating() on materializeForSet before computing the setter's isMutating() in the case where a setter was explicitly declared 'nonmutating'. Fix that by replacing the setter->isMutating() call with a direct computation of the expected result. It seems that the materializeForSet of protocol protocol requirements has to be mutating, even if the protocol is a class protocol or the property is nonmutating -- I need to investigate why and fix SILGen to not make this assumption, but in the meantime, opt-out of the new logic with protocol requirements to avoid more breakage. | 20 January 2016, 08:55:05 UTC |
1ba8f45 | Slava Pestov | 20 January 2016, 08:40:45 UTC | SILGen: Targeted fix for MaterializeForSet use If we're accessing a field of a struct, ignore materializeForSet; it will only have been synthesized if the struct conforms to a protocol, and we don't want the presence of a conformance to affect generated code. A more principled fix would change the SILGen logic to use a materializeForSet if it would have been synthesized anyway, asserting that it was synthesized in that case. I'll clean this up in the 'master' branch once these fixes settle. | 20 January 2016, 08:55:05 UTC |
87fed07 | Slava Pestov | 20 January 2016, 08:49:50 UTC | Sema: Fix checkMutating() for stored properties that had accessors synthesized for them already Otherwise, a struct's static property cannot witness two different protocol requirements, because the second one thinks it is mutating. Probably we should push the isInstanceMember() check up into the code that computes isMutating(). A more principled cleanup for this code is coming soon. | 20 January 2016, 08:55:05 UTC |
61d5657 | Slava Pestov | 20 January 2016, 06:50:23 UTC | Tweak some -parse-stdlib tests While making some changes that resulted in materializeForSet getting synthesized more often, I noticed that some tests started failing because they could not find the Optional type, which appears in the type of materializeForSet. Fix these tests by defining an Optional and passing in '-module-name Swift' so that we can find it. This feels a little bit gross, but Dmitri says it is preferrable to importing Swift from a test. | 20 January 2016, 08:55:04 UTC |
9fd2957 | Dmitri Gribenko | 20 January 2016, 05:01:10 UTC | Adjust tests for new deprecation annotations | 20 January 2016, 05:01:10 UTC |
4105067 | Dmitri Gribenko | 20 January 2016, 03:47:24 UTC | Swift 3 migration: standardize annotations for to-property migration | 20 January 2016, 03:47:24 UTC |
9df25b5 | Dmitri Gribenko | 20 January 2016, 03:44:41 UTC | Swift 3 migration: use the same annotations on all init()s for unsafe pointers | 20 January 2016, 03:45:10 UTC |
82ca8f8 | Dmitri Gribenko | 19 January 2016, 20:23:56 UTC | Add migration attributes for Swift 3 | 20 January 2016, 03:45:10 UTC |
88cedc3 | Jordan Rose | 15 January 2016, 17:54:49 UTC | [test] Add explicit targets to Driver/filelists.swift. These checks depend on Darwin-style linker invocations. | 20 January 2016, 01:36:12 UTC |
1641d7b | Jordan Rose | 14 January 2016, 02:33:31 UTC | [Driver] Write per-job filelists even under -save-temps. Same as the previous commit; here's a test case that triggers it. Last bit of https://bugs.swift.org/browse/SR-280. | 20 January 2016, 01:36:12 UTC |
93e68e9 | Jordan Rose | 14 January 2016, 02:15:02 UTC | Teach the frontend about -output-filelist. Finishes up https://bugs.swift.org/browse/SR-280. At this point, nothing should be passing O(N) arguments on the command line, where N is the number of input source files, unless N is small. (There are other inputs which are passed through to subtools, and these are not put into filelists. That's fine.) | 20 January 2016, 01:36:12 UTC |
8fb064c | Jordan Rose | 14 January 2016, 01:56:56 UTC | [Driver] Use filelists for the *output* of a threaded WMO build. This is only the driver side of the work; the frontend doesn't understand this new -output-filelist option yet. Next commit. More https://bugs.swift.org/browse/SR-280. | 20 January 2016, 01:36:12 UTC |
7523a16 | Jordan Rose | 14 January 2016, 00:47:53 UTC | [Driver] Use -filelist for the merge-module step. More https://bugs.swift.org/browse/SR-280. | 20 January 2016, 01:36:12 UTC |
