aboutsummaryrefslogtreecommitdiff
path: root/src/codegen.cpp
AgeCommit message (Collapse)Author
2019-08-16fix and test case for returning from suspend blockAndrew Kelley
See #3063
2019-08-16codegen: LLVMConstSub instead of LLVMBuildSub in one placeAndrew Kelley
2019-08-15fix crash with sometimes type not being resolvedAndrew Kelley
2019-08-15add assertion about control flow to fix gcc warningAndrew Kelley
2019-08-15fix error return traces for async calls of blocking functionsAndrew Kelley
2019-08-15remove `cancel`Andrew Kelley
2019-08-14codegen for async call of blocking functionAndrew Kelley
2019-08-13avoid the word "coroutine", they're "async functions"Andrew Kelley
2019-08-13alignment of structs no longer depends on LLVMAndrew Kelley
fixes async function tests in optimized builds
2019-08-13Merge remote-tracking branch 'origin/master' into rewrite-coroutinesAndrew Kelley
2019-08-13flip the order of fields in error unionsAndrew Kelley
to prepare for fixing u128 alignment issues
2019-08-11fix no-longer-correct `nonnull` attribute on merge err ret traces fnAndrew Kelley
2019-08-11fix cancel invoking branch on undefined memoryAndrew Kelley
2019-08-11all tests passingAndrew Kelley
2019-08-11fix canceling async functions which have error return tracingAndrew Kelley
2019-08-10fix returning a const error from async functionAndrew Kelley
2019-08-10fix try in an async function with error union and non-zero-bit payloadAndrew Kelley
2019-08-09fix combining try with errdefer cancelAndrew Kelley
2019-08-09fix cancel inside an errdeferAndrew Kelley
2019-08-08fix passing string literals to async functionsAndrew Kelley
2019-08-08async functions in single threaded mode do not use atomic opsAndrew Kelley
2019-08-08more debuggable safety for awaiting twiceAndrew Kelley
2019-08-08cancel, defer, errdefer all working as intended nowAndrew Kelley
2019-08-07modify header precedence for zig cc, resolves intrinsics issues (#3027)Ryan Saunderson
2019-08-07clean up the bitcasting of awaiter fn ptrAndrew Kelley
2019-08-07implement cancelAndrew Kelley
all behavior tests passing in this branch
2019-08-06fix awaiting when result type is a structAndrew Kelley
2019-08-06error return trace across suspend pointsAndrew Kelley
2019-08-06passing the error return trace async function testAndrew Kelley
2019-08-06improve async function semanticsAndrew Kelley
* add safety panic for resuming a function which is returning, pending an await * remove IrInstructionResultPtr * add IrInstructionReturnBegin. This does the early return in async functions; does nothing in normal functions. * `await` gets a result location * `analyze_fn_async` will call `analyze_fn_body` if necessary. * async function frames have a result pointer field for themselves to access and one for the awaiter to supply before the atomic rmw. when returning, async functions copy the result to the awaiter result pointer, if it is non-null. * async function frames have a stack trace pointer which is supplied by the awaiter before the atomicrmw. Later in the frame is a stack trace struct and addresses, which is used for its own calls and awaits. * when awaiting an async function, if an early return occurred, the awaiter tail resumes the frame. * when an async function returns, early return does a suspend (in IrInstructionReturnBegin) before copying the error return trace data, result, and running the defers. After the last defer runs, the frame will no longer be accessed. * proper acquire/release atomic ordering attributes in async functions.
2019-08-05async functions have error return traces where appropriateAndrew Kelley
however the traces are not merged on `await` or async function calls yet. When an async function has an error set or error union as its return type, it has a `StackTrace` before the args in the frame, so that it is accessible from `anyframe->T` awaiters. However when it does not have an errorable return type, but it does call or await an errorable, it has a stack trace just before the locals. This way when doing an `@asyncCall` on an async function pointer, it can populate the args (which are after the `StackTrace`) because it knows the offset of the args based only on the return type. This sort of matches normal functions, where a stack trace pointer could be supplied by a parameter, or it could be supplied by the stack of the function, depending on whether the function itself is errorable.
2019-08-05delete IrInstructionMarkErrRetTracePtrAndrew Kelley
this IR instruction is no longer needed
2019-08-05support async functions with inferred error setsAndrew Kelley
2019-08-04refactor logic for determining if there is a frame pointerAndrew Kelley
2019-08-04simpler, less memory intensive suspend/resume implementationAndrew Kelley
2019-08-04move state from ZigFn to CodeGen to save memoryAndrew Kelley
2019-08-04suspension points inside branching control flowAndrew Kelley
2019-08-03implement `@asyncCall` which supports async function pointersAndrew Kelley
2019-08-03add runtime safety for resuming an awaiting functionAndrew Kelley
2019-08-02implement async functions returning structsAndrew Kelley
2019-08-02Merge remote-tracking branch 'origin/master' into rewrite-coroutinesAndrew Kelley
2019-08-02Merge branch 'master' into rewrite-coroutinesAndrew Kelley
2019-08-02fix cmpxchg with discarded resultAndrew Kelley
2019-08-02implement async await and returnAndrew Kelley
2019-08-01fix calling an inferred async functionAndrew Kelley
2019-08-01reimplement async with function splitting instead of switchAndrew Kelley
2019-07-29improve support for anyframe and anyframe->TAndrew Kelley
* add implicit cast from `*@Frame(func)` to `anyframe->T` or `anyframe`. * add implicit cast from `anyframe->T` to `anyframe`. * `resume` works on `anyframe->T` and `anyframe` types.
2019-07-26add the `anyframe` and `anyframe->T` typesAndrew Kelley
2019-07-26async functions return void, no more GetSize resume blockAndrew Kelley
2019-07-25`@frameSize` works via PrefixDataAndrew Kelley
generated by cgit v1.2.3 (git 2.39.1) at 2025-12-28 10:31:51 +0000