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.
Match the new SILGen pattern, where only the box parameter is partially applied to the closure, and the address of the value is projected on the callee side.
And include some supplementary mangling changes:
- Give the first generic param (depth=0, index=0) a single character mangling. Even after removing the self type from method declaration types, 'Self' still shows up very frequently in protocol requirement signatures.
- Fix the mangling of generic parameter counts to elide the count when there's only one parameter at the starting depth of the mangling.
Together these carve another 154KB out of a debug standard library. There's some awkwardness in demangled strings that I'll clean up in subsequent commits; since decl types now only mangle the number of generic params at their own depth, it's context-dependent what depths those represent, which we get wrong now. Currying markers are also wrong, but since free function currying is going away, we can mangle the partial application thunks in different ways.
Swift SVN r32896
Canonical dependent member types are always based from a generic parameter, so we can use a more optimal mangling that assumes this. We can also introduce substitutions for AssociatedTypeDecls, and when a generic parameter in a signature is constrained by a single protocol, we can leave that protocol qualification out of the unsubstituted associated type mangling. These optimizations together shrink the standard library by 117KB, and bring the length of the longest Swift symbol in the stdlib down from 578 to 334 characters, shorter than the longest C++ symbol in the stdlib.
Swift SVN r32786
A microoptimization; since the module is likely to come up often in the subsequent mangling, we want to make it more likely to get the coveted S_ substitution.
Swift SVN r32784
'Ss' appears in manglings tens of thousands of times in the standard library and is also incredibly frequent in other modules. This alone is enough to shrink the standard library by 59KB.
Swift SVN r32409
This is more resilient, since we want to be able to add more information behind the address point of type objects. The start of the metadata object is now an internal "full metadata" symbol.
Note that we can't do this for known opaque metadata from the C++ runtime, since clang doesn't have a good way to emit offset symbol aliases, so for non-nominal metadata objects we still emit an adjustment inline. We also aren't able to generate references to aliases within the same module due to an MC bug with alias refs on i386 and armv7 (rdar://problem/22450593).
Swift SVN r31523
This is more resilient, since we want to be able to add more information behind the address point of type objects, and also makes IR a lot less cluttered. The start of the metadata object is now an internal "full metadata" symbol.
Note that we can't do this for known opaque metadata from the C++ runtime, since clang doesn't have a good way to emit offset symbol aliases, so for non-nominal metadata objects we still emit an adjustment inline.
Swift SVN r31515
The demangler recently regressed to not printing any context
names, including nominal type contexts. This means that symbols
like Optional.init where only printed as init. Continue printing
contexts but not modules (per the original simplified demangling
design).
rdar://problem/19312992
Swift SVN r31066
Break up "Simplified" demangling mode (shortened demangled descriptions
for the sake of displaying in UI with small areas) into more
fine-grained options instead of an opaque "Simplified" option and
provide a static preset of options for displaying stack traces in
Xcode UI and other tools, for example.
- Don't print unmangled suffixes
- Don't print module names
- Shorten various generic specialization descriptions as just
"specialized"
- Don't display long protocol conformances
- Truncate where clauses
- Don't display so-called "entity" types
- Shorten "partial apply *"
- Shorten thunk phrases
- Shorten value witness phrases
- Truncate archetype references
rdar://problem/21753651
Swift SVN r30247
This enables dead argument elimination to be paired with @owned -> @guaranteed
optimization. It has the additional advantage of allowing us to potentially
eliminate additional retains, releases since the fact that the use is dead
implies that the lifetime of the value no longer needs to be live across the
function call.
Since dead argument elimination can be composed with @owned -> @guaranteed, I
had to modify the mangler, remangler, demangler, to be able to handle a mangling
that combines the two.
I just saw noise in the perf test suite.
rdar://21114206
Swift SVN r29966
Constrained and protocol extensions should always include the extension context in their mangling, since they are never equivalent to definitions in the original type context. Have them use the extension mangling, and include the generic signature of the extension in its mangling, which is necessary to disambiguate properties and other definitions that are defined with the same name and type in differently constrained extensions. Fixes rdar://problem/21027215.
Swift SVN r29209
As part of this, I've made the demangler base the colon-vs.-not
decision on the entity kind instead of assuming that anything
with a function type must be a function. It also looks through
new-style generics when it didn't before.
Swift SVN r28814
To support UI applications displaying demangled names in a limited
amount of screen space, provide a new SwiftDemangle API and Demangler
option to do the following:
- Skip all module name prefixes when printing contexts
- Don't print implicit self/metatype parameters when printing
function types
Add a '-simplified' flag to swift-demangle to support testing at the
command line.
Swift SVN r28727
Mangle the generic signature and interface type of the conforming type signature. Fix the demangler to reset the generic context between mangling a witness's conformance and requirement; they have independent generic contexts, and the bookkeeping for generic signatures notes the discrepancy.
Swift SVN r28377
Share the logic we use to map archetype depth-index pairs to friendly unique names like 'A', 'B', so that demangle generic signatures are still somewhat readable, and so that archetype references into outer contexts with interface type manglings still make sense. Change the remangler to mangle archetypes and dependent generic params using nested index nodes instead of trying to parse the depth and index from the arbitrary names we give them.
Swift SVN r28343
Single generic parameters are common, as are a lack of requirements, so tweak things so that generic parameter counts are mangled as (count - 1), with a special mangling for zero, and give a single generic parameter at depth zero the empty-string mangling. Most requirements are protocol constraints, so use a better mangling for them that doesn't require the 'P..._' wrapping of the general type mangling. On the other hand, dependent member types ought to mangle in the protocol of the associated type, which adds some length, but isn't too bad since the protocol will almost definitely have a substitution introduced by a preceding protocol constraint.
Swift SVN r28296
This lets us disambiguate the symbols for static and instance properties, and enables us to eventually leave the useless "self" type mangling out of method symbols. Fixes rdar://19012022 and dupes thereof, including crasher #1341.
Swift SVN r25111
To get this to work, delay some "cleanup" work in the
demangler. For example, we now preserve in the tree
whether something was mangled as an allocating
initializer, and we only special-case the class vs.
non-class cases in the pretty printer.
Also fixes a number of remangling bugs, of course.
Swift SVN r24534
Currently this only handles top-level nominal types. We're just trying to
emulate what the debugger does when it needs to go from a mangled name to
an AST node, so it's okay that the cases handled here are very restricted.
We just want to make sure that the debugger is /able/ to do what it needs
to do.
This does not yet handle nested (non-top-level) values; that will require
changes to DeclContext::lookupQualified.
Part of rdar://problem/17632175
Swift SVN r21690