7cc1638 | Jordan Rose | 14 January 2016, 00:42:45 UTC | [Driver] Write the all-sources file even under -save-temps. We optimize subprocess invocation to a simple execve() if there's no cleanup work to do, but that doesn't get us out of doing /setup/ work. | 20 January 2016, 01:36:12 UTC |
adb5dd9 | Jordan Rose | 14 January 2016, 00:25:40 UTC | [Driver] Emit filelists for Jobs that want them (inputs). This is the other half of the previous commit: we can use -filelist for linking on Darwin. More https://bugs.swift.org/browse/SR-280. | 20 January 2016, 01:36:12 UTC |
1514403 | Jordan Rose | 13 January 2016, 23:20:23 UTC | [Driver] Pass -filelist to Darwin 'ld' too. Start sketching out a way for individual jobs to request filelists for their inputs or their outputs. This should cover all the cases mentioned in ad945426. More https://bugs.swift.org/browse/SR-280. | 20 January 2016, 01:36:12 UTC |
9618eb3 | Jordan Rose | 13 January 2016, 21:13:44 UTC | [Driver] Pass around 'const char *' instead of 'const std::string &'. No functionality change -- lifetime extension is handled by the ArgumentList already owned by the Compilation. | 20 January 2016, 01:36:12 UTC |
8e6f58d | Dmitri Gribenko | 20 January 2016, 01:21:42 UTC | Merge pull request #1018 from shahmishal/swift-2.2-branch Remove llbuild, swiftpm, xctest and foundation from preset | 20 January 2016, 01:21:42 UTC |
b05e392 | Mishal Shah | 20 January 2016, 01:16:33 UTC | Remove llbuild, swiftpm, and foundation from preset | 20 January 2016, 01:16:33 UTC |
f012518 | Slava Pestov | 19 January 2016, 22:10:47 UTC | Revert "SILGen: Correctly emit accessors synthesized to witness protocol requirements" This reverts commit 10c8ce824fe6732152b514b03b1c2cc2f37a88e3. | 19 January 2016, 22:22:15 UTC |
8280bb3 | Slava Pestov | 16 January 2016, 10:02:23 UTC | SILGen: Fix materializeForSet emission for static stored properties Lower the metatype instead of brazenly declaring it @thick. This was triggering the SIL verifier with the newly-added test case. | 19 January 2016, 22:22:14 UTC |
4d903f9 | Slava Pestov | 16 January 2016, 02:37:57 UTC | IRGen: Fix emission of conformances for concrete subclasses of generic classes In master, this was fixed by Luke Howard as part of some other changes in the following patch: <https://github.com/apple/swift/commit/b5880f386b814f8b5cb220d1c4e65e74c04eccf4> This patch back-ports the relevant part of the above (checking for a true return value from hasMetadataPattern() vs checking for a BoundGenericType). Fixes <rdar://problem/24183374>. | 19 January 2016, 22:22:14 UTC |
b971595 | Michael Ilseman | 14 January 2016, 02:38:26 UTC | [Diagnostics] -suppress-warnings and -warnings-as-errors flags Exposes the global warning suppression and treatment as errors functionality to the Swift driver. Introduces the flags "-suppress-warnings" and "-warnings-as-errors". Test case include. | 19 January 2016, 21:53:28 UTC |
a139ad4 | Michael Ilseman | 14 January 2016, 01:30:20 UTC | [Diagnostics] Add in treating warnings as errors | 19 January 2016, 21:53:27 UTC |
36d706e | Michael Ilseman | 13 January 2016, 00:40:52 UTC | [Diagnostics] Remove all categories Diagnostic categories are entirely unused and arguably useless as implemented, as they merely denote the sub-component of the compiler. As far as categorizing warnings are concerned, I'm abandoning the effort for now, as the utility is marginal and Swift and the Swift compiler are probalby not ready for these to be nailed down. For the sake of cleanliness, the CATEGORY field is also stripped from WARNINGS. If there's a need for automatic identifying of compiler sub-components for diagnstics in the future, there are better ways to do this. NFC Swift-2.2: resovled merge conflicts, and updated some new diagnostics. | 19 January 2016, 21:25:25 UTC |
14bd364 | Michael Ilseman | 13 January 2016, 00:09:16 UTC | [Diagnostics] Refactor and clarify diagnostic behavior rules. Restores StoredDiagnosticInfo, which is useful to help distinguish when the user explicitly modifies the behavior of a diagnostic vs we're just picking up the default kind. Adds some clarifying comments, and lays out the suppression workflow, whereby different types of suppression (per-diagnostic, per-category, etc) have different precedence levels. | 19 January 2016, 21:25:25 UTC |
