If you had something like:
struct G<T> {
func f<each U>(_: repeat each U) where T == (repeat each U) {}
}
We would fulfill 'each U' from the metadata for 'G<(repeat each U)>',
by taking apart the tuple metadata for `(repeat each U)` and forming
a pack.
However this code path was only intended to kick in for a tuple
conformance witness thunk. In the general case, this optimization
is not correct, because if 'each U' is substituted with a
one-element pack, the generic argument of `G<(repeat each U)>` is
just that one element's metadata, and not a tuple. In fact, we
cannot distinguish the one-element tuple case, because the wrapped
element may itself be a tuple.
The fix is to just split off FulfillmentMap::searchTupleTypeMetadata()
from searchTypeMetadata(), and only call the former when we're in
the specific situation that requires it.
- Fixes https://github.com/swiftlang/swift/issues/78191.
- Fixes rdar://problem/135325886.
In `addAbstractForFulfillments()`, FulfillmentMap (DenseMap) is
iterated to compute all the cache entries. Even the
non-deterministic iteration order doesn't affect the computed cache
entries, the requests to the evaluator during the process will be in
non-deterministic order. This will cause the fine-grain dependency file
generated to be non-deterministic. Use a `MapVector` to make sure
iteration order is deterministic.
rdar://147971980
A lot of the fixes here are adjustments to compensate in the
fulfillment and metadata-path subsystems for the recent pack
substitutions representation change. I think these adjustments
really make the case for why the change was the right one to make:
the code was clearly not considering the possibility of packs
in these positions, and the need to handle packs makes everything
work out much more cleanly.
There's still some work that needs to happen around type packs;
in particular, we're not caching them or fulfilling them as a
whole, and we do have the setup to do that properly now.
Most of the work of this patch is just propagating metadata states
throughout the system, especially local-type-data caching and
metadata-path resolution. It took a few design revisions to get both
DynamicMetadataRequest and MetadataResponse to a shape that felt
right and seemed to make everything easier.
The design is laid out pretty clearly (I hope) in the comments on
DynamicMetadataRequest and MetadataResponse, so I'm not going to
belabor it again here. Instead, I'll list out the work that's still
outstanding:
- I'm sure there are places we're asking for complete metadata where
we could be asking for something weaker.
- I need to actually test the runtime behavior to verify that it's
breaking the cycles it's supposed to, instead of just not regressing
anything else.
- I need to add something to the runtime to actually force all the
generic arguments of a generic type to be complete before reporting
completion. I think we can get away with this for now because all
existing types construct themselves completely on the first request,
but there might be a race condition there if another asks for the
type argument, gets an abstract metadata, and constructs a type with
it without ever needing it to be completed.
- Non-generic resilient types need to be switched over to an IRGen
pattern that supports initialization suspension.
- We should probably space out the MetadataStates so that there's some
space between Abstract and Complete.
- The runtime just calmly sits there, never making progress and
permanently blocking any waiting threads, if you actually form an
unresolvable metadata dependency cycle. It is possible to set up such
a thing in a way that Sema can't diagnose, and we should detect it at
runtime. I've set up some infrastructure so that it should be
straightforward to diagnose this, but I haven't actually implemented
the diagnostic yet.
- It's not clear to me that swift_checkMetadataState is really cheap
enough that it doesn't make sense to use a cache for type-fulfilled
metadata in associated type access functions. Fortunately this is not
ABI-affecting, so we can evaluate it anytime.
- Type layout really seems like a lot of code now that we sometimes
need to call swift_checkMetadataState for generic arguments. Maybe
we can have the runtime do this by marking low bits or something, so
that a TypeLayoutRef is actually either (1) a TypeLayout, (2) a known
layout-complete metadata, or (3) a metadata of unknown state. We could
do that later with a flag, but we'll need to at least future-proof by
allowing the runtime functions to return a MetadataDependency.
A concrete conformance may involve conditional conformances, which are witness
tables that we can access from the original conformance's one. We need to track
metadata and be able to follow it in a metadata path.
not have access to their type arguments at runtime. Use this to
fix the emission of native thunks for imported ObjC-generic
initializers, since they may need to perform bridging.
