| Stackful coroutines in C. |
|
| * `Task` & `Future` coroutines which can pause, waiting for a future. |
| * `ASleep` an example of pausing `Task`s using `Future`s. |
| * `Generator` coroutines used as generators for loops. |
| * `Coroutine` the base coroutine engine. |
|
| Your code doesn't need to do anything special to be a coroutine. Only standard, or commonly available libraries are needed. |
|
| ## Prerequisites |
|
| These libraries rely on as much as possible on C's cross-platform comfort zone. C's standard libraries are used as far as possible, but, as `threads.h` is not usually supported, `pthread.h` has been used instead. |
|
| You will need to build & link the code, `coroutine/*.c`, as part of your system, and ensure the headers, `include/*`, are available on your include path. |
|
| If your system doesn't have pthread, all the system-specific bits have been collected into cor_platform.c & .h. Replace these with versions which work with your platform. |
|
| ## Quick Start |
|
| ### Tasks |
|
| To run `Task`s: |
|
| #!C |
| #include "coroutine.h" |
| #include "task.h" |
| main(){ |
| size_t mytask_stack_size = 8192 * sizeof(void *); |
| void *res = NULL; |
| bool canceled = Task_Run(mytask_stack_size, maintask, ¶m, &res); |
| } |
|
| `Task_Run` runs tasks, switching between them when the current task waits on an `Future`. `maintask()` is run as a task. The start function for any task looks like this: |
|
| #!C |
| bool mytask(void *param, void **res){ |
|
| // do your thing here |
|
| return canceled; |
| } |
|
| When `Task` returns from its start function, it returns whether it was canceled. Canceled `Task`s are assumed to have not finished what they were doing. |
|
| Within your main task, create `Task`s and `Task_Await()` them when you want to wait for their result: |
|
| #!C |
| Task task1; |
| Task_ctor(&task1, mytask_stack_size, adifferenttask, &task1param); |
|
| void *result; |
| bool canceled = Task_Await(&task1, &result); |
|
| Task_dtor(&task1); |
|
| // use the result |
|
| When a task needs to wait for something, and wants to allow other tasks to run, it should use a `Future`: |
|
| #!C |
| Future future; |
| Future_ctor(&future); |
|
| // pass the future to the background-thing-which-might-take-a-while |
|
| void *res; |
| bool canceled = Future_Await(&future, &res); |
|
| Future_dtor(&future); |
|
| When the background-thing-which-might-take-a-while has a result: |
|
| #!C |
| Future_SetResult(future, false, result); |
|
| ### ASleep |
|
| `ASleep()` needs its own system to be started to work: |
|
| #!C |
| ASleep_StartSystem() |
| // Run tasks here which may now use ASLeep() |
| ASleep_StopSystem(); |
|
| Note that `ASleep_StartSystem()` / `ASleep_StopSystem()` is only needed once per process. |
|
| Sleeping in a task: |
|
| #!C |
| bool mytask(void *param, void **result){ |
| .. |
| ASleep(time_to_sleep); |
| .. |
| } |
|
| ### Generators |
|
| Your code needs to be in a `Coroutine` to use a `Generator`: |
|
| #!C |
| void *mycoroutine(void *param){ |
| // You can use a Generator here |
| } |
| |
| Coroutine_Run(StackSpace, mycoroutine, NULL, NULL); |
|
| You will need a generator function: |
|
| #!C |
| void *yield_my_things(void *param){ |
| bool domore = true; |
|
| // loop/call functions to find more values to yield, and when you have one: |
| domore = Generator_Yield(thing); |
| // .. if domore is false, exit your generator - it is being destructed |
|
| // not actually used by generators, but this is a useful convention for bubbling |
| // the flag out to calling functions. |
| return (void *)domore; |
| } |
|
| And to use it: |
|
| #!C |
| Generator gen; |
| Generator_ctor(&gen, generator_stack_size, yield_my_things, ".."); |
| void *thing; |
| while(Generator_Next(&gen, &thing)){ |
| // use thing - a value yielded by your generator |
| } |
| Generator_dtor(&gen); |
|
| ### Coroutines |
|
| While you can use coroutines directly, it's designed as a system to support more useful patterns, like `Async` and `Generators`. |
|
| Your coroutine will need to have a start function: |
|
| #!C |
| void *start(void *param){ |
| ... |
| } |
|