6824879 | Max Moiseev | 15 January 2016, 00:50:52 UTC | applying swift3_migration attribute to things being renamed in Swift3 | 19 January 2016, 20:21:12 UTC |
07b2c88 | Doug Gregor | 19 January 2016, 05:26:52 UTC | [Clang importer] Always record module prefixes for stripping. The Swift 3 migration mode requires us to be able to compute the Swift 3 names for entities even when we aren't in omit-needless-words mode; recording module prefixes unconditionally ensures that we can do prefix stripping under these circumstances. | 19 January 2016, 05:28:59 UTC |
b5cb825 | Doug Gregor | 18 January 2016, 07:14:46 UTC | [Clang importer] Strip the "NS" prefix from entities in Foundation As part of the improved import of Objective-C APIs into Swift, strip the "NS" prefix from entities defined in the Foundation framework. Addresses rdar://problem/24050011, which is part of SE-0005. Naturally, this is hidden behind -enable-omit-needless-words. | 19 January 2016, 05:28:59 UTC |
651f025 | Doug Gregor | 18 January 2016, 05:58:45 UTC | Also map -enable-omit-needless-words to a language option. Although omit-needless-words is almost entirely a Clang importer task, there are a handful of other places in the compiler that will need to query this flag as well. NFC for now; those changes will come soon. | 19 January 2016, 05:28:59 UTC |
31eaa2b | Doug Gregor | 18 January 2016, 05:56:18 UTC | Abstract the set of known Foundation entities into a .def-driven enum. NFC Specifically, we don't want to hard-code the Swift names of these Objective-C entities, because the importer renaming will affect them. | 19 January 2016, 05:28:59 UTC |
ec881e3 | Ted Kremenek | 18 January 2016, 20:12:08 UTC | Merge pull request #1007 from tkremenek/fix-swift2.1-regressions Fix two regressions from Swift 2.1 related to parameter parsing. | 18 January 2016, 20:12:08 UTC |
cb2ffe3 | Chris Lattner | 18 January 2016, 17:02:56 UTC | fix SR-573 - Crash with invalid parameter declaration This makes sure to diagnose the error of not having a type on a parameter in the parser, which is what sema expects. # Conflicts: # include/swift/AST/DiagnosticsParse.def | 18 January 2016, 20:02:27 UTC |
59b9df0 | Chris Lattner | 18 January 2016, 16:49:09 UTC | Simplify mapParsedParameters/createParam by passing the ParsedParameter into it, instead of passing a bunch of pieces into it. NFC. | 18 January 2016, 19:55:40 UTC |
60f624e | Andrew Trick | 17 January 2016, 22:55:42 UTC | [SILOpt] Fix speculative devirtualization miscompile. Properly handle limiting the number of speculative checks. This bug dates back to 2015-09-15, Swift 2.1. Fixes rdar:23228386. (cherry picked from commit e8fc64cdaec6e3f0312ada52d653ceb7fa1246a9) | 18 January 2016, 02:59:16 UTC |
9c08d14 | Doug Gregor | 17 January 2016, 05:10:34 UTC | [Swift 3 migration] Infer @swift3_migration for witnesses from requirements. | 17 January 2016, 05:10:34 UTC |
0cabae5 | Doug Gregor | 16 January 2016, 06:47:51 UTC | [Swift 3 migration] Handle renaming for dynamic member references. This deals with uses of members of "AnyObject" that have changed names. | 16 January 2016, 06:48:36 UTC |
4a5fe8c | Doug Gregor | 16 January 2016, 05:56:06 UTC | [Swift 3 migration] Handle OtherConstructorDeclRefExpr ... for "super.init" and "self.init"'s to renamed initializers. | 16 January 2016, 06:48:36 UTC |
03b78d7 | Doug Gregor | 16 January 2016, 05:50:41 UTC | [Clang importer] Don't infer @swift3_migration to typedefs. ... because Swift 3 does *not* rename AnyObject to id :) | 16 January 2016, 06:48:36 UTC |
2ed463a | gregomni | 13 January 2016, 17:24:49 UTC | [SR-533] Improve C-style for fixits by also removing 'var' keyword from loop var decl. | 16 January 2016, 05:51:56 UTC |
258f05c | Chris Hanson | 16 January 2016, 01:07:57 UTC | XCTest Overlay: Treat ObjC exceptions as unexpected failures When an assertion fails due to an Objective-C exception, that should be treated as an unexpected failure, not an expected failure. Addresses rdar://problem/24114025. | 16 January 2016, 01:07:57 UTC |
c0e85ce | Doug Gregor | 16 January 2016, 00:16:43 UTC | [Swift 3 Migration] Add initial Swift 3 migration pass. When -swift3-migration is passed to the front end, perform the transformations specified by the @swift3_migration attribute. For now, we support renaming in a number of places in the AST. More transformations and better coverage to come. | 16 January 2016, 00:20:43 UTC |
