Slava Pestov
b2bc2c72ec
AST: Introduce PackElementType
2023-05-25 11:17:30 -04:00
John McCall
d25a8aec8b
Add explicit lowering for value packs and pack expansions.
...
- SILPackType carries whether the elements are stored directly
in the pack, which we're not currently using in the lowering,
but it's probably something we'll want in the final ABI.
Having this also makes it clear that we're doing the right
thing with substitution and element lowering. I also toyed
with making this a scalar type, which made it necessary in
various places, although eventually I pulled back to the
design where we always use packs as addresses.
- Pack boundaries are a core ABI concept, so the lowering has
to wrap parameter pack expansions up as packs. There are huge
unimplemented holes here where the abstraction pattern will
need to tell us how many elements to gather into the pack,
but a naive approach is good enough to get things off the
ground.
- Pack conventions are related to the existing parameter and
result conventions, but they're different on enough grounds
that they deserve to be separated.
2023-01-29 03:29:06 -05:00
Slava Pestov
7d8f3e6b63
AST: Change return type of Requirement::subst() to Requirement
...
Instead of returning None, let callers check hasError() if they need to.
Fixes rdar://problem/98565072.
2022-08-12 14:03:57 -04:00
John McCall
e60095e36a
Remove an unnecessary assertion.
...
I did check that the code doesn't actually rely on not seeing a
type parameter.
rdar://96571508
2022-07-13 16:17:27 -04:00
Michael Gottesman
198b974622
Merge pull request #59859 from gottesmm/pr-276c4e4220a25490659a285e8b94a36bd28ffede
...
[no-implicit-copy] Rename SILMoveOnlyType -> SILMoveOnlyWrappedType.
2022-07-01 23:17:43 -07:00
Michael Gottesman
8f3fe63fed
[no-implicit-copy] Rename SILMoveOnlyType -> SILMoveOnlyWrappedType.
...
Since I am beginning to prepare for adding real move only types to the language,
I am renaming everything that has to do with copyable types "move only wrapped"
values instead of move only. The hope is this reduces/prevents any confusion in
between the two.
2022-07-01 17:26:13 -07:00
John McCall
175f74d38f
Implement symbolic demangling for extended existential metadata
...
Fixes rdar://96268090.
2022-07-01 11:21:53 -04:00
Michael Gottesman
c66061a783
[moveOnly] Add a new SIL specific AST type called SILMoveOnly.
...
It can only be used in SIL contexts to express that a wrapped type is
MoveOnly.
2022-06-03 11:27:16 -07:00
Slava Pestov
4641ffc5ea
AST: Refactor existential generalization to use buildGenericSignature() instead of GenericSignatureBuilder
...
The GenericSignatureBuilder is no longer used for anything by default, and we
don't want to create instances of it directly.
Instead, use the buildGenericSignature() utility method, which is backed by
the Requirement Machine.
2022-05-10 01:13:18 -04:00
Josh Soref
81d3ad76ac
Spelling ast ( #42463 )
...
* spelling: accessor
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: accommodates
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: argument
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: associated
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: availability
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: available
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: belongs
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: bookkeeping
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: building
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: clazz
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: clonable
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: closure
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: concatenated
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: conformance
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: context
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: conversion
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: correspondence
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: declarations
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: declared
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: defining
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: delayed
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: dependency
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: deployed
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: descendants
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: diagnose
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: diagnostic
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: equitable
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: evaluation
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: exclusivity
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: existence
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: existential
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: explicit
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: expressed
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: for
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: foreign
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: function
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: identifier
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: implicit
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: indices
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: information
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: instance
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: interchangeable
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: interface
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: introduced
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: invalid
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: kind-in
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: least
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: library
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: location
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: namespace
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: necessary
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: nonexistent
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: not
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: number
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: obtains
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: occurs
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: opaque
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: overridden
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: parameter
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: precede
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: preceding
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: property
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: protocol
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: qualified
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: recognized
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: recursively
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: references
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: relaxing
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: represented
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: request
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: requirement
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: requirements
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: retrieve
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: returned
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: satisfied
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: satisfy
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: scanner
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: siblings
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: simplified
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: something
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: source
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: specializations
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: specially
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: statement
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: stripped
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: structure
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: substitution
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: the
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: transform
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: transformed
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: transitively
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: transparent
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: typecheck
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: unknown
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: unlabeled
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: unqualified
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: whether
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: with
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
* spelling: scanner
Signed-off-by: Josh Soref <jsoref@users.noreply.github.com >
Co-authored-by: Josh Soref <jsoref@users.noreply.github.com >
2022-04-21 12:57:16 -07:00
John McCall
124fe99c2a
Implement the existential generalization algorithm.
...
The basic concept here was previously laid out in the metadata
system commits adding ExtendedExistentialTypeShape, but to recap,
we want to produce (for any existential type, but we're really
only going to use it for constrained or otherwise generalized
existentials) a "shape" signature and type that will be agreed
upon by all possible abstractions of the type. In this
signature+type pair, any substitutable position in the original
type is abstracted as a parameter in the signature, which the
original type is a concrete application of. This permits
agreement on the type in the face of runtime generic execution.
I have made every effort to support protocol compositions in
the code I've written, but of course until you're allowed to
actually write them, that's all untested.
I have not yet implemented signature minimization of the
generalization signature. I think this will be necessary in order
to ensure that there are no non-redundant parameters, especially
for compositions, which is something the runtime currently assumes
and is better for code size anyway. But it can wait a few days,
I think.
2022-04-11 22:15:21 -04:00