For now, pseudo-genericity is all-or-nothing, but we may want to
make it apply only to certain type arguments.
Also, clean up some code that was using dead mangling nodes.
We prevent the generic parameters to an ObjC generic from being used in an extension method at the sema level, and we don't want the polymorphic convention to try to generate independent parameters since that would break the method convention. Trick the polymorphic convention by producing "impossible" fulfillments for the generic parameters derivable from an ObjC generic method's self type.
fulfillments, because we don't actually fill them in.
This is a minimal, temporary, speculative fix for the failure
described in rdar://25069637, for which I unfortunately do not have
a minimal test case and which I have been unable to duplicate in
my own testing due to various language restrictions that I hope
to lift given the better (but temporarily reverted) fix.
"minimal" is defined as the set of requirements that would be
passed to a function with the type's generic signature that
takes the thick metadata of the parent type as its only argument.
There are several interesting new features here.
The first is that, when emitting a SILFunction, we're now able to
cache type data according to the full dominance structure of the
original function. For example, if we ask for type metadata, and
we've already computed it in a dominating position, we're now able
to re-use that value; previously, we were limited to only doing this
if the value was from the entry block or the LLVM basic block
matched exactly. Since this tracks the SIL dominance relationship,
things in IRGen which add their own control flow must be careful
to suppress caching within blocks that may not dominate the
fallthrough; this mechanism is currently very crude, but could be
made to allow a limited amount of caching within the
conditionally-executed blocks.
This query is done using a proper dominator tree analysis, even at -O0.
I do not expect that we will frequently need to actually build the
tree, and I expect that the code-size benefits of doing a real
analysis will be significant, especially as we move towards making
more metadata lazily computed.
The second feature is that this adds support for "abstract"
cache entries, which indicate that we know how to derive the metadata
but haven't actually done so. This code isn't yet tested, but
it's going to be the basis of making a lot of things much lazier.
of associated types in protocol witness tables.
We use the global access functions when the result isn't
dependent, and a simple accessor when the result can be cheaply
recovered from the conforming metadata. Otherwise, we add a
cache slot to a private section of the witness table, forcing
an instantiation per conformance. Like generic type metadata,
concrete instantiations of generic conformances are memoized.
There's a fair amount of code in this patch that can't be
dynamically tested at the moment because of the widespread
reliance on recursive expansion of archetypes / dependent
types. That's something we're now theoretically in a position
to change, and as we do so, we'll test more of this code.
This speculatively re-applies 7576a91009,
i.e. reverts commit 11ab3d537f.
We have not been able to duplicate the build failure in
independent testing; it might have been spurious or unrelated.
of associated types in protocol witness tables.
We use the global access functions when the result isn't
dependent, and a simple accessor when the result can be cheaply
recovered from the conforming metadata. Otherwise, we add a
cache slot to a private section of the witness table, forcing
an instantiation per conformance. Like generic type metadata,
concrete instantiations of generic conformances are memoized.
There's a fair amount of code in this patch that can't be
dynamically tested at the moment because of the widespread
reliance on recursive expansion of archetypes / dependent
types. That's something we're now theoretically in a position
to change, and as we do so, we'll test more of this code.
This reverts commit 6528ec2887, i.e.
it reapplies b1e3120a28, with a fix
to unbreak release builds.
This reverts commit b1e3120a28.
Reverting because this patch uses WitnessTableBuilder::PI in NDEBUG code.
That field only exists when NDEBUG is not defined, but now NextCacheIndex, a
field that exists regardless, is being updated based on information from PI.
This problem means that Release builds do not work.
of associated types in protocol witness tables.
We use the global access functions when the result isn't
dependent, and a simple accessor when the result can be cheaply
recovered from the conforming metadata. Otherwise, we add a
cache slot to a private section of the witness table, forcing
an instantiation per conformance. Like generic type metadata,
concrete instantiations of generic conformances are memoized.
There's a fair amount of code in this patch that can't be
dynamically tested at the moment because of the widespread
reliance on recursive expansion of archetypes / dependent
types. That's something we're now theoretically in a position
to change, and as we do so, we'll test more of this code.