b3bfe58 | Doug Gregor | 16 January 2016, 00:09:52 UTC | Remove redundant, bogus Fix-It | 16 January 2016, 00:20:43 UTC |
b0d95b7 | Doug Gregor | 15 January 2016, 19:02:42 UTC | Parse attributes on generic type parameter declarations and check availability. | 15 January 2016, 23:24:52 UTC |
05fc8c9 | Doug Gregor | 15 January 2016, 00:36:50 UTC | Deserialize swift3_migration attribute with an empty "renamed" string. | 15 January 2016, 00:38:00 UTC |
9e83511 | Erik Eckstein | 14 January 2016, 22:59:05 UTC | EscapeAnalysis: don't crash in case a function_ref is passed to a C-function pointer argument. Fixes rdar://problem/24183323 | 14 January 2016, 23:24:16 UTC |
202409e | Doug Gregor | 14 January 2016, 21:13:32 UTC | [Clang importer] Infer swift3_migration attribute for imported declarations. When the name of an imported declaration will change when going from Swift 2 to Swift 3, add the appropriate swift3_migration attribute. | 14 January 2016, 21:39:30 UTC |
df0c5fc | Chris Hanson | 14 January 2016, 20:57:06 UTC | XCTest Overlay: Allow assertion expressions to throw When an assertion expression throws an error, treat that as an unexpected failure of the assertion. Also generate the failure message for the error directly in the overlay, rather than routing through _XCTFailureDescription, to avoid revlock between the overlay and XCTest.framework's implementation. Addresses rdar://problem/23789893. | 14 January 2016, 21:15:57 UTC |
8a68f84 | Chris Hanson | 14 January 2016, 02:07:31 UTC | XCTest Overlay: Add XCTAssertThrowsError assertion Add an assertion that can be used to check that an expression throws an error, and if it does to run a block against that error which may contain further checks. Addresses rdar://problem/23789904. | 14 January 2016, 21:15:52 UTC |
2b7f36f | Todd Fiala | 14 January 2016, 18:00:13 UTC | lldb build: allow optional xcodebuild flags to be passed to build-script-impl | 14 January 2016, 20:02:18 UTC |
c9c8def | Doug Gregor | 14 January 2016, 19:44:02 UTC | [Clang importer] Teach importFullName to provide argument labels for functions. importFullName was inconsistent in the kinds of names it produced for imported C functions. swift_name'd functions would get argument labels, but non-swift_name'd functions would not, and other parts of the important were working around the oddity. Make importFullName() always provide the argument labels. | 14 January 2016, 19:51:02 UTC |
b52abd5 | Doug Gregor | 14 January 2016, 16:49:18 UTC | Fix link ordering. | 14 January 2016, 19:51:02 UTC |
8bd5987 | Doug Gregor | 14 January 2016, 06:46:10 UTC | Add a "-swift3-migration" frontend option and corresponding language option. This is intended to enable various Fix-Its to perform the one-way migration from Swift 2.x to Swift 3, with a focus on the issues involving naming. | 14 January 2016, 16:21:22 UTC |
2274db4 | Doug Gregor | 14 January 2016, 06:22:21 UTC | Remove the "-disable-self-type-mangling" flag and corresponding language option NFC | 14 January 2016, 16:21:21 UTC |
ffd6054 | Doug Gregor | 14 January 2016, 05:42:05 UTC | Attempt to fix linking on Linux | 14 January 2016, 16:21:06 UTC |
1a8427b | Doug Gregor | 14 January 2016, 00:18:47 UTC | Add a swift3_migration attribute to describe how an API gets migrated. Introduce a new attribute, swift3_migration, that lets us describe the transformation required to map a Swift 2.x API into its Swift 3 equivalent. The only transformation understood now is "renamed" (to some other declaration name), but there's a message field where we can record information about other changes. The attribute can grow somewhat (e.g., to represent parameter reordering) as we need it. Right now, we do nothing but store and validate this attribute. | 14 January 2016, 00:58:27 UTC |
ce45559 | Doug Gregor | 13 January 2016, 22:10:11 UTC | Revert "[AST] Introduce internal attribute '_migration_id'." This reverts commit 042efbfb267febc4bf598cc65ce87996b8222a37. We're going to take a different approach to the migration attribute. | 14 January 2016, 00:58:26 UTC |