| When there is no coroutine running, start your 'main' coroutine: |
|
| #!C |
| if (Coroutine_Run(coroutine_stack_size, comain, param, &result)){ |
| // handle the failure |
| } |
|
| Create other coroutines like this: |
|
| #!C |
| Coroutine *cor = Coroutine_New(start); |
|
| When you want a Coroutine to be queued to run: |
|
| #!C |
| Coroutine_Continue(cor, value, run_early); |
|
| `value` will be start function's parameter, or the value returned from the yield. |
|
| Within the Coroutine, to yield a value, and allow other `Coroutine`s to run: |
|
| #!C |
| void *Coroutine_Yield(value, on_yield, void *me); |
|
| The on_yield function is called after the coroutine has been 'wait'ed, but before the next coroutine is resumed. |
|
| ## How it Works |
|
| The coroutine system uses the stack divided into smaller stacks for the coroutines. This means you may need to consider whether each coroutine's stack size, is right for each Coroutine, and whether your C stack size is enough for the number of coroutines you might run. |
|
| Each of your threads has its own stack - the coroutine system can be run (or not) independantly on each thread. For some special cases, you may want to adjust each of your thread's stack sizes depending on how it is used. |
|
| ## Style |
|
| The style is influenced by C++. For example, where possible, a `Something *Something_New(a, b, c)` and `Something_Delete(Something *)` will have corresponding `Somthing_ctor(Somthing *, a, b, c)` and `Something_dtor(Something *)` to initialise and finalise a `Something` on the stack, or within another object. Using `.._ctor()` and `.._dtor()` will be faster as they avoid the `malloc()` and `free()`. |
|
| #!C |
| Something *oneofthem = Something_New(); |
| // use oneofthem |
| Something_Delete(oneofthem); |
|
| Can be also be done like this, and this will run faster: |
|
| #!C |
| Something oneofthem; |
| Something_ctor(&oneofthem); |
| // use oneofthem |
| Something_dtor(&oneofthem); |
|
| The exception is `Coroutine_New()` and `Coroutine_Delete()`. The returned `Coroutine` is somewhere on your thread's stack - its memory is managed by the coroutine system, and is allocated and freed quickly. |
|
| ## Usage |
|
| When you are using coroutines or generators: |
|
| #!C |
| void *myfunc(void *){ |
| // your function here |
| } |
|
| size_t coroutine_stack_size = 8192 * sizeov(void *); |
| if (Coroutine_Run(coroutine_stack_size, myfunc, (void *)myparam, NULL)){ |
| // handle the failure |
| } |
|
| You can make many calls to `Coroutine_Run()` or `Task_Run()`. `Coroutine_Run()` ensures the system is started, and that `myfunc` is called |
| from inside a Coroutine. In paeticular, if the Coroutine system is running and `Coroutine_Run()` is called from inside a coroutine, then `myfunc` is simply called. |
|
| ## Stack Overruns |
|
| The C stack is divided into smaller stacks. There's one, the startup stack, to give some room for `start` in `Coroutine_RunSystem` to work, and then each `Coroutine` has its own stack. These have guard markers which are checked to see if the stack has overrun. If there is a stack overrun, the system cannot continue - a message is output and the programe exited. There's a number of ways to avoid this issue: |
|
| * Use less stack. This is, sometimes, the right advice, especially if the startup stack overruns. The expectation is that very little is done by `start` in `Coroutine_RunSystem`. If your situation needs more doing, you can... |
|
| * increase the stack size for your `Coroutine`. If your use case is even more demanding, such as if you want 1000s of coroutines (so you need small stack chunks), /and/ some of them can recurse an unknown amount (so you need a deep stack for that `Coroutine`), then you can... |
|
| * monitor stack headroom, and add another stack chunk if you need to: |
|
| In this last case you'll need to add some code at key points: |
|
| #!C |
| void *myfunction(void *param){ |
| if (Coroutine_GetStackHeadroom() < MIN_ALLOWED_STACK){ |
| void *result; |
| Coroutine_Err err = Coroutine_Chain(my_stack_size, myfunction, param, &result); |
| if (err){ |
| // handle failure |
| } |
| return result; |
