Three minor changes:
* Remove unneeded stdint.h inclusion
* Use __FILE_NAME__ instead of __FILE__ to reduce code size
* Write location as "file:line" for better compatibility with existing tools
The handling of multi-basic-block control flow in `defer` blocks looks like it
was left incomplete and completely untested; I fixed a few obvious problems but
it still completely lacks any analysis of conditional reinitializations. For now,
change it to treat attempted reinitializations as uses-after-consumes so we raise
reliable errors now instead of emitting code that causes memory corruption at
runtime. Fixes rdar://129303198.
TLDR: This makes it so that we always can parse sending/transferring but changes
the semantic language effects to be keyed on RegionBasedIsolation instead.
----
The key thing that makes this all work is that I changed all of the "special"
semantic changes originally triggered on *ArgsAndResults to now be triggered
based on RegionBasedIsolation being enabled. This makes a lot of sense since we
want these semantic changes specifically to be combined with the checkers that
RegionBasedIsolation turns on. As a result, even though this causes these two
features to always be enabled, we just parse it but we do not use it for
anything semantically.
rdar://128961672
Updates swift-symbolgraph-extract to parse "-cxx-interoperability-mode"
flags and update the underlying compiler invocation. This fixes a bug
where were are unable to extract the symbol graph from swiftmodules with
transitive cxx modules because we parsed cxx headers as c headers.
rdar://128888548 (Add support for parsing cxx headers)
Some compilers have the NoncopyableGenerics feature enabled via
interesting mechanisms but do not have ConformanceSuppression. To
support such compilers, the NoncopyableGenerics feature must appear
before ConformanceSuppression in the list of features. Otherwise, when
parsing the portion of the swiftinterface corresponding to an entity
which involves both features, the first check will be for
NoncopyableGenerics (which that old compiler has) and the code inside
will involve ConformanceSuppression (which that old compiler does not
have).
rdar://128611158
When caching is enabled, using swift style diagnotics can lead to
crashes due to some uninitialized variables. Even more, the swift style
diagnostics is not going to render the same when replay from the cache
since the currect caching diagnostics processor is not capture all
information that is needed to render diagnostis from SwiftSyntax.
Temprarily using llvm style for caching builds.
rdar://127530204
A few things:
1. Internally except for in the parser and the clang importer, we only represent
'sending'. This means that it will be easy to remove 'transferring' once enough
time has passed.
2. I included a warning that suggested to the user to change 'transferring' ->
'sending'.
3. I duplicated the parsing diagnostics for 'sending' so both will still get
different sets of diagnostics for parsing issues... but anywhere below parsing,
I have just changed 'transferring' to 'sending' since transferring isn't
represented at those lower levels.
4. Since SendingArgsAndResults is always enabled when TransferringArgsAndResults
is enabled (NOTE not vis-a-versa), we know that we can always parse sending. So
we import "transferring" as "sending". This means that even if one marks a
function with "transferring", the compiler will guard it behind a
SendingArgsAndResults -D flag and in the imported header print out sending.
rdar://128216574
This adds three new assertion macros:
* `ASSERT` - always compiled in, always checked
* `CONDITIONAL_ASSERT` - always compiled in, checked whenever the `-compiler-assertions` flag is provided
* `DEBUG_ASSERT` - only compiled into debug builds, always checked when compiled in (functionally the same as Standard C `assert`)
The new `-compiler-assertions` flag is recognized by both `swift-frontend` and
`swiftc`.
The goal is to eventually replace every use of `assert` in the compiler with one of the above:
* Most assertions will use `ASSERT` (most assertions should always be present and checked, even in release builds)
* Expensive assertions can use `CONDITIONAL_ASSERT` to be suppressed by default
* A few very expensive and/or brittle assertions can use `DEBUG_ASSERT` to be compiled out of release builds
This should:
* Improve quality by catching errors earlier,
* Accelerate compiler triage and debugging by providing more accurate crash dumps by default, and
* Allow compiler engineers and end users alike to add `-compiler-assertions` to get more accurate failure diagnostics with any compiler
