This flag will enable all experimental differentiable programming features.
The default will be `true` on tensorflow branch but `false` on master branch.
Features will first be updated on tensorflow branch to use this flag, before
being upstreamed to master. The goal is to achieve a minimal code diff between
the two branches.
The [TF-820](https://bugs.swift.org/browse/TF-820) master issue tracks upstreaming differentiable programming.
---
Rationale: we chose to add a frontend flag rather than a `build-script`/CMake flag for easier testing. Differentiable programming `lit` tests can be run by specifying this additional flag without recompiling the compiler and standard library.
[Forum discussion on upstreaming differentiable programming.](https://forums.swift.org/t/upstreaming-differentiable-attribute-and-differentiable-protocol/26821)
This flag, currently staged in as `-experimental-skip-non-inlinable-function-bodies`, will cause the typechecker to skip typechecking bodies of functions that will not be serialized in the resulting `.swiftmodule`. This patch also includes a SIL verifier that ensures that we don’t accidentally include a body that we should have skipped.
There is still some work left to make sure the emitted .swiftmodule is exactly the same as what’s emitted without the flag, which is what’s causing the benchmark noise above. I’ll be committing follow-up patches to address those, but for now I’m going to land the implementation behind a flag.
Add `-no-toolchain-stdlib-rpath` flag: the negative version of
`-toolchain-stdlib-rpath`.
Make `-no-toolchain-stdlib-rpath` be the default: use `/usr/lib/swift` as
default RPATH on Darwin platforms instead of toolchain standard library.
Adapted from https://github.com/apple/swift/pull/27206.
tensorflow branch requires the opposite default (use toolchain standard
library as RPATH) because some stdlib modules like TensorFlow do not exist in
`/usr/lib/swift`.
This flag adds diagnostic names to the end of their messages, e.g. 'error: cannot convert value of type '[Any]' to specified type '[Int]' [cannot_convert_initializer_value]'. It's intended to be used for debugging purposes when working on the compiler.
We use one bit of the third reserved swift private tls key.
Also move the functionality into a separate static archive that is
always linked dependent on deployment target.
Add the command line option -require-explicit-availability to detect public
or `@usableFromInline` declarations and warn if they don't declare
an introduction OS version. This option should catch forgotten `@available`
attributes in frameworks where all services are expected to be introduced
by an OS version.
The option -require-explicit-availability-target "macOS 10.14, iOS 12.0"
can be specified for the compiler to suggest fix-its with the missing
attributes `@available(macOS 10.14, iOS 12.0, *)`.
rdar://51001662
Many build systems that support Swift don't use swiftc to drive the linker. To make things
easier for these build systems, also use autolinking to pull in the needed compatibility
libraries. This is less ideal than letting the driver add it at link time, since individual
compile jobs don't know whether they're building an executable or not. Introduce a
`-disable-autolink-runtime-compatibility` flag, which build systems that do drive the linker
with swiftc can pass to avoid autolinking.
rdar://problem/50057445
On Windows, there are multiple variants of the C runtime that must be
explicitly specified and consistently used from the runtime to the
application. The new `-libc` option allows us to control the linking
phase by correctly embedding the requested library to be linked. It is
made into a required parameter on Windows and will add in the
appropriate flags for the imported C headers as well. This ensures that
the C library is not incorrectly linked.
Otherwise the overlays won't be force-loaded when you import them.
(This feature could possibly be scoped back at this point, but it's
not something that should be changed just by using module interfaces!)
This allows us to mark this flag as DoesNotAffectIncrementalBuild, so
that switching between a tty and non-tty doesn't cause full rebuilds
with swiftpm.
https://bugs.swift.org/browse/SR-7982
Leave the old flag in as an alias to the new flag, for transition
purposes. Also go ahead and remove the long-deprecated and unused
`emit-interface-path`.
Part of rdar://49359734
...and remove the option. This is ~technically~ CLI-breaking because
Swift 5 shipped this as a hidden driver option, but it wouldn't have
/done/ anything in Swift 5, so I think it's okay to remove.
Note that if a parseable interface (.swiftinterface) and a binary
interface (.swiftmodule) are both present, the binary one will still
be preferred. This just /allows/ parseable interfaces to be used.
rdar://problem/36885834
By default, we compile the standard library with
`-enforce-exclusivity=unchecked`. If we don't preserve this argument,
then the standard library compiled from an interface includes
exclusivity enforcement, which pessimizes inlining those functions,
which decreases performance for clients.
Part of rdar://46431767
Windows doesn't know what a shebang is, so it's unable to run tests that
use -driver-use-frontend-path with a script. This allows the script
interpreter to be run as the executable with the script as its first
argument. e.g. --driver-use-frontend-path "python;my-script.py"
This patch allows `-serialize-diagnostics-path` for the interpret mode.
There is one file compiled in such mode, so it makes sense to support
this flag to specify an explicit output path for diagnostics emission.
Resolves: SR-9670
<rdar://problem/46548531> Extend @available to support PackageDescription
This introduces a new private availability kind "_PackageDescription" to
allow availability testing by an arbitary version that can be passed
using a new command-line flag "-swiftpm-manifest-version". The semantics
are exactly same as Swift version specific availability. In longer term,
it maybe possible to remove this enhancement once there is
a language-level availability support for 3rd party libraries.
Motivation:
Swift packages are configured using a Package.swift manifest file. The
manifest file uses a library called PackageDescription, which contains
various settings that can be configured for a package. The new additions
in the PackageDescription APIs are gated behind a "tools version" that
every manifest must declare. This means, packages don't automatically
get access to the new APIs. They need to update their declared tools
version in order to use the new API. This is basically similar to the
minimum deployment target version we have for our OSes.
This gating is important for allowing packages to maintain backwards
compatibility. SwiftPM currently checks for API usages at runtime in
order to implement this gating. This works reasonably well but can lead
to a poor experience with features like code-completion and module
interface generation in IDEs and editors (that use sourcekit-lsp) as
SwiftPM has no control over these features.
In the Darwin toolchain the linker is invoked directly, and compiler_rt
is used if it is found, but in Unix platforms, clang++ is invoked
instead, and the clang driver will invoke the linker. Howerver there was
no way of modifying this clang++ invocation, so there's no way of
providing `--rtlib=` and change the platform default (which is normally
libgcc). The only workaround is doing the work that the Swift driver is
doing "manually".
The change adds a new option (with help hidden, but we can change that)
to allow providing extra arguments to the clang++ invocation. The change
is done in the two places in the Unix and Windows toolchains that I
found the clang driver was being used.
Includes some simple tests.
A module compiled with `-enable-private-imports` allows other modules to
import private declarations if the importing source file uses an
``@_private(from: "SourceFile.swift") import statement.
rdar://29318654
Currently, the check for whether to serialize parseable interface
arguments doesn't handle the case where a supplementary output file map
is used, preferring only to check if the frontend is passed
`-emit*interface`. Instead, check if the frontend inputs and outputs
contains a parseable interface, and use that to determine if we need to
save args.
This also puts `-module-link-name` in the parseable interface arg list.