| } |
| // do everything normally |
| } |
|
| More realistically: |
|
| #!C |
| struct myfunctionparams { |
| int a; |
| char *b; |
| struct dog *d; |
| } |
|
| void *mychain(void *param){ |
| struct myfunctionparams *myparams = (struct myfunctionparams *)params; |
| return (void *)myfunction(myparams->a, myparams->b, *myparams->d); |
| } |
|
| int myfunction(int a, char *b, struct dog d){ |
| if (Coroutine_GetStackHeadroom() < MIN_ALLOWED_STACK){ |
| struct myfunctionparams params = { |
| a, |
| b, |
| &d |
| }; |
| void *result; |
| Coroutine_Err err = Coroutine_Chain(my_stack_size, mychain, ¶ms, &result); |
| if (err){ |
| // handle failure |
| } |
| return (int)(intptr_t)result; |
| } |
| } |
|
| And if you want to panic if the C stack overruns: |
|
| #!C |
| if (Coroutine_GetStackHeadroom() < MIN_ALLOWED_COROUTINE_STACK){ |
| if (Coroutine_HasCoroutinesInFreePool() || |
| (char *)Coroutine_GetCStackTop() - c_stack_end >= MIN_ALLOWED_C_STACK) { |
| struct myfunctionparams params = { |
| a, |
| b, |
| &d |
| }; |
| void *result; |
| if (Coroutine_Chain(my_stack_size, mychain, ¶ms, &result)){ |
| // handle failure |
| } |
| return (int)(intptr_t)result; |
| } |
| // panic now |
| } |
|
| ## Configuring for Your Use Case |
|
| There's a number of adjustments which you may need to make for your situation. This are, mostly, in `cor_platform.h`. |
|
| There's some options in `coroutine.h` which you may need to adjust: |
|
| Coroutine_NS(name) |
| : This allows the `Coroutine_???()` functions to be given a different naming convention. |
|
| COROUTINE_STARTUP_STACK_SIZE |
| : The amount of stack set aside for `start` in `Coroutine_RunSystem()`. |
|
| COROUTINE_MINIMUM_STACK_SIZE |
| : The minimum stack size you'll ask for. The C stack is managed as a heap. If one of the free blocks in that |
| heap is big enough for your new Coroutine, and has spare, if that spare is too small for a `Coroutine` wanting |
| `COROUTINE_MINIMUM_STACK_SIZE` of stack, then the whole free block is given to your new Coroutine, instead |
| of being split into two. |
|
| `cor_platform.h` focuses on customisation for your particular use case. |
|
| _Cor_thread_local |
| : How to declare a variable to be thread local |
|
| COROUTINE_HAVE_ALLOCA_H |
| : Whether your system can `#incude <alloca.h>`. |
|
| _Cor_Mutex and related routines |
| : Your system's mutex. |
|
| _Cor_Realtime_Now |
| : Return a realtime clock value compatible with _Cor_Semaphore_Wait. |
|
| _Cor_Semaphore and related routines |
| : Semaphores on your system. |
|
| _Cor_Thread and related routines |
| : Threads on your system. |
|
| # API |
|
| ## Task & Future |
|
| The pattern for using async is: |
|
| #!C |
| bool mymaintask(void *param, void **result){ |
| // do your main task things here, like starting more tasks |
| } |
|
| void *res = NULL; |
| bool canceled = Task_Run(mymaintask, NULL, &res); |
|
| To create and wait for a task: |
|
| #!C |
| Task task1; |
| Task_ctor(&task1, asynctask1, &task1param); |
| void *res = NULL; |
| bool canceled = Task_Await(&task1, void **res) |
| Task_dtor(&task1); |
|
| or, if you prefer new & delete: |
|
| #!C |
| Task *task1 = Task_New(asynctask1, &task1param); |
| void *res = NULL; |
| bool canceled = Task_Await(task1, void **res) |
| Task_Delete(task1); |
|
| Inside your task, when there is something to wait for and you want other tasks to run while your task is waiting, you will need a future: |
|
| #!C |
| Future future; |
| Future_ctor(&future); |
|
| // keep &future to hand for when the background thing completes |
| bool canceled = Future_Await(&future, NULL); |
|
| Future_dtor(&future); |
|
| `Future_New()` and `Future_Delete()` are also available if you prefer that style. |
|
| Inside the callback when the background thing is complete: |
|
| #!C |
| // result is a void * |
| Future_SetResult(future, result, false); |
|
| or, if something went wrong: |
|
| #!C |
| // exception is a void * |
| Future_SetResult(future, exception, true); |
|
| Back in the task, you can respond to the future: |
|
| #!C |
| ... Future_Await has returned |
| if (canceled){ |
| // exit quickly - you've been canceled |
