The heuristics to decide whether fully qualified type names should be printed may
work differently when generating the baseline and when importing from just built
frameworks. This patch makes it consistent so that such false positives won't happen.
rdar://54276347
Inlinable decls from swift interface files allow us to construct the underlying types
of opaque result types, whose changes can break ABI.
rdar://52273137
Removing accessors other than getter and setter can be ABI breaking. This
patch starts to formally include all accessor decls in the tree and diagnose
their removal. This change only applies to the ABI checker since we still
exclude accessors other than getter and setter when diagnosing source
compatibility.
Including accessors formally can also allow us to check the missing
of availability attributes for newly added accessors.
rdar://52063421
With module stability from the compiler, we can now use an earlier version
of the SDK as the baseline and compare the new SDK with it. This patch
adds a new -bsdk flag to the tool to specify the baseline SDK path.
We used to use a Json format to capture the ABI/API detail of Swift modules to
walk-around module compatibility issues across different compiler versions. Since
Swift module now is in stable format, we can compare two swiftinterface files
directly without dumping the content into Json format.
We started to diagnose new APIs such as the missing of @available attributes. This diverges
the expected change lists for builds with/without assertions enabled because some stdlib
symbols are included only when asserts are enabled. We start to require
asserts here to make sure the test doesn't fail on no-assertion settings.
New ABIs should have an @available attribute to describe the introducing
version. This patch teaches the tool to diagnose its missing.
Decls with @_AlwaysEmitIntoClient are excluded from the blaming lists since
they are essentially available all the time.
rdar://51089418
External type declarations are synthesized to incorporate members in extensions to types
of external modules. In diagnostics, we should use 'extension' instead of 'struct/class'
for these decls to avoid confusion.
To incorporate extensions to types from other modules, the tool sometimes needs
to pull type declaration from external modules even though these modules
are not under checking. We need a flag to explicitly mark such case.
SDK conformance nodes include only protocol
`name` and `printedName`, both are demangled.
USR field of the protocol type is added, to be able
to uniquely identify conformances.
This implements the protocols and static functions proposed in SE-0246, plus some initial test coverage. It also has some rough accompanying cleanup of tgmath. It does not include the globals (on scalars or SIMD types) nor does it deprecate much in tgmath.h.
For some unclear reasons, calling getAllConformances on Int.Words and
UInt.Words returns duplicate entries for conforming to RandomAccessCollection.
Since this isn't the case for swift-5.1-branch, we saw false positives shown
in rdar://49568079.
This patch fixes the ABI/API checker by de-duplicate results collected from
getAllConformances.
* Revert "Merge pull request #23791 from compnerd/you-know-nothing-clang"
This reverts commit 5150981150, reversing
changes made to 8fc305c03e.
* Revert "Merge pull request #23780 from compnerd/math-is-terrible"
This reverts commit 2d7fedd25f, reversing
changes made to 0205150b8f.
* Revert "Merge pull request #23140 from stephentyrone/mafs"
This reverts commit 777750dc51, reversing
changes made to 0c8920e747.
This commit implements SE-0246, by adding conformance to Real to the Float, CGFloat, Double, and Float80 types, implemented either in terms of the system's C math library, existing standard library functionality, or LLVM intrinsics. It includes basic test coverage for these new functions, and deprecates and obsoletes *some* existing functionality in the Platform overlay. We still need to make a decision about how to handle the remaining "tgmath" functions, because obsoleting them is technically a source-breaking change (if users have unqualified names like "exp(1)", it's fine, but it would break users who have used qualified names like "Darwin.exp(1)".)
ABI placeholders are decls with attribute '@available(macOS 9999, iOS
9999, tvOS 9999, watchOS 9999, *)'. The diagnostics phase could be
forgiving for ABI breakages on these decls since they are added
recently. This patch adds a new flag to the json file indicating whether
a declaration or a conformance is an ABI placeholder. The checking of
placeholder is transitive, meaning a decl is an ABI placeholder if its
decl context is one.
rdar://49502365
Protocol requirements may not necessarily add new entries to the witness table if
it's inherited from super protocol. This patch teaches the json dump to
include a flag indicating whether a protocol requirement requires new
witness table entry and diagnoses the change of such flag as ABI
breakages.
rdar://47657204
Otherwise we can get in trouble when a local type is named, say,
'Sequence'.
Also contains test updates and a fix for Harlan's previous commit,
which actually affects all typealiases, not just those in the Builtin
module.
This avoids us having to pattern match every source file which should
help speed up the CMake generation. A secondary optimization is
possible with CMake 3.14 which has the ability to remove the last
extension component without having to resort to regular expressions. It
also helps easily identify the GYB'ed sources.
Sorting of DeclContext-local protocols and conformances shouldn't ever
be necessary, because the underlying data structures that produce
these lists should be deterministic. Sorting can hide any
non-determinism, so stop doing it and we can address the underlying
nondeterminism.
Upon seeing an extension for a type outside the current module, the digester
creates a dummy type node and puts all the extensions’ members and conformances
in that type. This allows us to track new API endpoints that are retroactively
added.
Unfortunately, if there are no public members/conformances (only internal or
private ones), the type itself ends up in the SDK dump without any public
children. This causes an issue when you dump the SDK from a parseable
interface, where the internal extension was not printed.
Without this change, SILGen will crash when compiling a use of the
derived protocol's requirement: it will instead attempt to use
the base protocol's requirement, but the code will have been
type-checked incorrectly for that.
This has a potential for source-compatibility impact if anyone's
using explicit override checking for their protocol requirements:
reasonable idioms like overriding a mutating requirement with a
non-mutating one will no longer count as an override. However,
this is arguably a bug-fix, because the current designed intent
of protocol override checking is to not allow any differences in
type, even "covariant" changes like making a mutating requirement
non-mutating. Moreover, we believe explicit override checking in
protocols is quite uncommon, so the overall compatibility impact
will be low.
This also has a potential for ABI impact whenever something that
was once an override becomes a non-override and thus requires a
new entry. It might require a contrived test case to demonstrate
that while using the derived entry, but it's quite possible to
imagine a situation where the derived entry is not used directly
but nonetheless has ABI impact.
Furthermore, as part of developing this patch (earlier versions of
which used stricter rules in places), I discovered a number of
places where the standard library was unintentionally introducing
a new requirement in a derived protocol when it intended only to
guide associated type deduction. Fixing that (as I have in this
patch) *definitely* has ABI impact.
