This allows making global addressors fragile (They use globalinit_{token,func} for initialization of globals).
It has no noticable performance impact on our benchmarks, but it removes an ugly hack which explicitly
prevented addressors from being fragile.
Swift SVN r22795
Change the SILLocations for enum element dispatches to make the line table
more consistent. emitEnumElementDispatch may be invoked several times so
it should use the location of the first pattern rather than TheSwitch.
Swift SVN r22782
This reverts commit r22739 because it broke an unreachable code test
and is blocking the bots.
The problem can be reproduced with:
swift -frontend -disable-objc-attr-requires-foundation-module -emit-sil test/SILPasses/unreachable_code.swift -o /dev/null -verify
Swift SVN r22750
more consistent. emitEnumElementDispatch may be invoked several times so
it should use the location of the first pattern rather than TheSwitch.
Swift SVN r22739
as +0 when the client requests it. This avoids emitting some pointless
retain/releases.
This finishes up:
<rdar://problem/17207456> Unable to access dynamicType of an object in a class initializer that isn't done
Swift SVN r22736
emitting LoadExprs (unless they have a side effect that must be
emitted). This is good for -O0 code quality in general, but is important
to avoid materializing 'self' values in initializers that semantically
don't have to happen.
This is progress towards fixing:
<rdar://problem/17207456> Unable to access dynamicType of an object in a class initializer that isn't done
enabling the struct cases.
Swift SVN r22735
And in IRGen, fix up pointer cast codegen to work with types that explode to single scalars while being stored in named structs. Fixes rdar://problem/18604262
Swift SVN r22671
with user code and that the boilerplate is counted towards the prologue.
<rdar://problem/18563763> Setting a breakpoint on "main" in a Swift program doesn't stop at user code
Swift SVN r22611
layouts. Introduce new SIL instructions to initialize
and open existential metatype values.
Don't actually, y'know, lift any of the restriction on
existential metatypes; just pointlessly burn extra
memory storing them.
Swift SVN r22592
This commit adds SILGen for references to potentially unavailable global
variables. We emit .None if the global variable is not available and
.Some(rvalue) if the global variable is available. I will add SILGen
for references to other kinds of potentially unavailable declarations in a
later commit.
Swift SVN r22556
Use init_enum_data_addr and inject_enum_addr to construct optional values instead of the injection intrinsics, further simplifying -Onone IR. This not only avoids a call but also allows the frontend to emit optional payloads in-place in more cases, eliminating a lot of stack traffic.
Swift SVN r22549
When we've already established that the optional has a value, using unchecked_take_enum_data_addr to directly extract the enum payload is sufficient and avoids a redundant call and check at -Onone. Keep using the _getOptionalValue stdlib function for checked optional wrapping operations such as "x!", so that the stdlib can remain in control of trap handling policy.
The test/SIL/Serialization failures on the bot seem to be happening sporadically independent of this patch, and I can't reproduce failures in any configuration I've tried.
Swift SVN r22537
When we've already established that the optional has a value, using unchecked_take_enum_data_addr to directly extract the enum payload is sufficient and avoids a redundant call and check at -Onone. Keep using the _getOptionalValue stdlib function for checked optional wrapping operations such as "x!", so that the stdlib can remain in control of trap handling policy.
Swift SVN r22533
Eliminate the intermediate top_level_code function. Now that SIL is expressive enough to express a "main" function, there's no reason for it, and this eliminates a bunch of mystery code in IRGen to thunk from main to top_level_code by reaching for hardcoded symbol names. Demystify the special code for setting up C_ARGC and C_ARGV by having SILGen look for a transparent "_didEnterMain" hook in the stdlib and emit a call to it.
Swift SVN r22525
We can't expose the lazy initialization token across module boundaries because that might lead to it being duplicated, which would be bad, so if we were ever going to make them fragile, it would have to happen after a guaranteed globalopt pass. Fixes a crash that was blocking DaveA's work; his incoming patches should exercise this.
Swift SVN r22509
This is controlled by a new isWholeModule() attribute in SILModule.
It gives about 9% code size reduction on the benchmark executables.
For test-suite reasons it is currently not done for the stdlib.
Swift SVN r22491
Stub initializers don't get serialized, so this fixes a vtable layout inconsistency when a method of a subclass of a subclass of NSObject is accessed. Fixes rdar://problem/18498385.
Swift SVN r22480
llvm::Optional lives in "llvm/ADT/Optional.h". Like Clang, we can get
Optional in the 'swift' namespace by including "swift/Basic/LLVM.h".
We're now fully switched over to llvm::Optional!
Swift SVN r22477
In preparation for the switch to llvm::Optional, which doesn't have this.
I suggested it, but got some pushback from David Blaikie, which is
understandable because implicit conversions are dangerous.
std::experimental::optional also doesn't have this, and we don't use it
too much. The only cost is being slightly more explicit when relying on
conversions to give us a T to then wrap in an Optional.
Swift SVN r22472
In preparation for the switch to llvm::Optional, which doesn't have a 'cache'
method. Given how long we spent bikeshedding over the name and how few places
we ended up using it, I didn't feel like trying to push it through on the
LLVM side.
Swift SVN r22471
Use open_existential/witness_method instead of project_existential/protocol_method for property accessor lookups to, eliminating the remaining uses of protocol_method.
Swift SVN r22443
This simplifies the code generation path for existential methods by allowing it to shared more code with the generic case, (It'll be even simpler when Sema opens the existentials for SILGen...) turning protocol_method lookups into open_existential + witness_method sequences. In this patch, we handle normal generic method lookups, but property accesses still go through protocol_method.
Swift SVN r22437
This fixes bugs when we call the result of generic methods that return functions; we were advancing through orig types of the signature but not formal ones, causing us to fall out of sync when we exhausted curry levels of the original function and started calling function-typed results. Fixes rdar://problem/18143223 and rdar://problem/18495979, maybe others too.
Swift SVN r22418
properties.
The main design change here is that, rather than having
purportedly orthogonal storage kinds and has-addressor
bits, I've merged them into an exhaustive enum of the
possibilities. I've also split the observing storage kind
into stored-observing and inherited-observing cases, which
is possible to do in the parser because the latter are
always marked 'override' and the former aren't. This
should lead to much better consideration for inheriting
observers, which were otherwise very easy to forget about.
It also gives us much better recovery when override checking
fails before we can identify the overridden declaration;
previously, we would end up spuriously considering the
override to be a stored property despite the user's
clearly expressed intent.
Swift SVN r22381
This patch adds SILGen for the API availability (#os(...)) construct. To do so,
it (1) adds version range information to the AvailabilityQueryExpr AST in Sema,
during type refinement context construction; and (2) uses that version range,
during SILGen, to emit a call to the standard library's
_stdlib_isOSVersionAtLeast function.
Swift SVN r22348
"self" needs to be materialized in these cases. We were handling methods correctly, but not property accesses. Fixes rdar://problem/18454204.
Swift SVN r22309
semantically valid way.
Previously, this decision algorithm was repeated in a
bunch of different places, and it was usually expressed
in terms of whether the decl declared any accessor
functions. There are, however, multiple reasons why a
decl might provide accessor functions that don't require
it to be accessed through them; for example, we
generate trivial accessors for a stored property that
satisfies a protocol requirement, but non-protocol
uses of the property do not need to use them.
As part of this, and in preparation for allowing
get/mutableAddressor combinations, I've gone ahead and
made l-value emission use-sensitive. This happens to
also optimize loads from observing properties backed
by storage.
rdar://18465527
Swift SVN r22298