| // you could, for example, use the future's result as an exception, or error code here |
| } |
| // carry on - the future's result may be an actual result, that's up to you |
|
|
| ##### void Future_ctor(Future *fut) |
|
| fut |
| : The `Future` being constructed |
|
| Initialise a future. When you no longer need it, use `Future_dtor()`. |
|
| ##### Future *Future_New() |
|
| (returns) |
| : The new future |
|
| Allocates and initialises a future, When you no longer need it, use `Future_Delete()`. |
|
| ##### void Future_dtor(Future *fut) |
|
| fut |
| : The `Future` being destructed |
|
| Destruct a future previously constructed with `Future_ctor()`. |
|
| ##### void Future_Delete(Future *fut) |
|
| fut |
| : The `Future` to be destructed and freed |
|
| Delete (finalise and free) a future previously new'ed with `Future_New()` |
|
| ##### void Future_SetResult(Future *fut, bool canceled, void *value) |
|
| fut |
| : The `Future` whose result is being set |
|
| canceled |
| : The future's `canceled` setting |
|
| value |
| : The future's result `value` |
|
| Set the result of a future. This has an effect only the first time its done, ie a completed future can't be canceled and a canceled future can't be completed. When an `Future` has a result, its watchers are called back. |
|
| The `value` of a future might be a result if the future completes (when `canceled == false`), or could be some sort of exception value if `canceled == true`. The interpretation of a future's `value` is up to the user - as far as the async system is concerned, it's only a `void *`. |
|
| ##### bool Future_GetResult(Future *fut, void **res) |
|
| (returns) |
| : The `canceled` value of the `Future`. |
|
| res |
| : Where to store the value of the `Future`. This may be `NULL`. |
|
| Get the result of a future. |
|
| ##### typedef void (*Future_Watcher)(void *me, Future *fut) |
|
| A `Future_Watcher` is a callback called when a future has a result. The `me` parameter is the one passed to `Future_AddWatcher()`. `fut` is the future which has just got its result. |
|
| ##### void Future_AddWatcher(Future *fut, Future_Watcher watcher, void *me) |
|
| fut |
| : the `Future` to add a watcher to |
|
| watcher |
| : the callback to call when the `Future` has a result. |
|
| me |
| : the `me` value to pass to `watcher` when it is called back. |
|
| Add a watcher (callback) to be called when the future has a result. If the future is already complete, `watcher` is immediately called. The `me` value is passed to the watcher as its `me` parameter. It is assumed that a watcher, identified by the `(watcher, me)` pair, will only be added once. |
|
| ##### void Future_RemoveWatcher(Future *fut, Future_Watcher watcher, void *me) |
|
| fut |
| : the `Future` to remove a watcher from |
|
| watcher |
| : the callback of the watcher to remove. |
|
| me |
| : the `me` value of the watcher to remove. |
|
| Remove a watcher from a future. It is not an error if no watcher matching `(watcher, me)` is found - it has probably already been called back. |
|
| ##### bool Future_Await(Future *fut, void **res) |
|
| (returns) |
| : whether the `Future` was canceled. |
|
| fut |
| : The `Future` to wait for. |
|
| res |
| : Where to store the `value` of the future when it is has a result. May be `NULL`. |
|
| The current `Task` is paused until the `Future` has a result. Other `Task`s are run while this one is waiting. |
|
| ##### typedef bool (*Task_Entry)(void *param, void **res) |
|
| The entry function to an `Task`. |
|
| ##### void Task_ctor(Task *tsk, Task_Entry entry, void *param) |
|
| tsk |
| : The task to construct. |
|
| entry |
| : The entry function for the task. |
|
| param |
| : The value for `param` to pass to `entry`. |
|
| Initialises an `Task`. When you have finished with an `Task` you must finalise it using `Task_dtor()` |
|
| #!C |
| Task tsk; |
| Task_ctor(&tsk, mytask, myparam); |
| // tsk will run if you wait for a task or future |
| Task_Await(&tsk, NULL); |
| Task_dtor(&tsk); |
|
| ##### Task *Task_New(Task_Entry entry, void *param) |
|
| (returns) |
| : The new `Task`. |
|
| entry |
| : The entry function for the task. |
