While this is not yet used, Compilation will be able to adopt this for easier testing of driver-level task execution (instead of forcing all driver-level tests to invoke the frontend, linker, etc.).
Swift SVN r12349
For the Unix implementation, this returns NumberOfParallelTasks (unless that value is 0, in which case it returns the default value of 1).
For the Default implementation, this always returns 1, since it does not support parallelism.
Swift SVN r12348
Instead of always calling waitpid() and looking for the termination of a task, call poll() and watch for events on the pipes which are used for buffering subtask output. When we receive a POLLIN or POLLPRI event, ask the corresponding Task to read from the pipe; when we receive a POLLHUP or POLLERR event, wait for the subtask to finish and then ask the Task to perform its post-execution tasks.
Now that the Unix implementation properly handles large amounts of subtask output, re-enable the Unix implementation (if supported).
This fixes <rdar://problem/15795234>.
Swift SVN r12293
When we're using Objective-C's memory allocation, emit .cxx_construct
methods whenever we have instance variables with in-class
initializers. Presently, these methods are just empty stubs.
Swift SVN r12211
The Objective-C runtime executes the .cxx_destruct method after the
last -dealloc has executed when destroying an object, allowing the
instance variables to remain live even after the subclass's
destructor/-dealloc has executed, which is important for memory
safety. This fixes the majority of <rdar://problem/15136592>.
Note that IRGenModule::getAddrOfIVarDestroyer() contains an egregious
hack to find the ivar destructor SIL function via a linear
search. We need a better way to find SIL functions that we know exist,
because LinkEntity does not suffice.
Swift SVN r12206
The current implementation does not read() from each Task's pipe until execution
finishes. This causes subtasks to block in write() if the pipe is full, which
only happens if there is a large amount of output (such as in -dump-ast mode).
Fixing output buffering and reenabling the Unix TaskQueue implementation is
tracked by <rdar://problem/15795234>.
Swift SVN r12135
Due to the nature of this class, there are two implementations of TaskQueue:
a Unix-specific implementation which supports both parallel execution and output
buffering, and a default implementation which supports neither of these features.
(The default implementation uses the functions from llvm/Support/Program.h for execution.)
TaskQueue allows clients to provide a TaskBeganCallback and a TaskFinishedCallback,
each of which will be called when a new task begins or when a task finishes execution,
respectively. Clients may add tasks to the TaskQueue in either of these callbacks,
and clients can stop further execution by returning TaskFinishedResponse::StopExecution
from the TaskFinishedCallback.
Swift SVN r12059
Revert "add a hackaround for rdar://15753317. I don't unerstand the code enough to tell if this is the right fix."
This reverts commit 416983b0734dde6979c98971948068c7a157e336.
Swift SVN r11942
Not NFC: changes some of the demangling tree schemata (in
interest of regularity and expressivity) and, incidentally,
improves support for initializer manglings.
The main schema change is that Path is now gone, and instead
the first child of an entity is its context. All contexts
are now labelled with their kind (variable/function/etc.)
rather than that being much more obscure or missing.
A secondary change is that the top-level node is now a
single node with the <global> and all its attributes as
children, rather than being weirdly sibling-linked.
Also, the representation has changed so that nodes link
only to their children, not to siblings or parents, which
means that it is no longer necessary to clone nodes when
replacing substitutions.
Finally, dump/print was brought in from swift-demangler
and made part of the core API for debugging purposes.
Swift SVN r11470
- Added a couple of new targets:
- libswiftDriver, which contains most of the driver implementation
- swift_driver, which produces the actual executable
- Added centralized version information into libswiftBasic.
- Added a new "Driver Design & Internals" document, which currently describes
the high-level design of the Swift driver.
- Implemented an early version of the functionality of the driver, including
versions of the Parse, Pipeline, Bind, Translate, and Execute driver stages.
Parse, Pipeline, and Bind are largely implemented; Translate and Execute are
early placeholders. (Translate produces "swift_driver --version" and "ld -v"
commands, while Execute performs all subtasks sequentially, rather than in
parallel.)
This is just the starting point for the Swift driver. Tests for the existing
behavior are forthcoming.
Swift SVN r10933
iVars of a given type should be unique by name and so in certain cases, their types will not actually add information to a consumer of the demangled string
This mode is mostly useful for LLDB. When searching for a field offset by name in a generic scenario, the search can succeed or fail depending on whether archetypes are
obtained through the Demangler ("A","B","C") or through the module (which will vend proper archetype names when poked the right way)
This mode removes that ambiguity
Swift SVN r10761
decl context of the type alias.
This implements <rdar://problem/15290346> "typealias sugar needs to be
mangled into debug info mangled type names".
Swift SVN r10749
This "greedy" demangling should not be necessary anymore once <rdar://problem/15444866> goes in
As of now, this extension should help LLDB support generics in functions
Swift SVN r10493
The underlying mechanism uses darwin's libcache library (if not building on darwin there is a default implementation).
This was originally introduced in SourceKit, but we are moving it to libswiftBasic so it can be utilized by libswiftIDE
for caching code-completion results.
Swift SVN r9610
This checkin extends the Demangler to allow printing sugar on demangled types
Namely, it introduces a DemanglerOptions class with just one field SynthesizeSugarOnTypes which has the same functionality as PrintOptions::SynthesizeSugarOnTypes
This changes outputs like _TtGSqC5sugar7MyClass_ ---> swift.Optional<sugar.MyClass> into _TtGSqC5sugar7MyClass_ ---> sugar.MyClass?
By default this flag is false, so that existing clients of the Demangler API do not break
OTOH, the command-line tool swift-demangle flags sets the flag to true, unless the -no-sugar option is passed on the command-line
Test cases included
Swift SVN r9502