ae250af | Dmitri Gribenko | 13 January 2016, 16:32:10 UTC | Merge remote-tracking branch 'origin/master' into swift-2.2-branch | 13 January 2016, 16:32:10 UTC |
e711d5e | Slava Pestov | 13 January 2016, 00:07:18 UTC | Preliminary evolution tests Checking these in to get some discussion over FileCheck style and test file structure going before I write too many more tests. | 13 January 2016, 09:32:27 UTC |
199468f | Dmitri Gribenko | 13 January 2016, 09:08:06 UTC | Merge pull request #954 from apple/revert-950-gyb-without-encoding Revert "[gyb] Force UTF-8 encoding when parsing templates on Linux" | 13 January 2016, 09:08:06 UTC |
1b44809 | Dmitri Gribenko | 13 January 2016, 09:07:53 UTC | Revert "[gyb] Force UTF-8 encoding when parsing templates on Linux" | 13 January 2016, 09:07:53 UTC |
81dfb77 | rjmccall | 13 January 2016, 08:52:41 UTC | Merge pull request #953 from practicalswift/typo-fixes-20160113 [gardening] Remove recently introduced typo | 13 January 2016, 08:52:41 UTC |
4a49c2d | practicalswift | 13 January 2016, 08:33:23 UTC | [gardening] Remove recently introduced typo :-) | 13 January 2016, 08:33:23 UTC |
92175c4 | rjmccall | 13 January 2016, 08:24:25 UTC | Merge pull request #952 from practicalswift/remove-binary [gardening] Remove binary "nix" - should not be in repo | 13 January 2016, 08:24:25 UTC |
6f0ac35 | practicalswift | 13 January 2016, 08:09:48 UTC | Remove binary "nix" introduced in commit 94ebcb847b3aae799da1cdb9374ad090f75d86dc $ file nix nix: Mach-O 64-bit x86_64 executable $ git log nix commit 94ebcb847b3aae799da1cdb9374ad090f75d86dc Author: John McCall <rjmccall@apple.com> Date: Tue Jan 12 23:34:54 2016 -0800 Handle inherited concrete conformances correctly on witness_method instructions. rdar://24141848 | 13 January 2016, 08:09:48 UTC |
94ebcb8 | John McCall | 13 January 2016, 07:34:54 UTC | Handle inherited concrete conformances correctly on witness_method instructions. rdar://24141848 | 13 January 2016, 07:36:21 UTC |
094098a | Slava Pestov | 13 January 2016, 07:06:55 UTC | IRGen: Don't walk into resilient types when emitting static value witness tables references This was the first real bug caught by the library evolution tests that I'm working on! | 13 January 2016, 07:29:22 UTC |
1c14b49 | Ted Kremenek | 13 January 2016, 05:51:54 UTC | Merge pull request #950 from RLovelett/gyb-without-encoding [gyb] Force UTF-8 encoding when parsing templates on Linux | 13 January 2016, 05:51:54 UTC |
c8b1aa2 | Dmitri Gribenko | 13 January 2016, 04:39:02 UTC | Merge remote-tracking branch 'origin/master' into swift-2.2-branch | 13 January 2016, 04:39:02 UTC |
03f5f01 | Dmitri Gribenko | 13 January 2016, 04:38:31 UTC | test/Frontend/filelist.swift: don't fail if a temporary file does not exist | 13 January 2016, 04:38:31 UTC |
ad94542 | Jordan Rose | 13 January 2016, 02:05:43 UTC | Teach the frontend about -filelist for input files. With this, we're out of the business of passing large numbers of input files on the command line to the frontend, which means we no longer overflow argv with a mere 1100 input files under whole-module optimization. In order to make sure this doesn't happen again, I'd like to also get this working for - swiftmodule inputs to the merge-module build phase - /output/ files for multithreading single-frontend builds (WMO) - object file inputs to the linker on OS X (response files for binutils ld have different quoting rules) Part 3 of https://bugs.swift.org/browse/SR-280. | 13 January 2016, 03:20:50 UTC |
fe00083 | Jordan Rose | 13 January 2016, 01:04:25 UTC | [Driver] Actually write out the source file list file. This is simply a newline-separated list of files, matching the format from Darwin ld's -filelist option. Part 2 of https://bugs.swift.org/browse/SR-280. | 13 January 2016, 03:20:50 UTC |
b45a69e | Jordan Rose | 13 January 2016, 00:43:21 UTC | [Driver] Allow passing all source files in a filelist. Generate frontend commands with -filelist in them. This isn't actually implemented yet, but we can start testing at this point. Part 1 of https://bugs.swift.org/browse/SR-280. | 13 January 2016, 03:20:50 UTC |
67eaab6 | John McCall | 13 January 2016, 03:11:25 UTC | Give _swift_dead_method_stub internal linkage; the Mach-O linker doesn't support aliases to weak symbols. rdar://24121475 | 13 January 2016, 03:12:20 UTC |