All generic bridgeable types can bridge for all their instantiations now. Removing this ferrets out some now-unnecessary traps that check for unbridgeable parameter types.
In various cases where we had global operators for non-generic
concrete types (such as String + String), move those operators into
the type. This should not affect the sources, but makes the exposition
of the library cleaner.
Plus, it's a good test for the compiler, which uncovered a few issues
where the compiler was coupled with the library.
The C++ code was very fragile in terms of ABI dependencies and broke
when the standard library was built with -enable-resilience.
The actual reason it broke is that case numbering changes when
resilience is enabled, but instead of messing with that, it seemed
more logical to rewrite this routine in Swift instead, to avoid
ABI dependencies altogether.
This requires using the "shadow protocol" trick to call NSNumber
methods, since we cannot import NSNumber from the stdlib.
Make _MagicMirrorData @_fixed_layout, but not the concrete mirror
implementations.
To make the calling convention work in resilient builds, make the
runtime entry points into top-level functions that take a
_MagicMirrorData, instead of adding @_silgen_name attributes on
methods.
This involves changing the convention on the 'owner' parameter
from +0 to +1.
Also, there was a memory leak in the old enum code that I noticed
by inspection. We would copy the enum value into a box, strip
the enum tag bits, take the box contents but never free the box
itself.
The fix isn't very satisfying either -- we have to modify the
source enum in-place to strip tag bits, then we copy it into a
box having the right payload type, and add the tag bits back.
At least this way, we can free the box after.
Similarly to how we've always handled parameter types, we
now recursively expand tuples in result types and separately
determine a result convention for each result.
The most important code-generation change here is that
indirect results are now returned separately from each
other and from any direct results. It is generally far
better, when receiving an indirect result, to receive it
as an independent result; the caller is much more likely
to be able to directly receive the result in the address
they want to initialize, rather than having to receive it
in temporary memory and then copy parts of it into the
target.
The most important conceptual change here that clients and
producers of SIL must be aware of is the new distinction
between a SILFunctionType's *parameters* and its *argument
list*. The former is just the formal parameters, derived
purely from the parameter types of the original function;
indirect results are no longer in this list. The latter
includes the indirect result arguments; as always, all
the indirect results strictly precede the parameters.
Apply instructions and entry block arguments follow the
argument list, not the parameter list.
A relatively minor change is that there can now be multiple
direct results, each with its own result convention.
This is a minor change because I've chosen to leave
return instructions as taking a single operand and
apply instructions as producing a single result; when
the type describes multiple results, they are implicitly
bound up in a tuple. It might make sense to split these
up and allow e.g. return instructions to take a list
of operands; however, it's not clear what to do on the
caller side, and this would be a major change that can
be separated out from this already over-large patch.
Unsurprisingly, the most invasive changes here are in
SILGen; this requires substantial reworking of both call
emission and reabstraction. It also proved important
to switch several SILGen operations over to work with
RValue instead of ManagedValue, since otherwise they
would be forced to spuriously "implode" buffers.
* Switch to calling `putchar_unlocked()` instead of `putchar()` for
actual printing. We're already locking stdout with `flockfile()`, so
there's no need for the redundant lock that `putchar()` uses.
* Add an explicit lock to the output stream in `dump()`. This means the
entire dump is printed with the lock held, which will prevent the
output of `dump()` from mixing with prints on other threads.
* Use `_debugPrint_unlocked()` instead of `debugPrint()` in
`_adHocPrint()`. The output stream is already locked while this
function is executing. Rename the function to `_adHocPrint_unlocked()`
to make this explicit.
* Use `targetStream.write()` and `_print_unlocked()` instead of
`print()` in `_dumpObject()`. This removes the redundant locking, and
also eliminates the creation of intermediate strings. Rename the
function to `_dumpObject_unlocked()` to make this explicit.
* Use `targetStream.write()`, `_print_unlocked()`, and
`_debugPrint_unlocked()` in `_dumpSuperclass()`. This removes the
redundant locking, and also eliminates the creation of intermediate
strings. Rename the function to `_dumpSuperclass_unlocked()` to make
this explicit.
* Use `_debugPrint_unlocked()` instead of `debugPrint()` in
`String.init(reflecting:)`. This shouldn't really make much of a
difference but it matches the usage of `_print_unlocked()` in
`String.init(_:)`.
The net result is that all printing is still covered under locks like
before, but stdout is never recursively locked. This should result in
slightly faster printing. In addition, `dump()` is now covered under a
single lock so it can't mix its output with prints from other threads.
This pull request broke the following tests on several build configurations
(eg --preset=buildbot,tools=RA,stdlib=DA)
1_stdlib/Reflection.swift
1_stdlib/ReflectionHashing.swift
1_stdlib/UnsafePointer.swift.gyb
This reverts commit c223a3bf06, reversing
changes made to 5c2bb09b09.
Changes:
- Reverted commit reverting original SR-88 commit
- Removed mirror children helper collections and related code
- Rewrote some tests to keep them working properly
- Wrote two more tests for the three pointer APIs to ensure no crashes if created using a value > Int64.max
This reverts commit 8917eb0e5a.