|
| param |
| : The value for `param` to pass to `entry`. |
|
| This allocates and initialises a new `Task`. When you have finished with your task, you must `Task_Delete()` it. |
|
| #!C |
| Task *tsk = Task_New(mytask, myparam); |
| // tsk will run if you wait for a task or future |
| Task_Await(tsk, NULL); |
| Task_Delete(tsk); |
|
| ##### void Task_dtor(Task *tsk) |
|
| tsk |
| : The `Task` to destruct. |
|
| This finalises an `Task` you ealier initalised with `Task_ctor()`. It is an error to attempt to destruct a task which is running. |
|
| #!C |
| Task tsk; |
| Task_ctor(&tsk, mytask, myparam); |
| // use tsk |
| Task_dtor(&tsk); |
|
| ##### void Task_Delete(Task *tsk) |
|
| tsk |
| : The `Task` to delete. |
|
| This finalises and frees an `Task` you ealier new'ed with `Task_New()`. It is an error to attempt to delete a task which is running. |
|
| #!C |
| Task *tsk = Task_New(mytask, myparam); |
| // use tsk |
| Task_Delete(tsk); |
|
| ##### static inline bool Task_Await(Task *tsk, void **res) |
|
| (returns) |
| : Whether the task was canceled. |
|
| tsk |
| : The `Task` to wait for. |
|
| res |
| : Where to store the `Task`'s value when it finishes. This may be NULL. |
|
| The current `Task` waits for `tsk` to finish, and returns the result. |
|
| ##### void Task_Cancel(Task *tsk, void *cancel_value) |
|
| tsk |
| : The task to cancel. |
|
| cancel_value |
| : The value to set on any future this task waits on. |
|
| This marks a task as canceled. When that task waits on a future that future will be canceled too, using `cancel_value`. |
|
| ##### static inline bool Task_IsCanceled(Task *tsk) |
|
| (returns) |
| : Whether the task is canceled. |
|
| tsk |
| : The task to get its canceled setting from. |
|
| ##### static inline Future *Task_GetAwaitedFuture(Task *tsk) |
|
| (returns) |
| : The future the task is waiting on. May be NULL. |
|
| tsk |
| : Teh task to read the future it is waiting on. |
|
| Return the future a task is waiting on. |
|
| ##### bool Task_Run(Task_Entry start, void *value, void **res) |
|
| (returns) |
| : Whether `start` was canceled. |
|
| start |
| : The function to use as the main task. |
|
| value |
| : The value to pass to `start`. |
|
| res |
| : Where to store the result of `start`. |
|
| Runs `start` as an `Task`. When `start` returns all other tasks must have been destructed, using `Task_dtor()` or `Task_Delete()`. |
|
| ## ASleep |
|
| ##### void ASleep_StartSystem() |
|
| You must start the `ASleep` system to use it. This needs to happen per process. Once you've finished with `ASleep` you must `ASleep_StopSystem()`. |
|
| #!C |
| ASleep_StartSystem(); |
| // Now you can use ASleep() on any thread |
| ASleep_StopSystem(); |
|
| ##### void ASleep_StopSystem() |
|
| Call this to stop the `ASleep` system. |
|
| ##### bool ASleep(float delay, void **value) |
|
| (returns) |
| : Whether the task was canceled. |
|
| delay |
| : How many seconds to delay for. |
|
| value |
| : Where to store the cancellation value. This may be NULL. |
|
| Sleep for `delay` seconds. `*value` will be set to `NULL` if the sleep is successful, and the `cancel_value` if the task is canceled. |
|
| ## Generator |
|
| The pattern for a `Generator` is: |
|
| #### A loop which uses the `Generator |
|
| #!C |
| Generator gen; |
| Generator_ctor(&gen, generator_stack_size, mygen, ¶m); |
|
| void *value; |
| while(Generator_Next(&gen, &value)){ |
| // use value here |
| } |
| // value is now the return value from the Generator |
|
| Generator_dtor(&gen); |
|
| Or: |
|
| #!C |
| Generator *gen = Generator_New(generator_stack_size, mygen, ¶m); |
|
| void *value; |
| while(Generator_Next(gen, &value)){ |
| // use value here |
| } |
|
| Generator_Delete(gen); |
|
| `Generator`s yield a series of `void *`s - what the `void *`s mean is up to you. `Generator_Next()` returns a `bool` to indicate whether the `Generator` has finished. |
| The `generator_stack_size` is the stack amount made available to your generator. |
|
| #### A generator function |
|
| #!C |
| void *mygen(void *param){ |
| bool domore = true; |
| // The parameter is a pointer to a string of chars |
| for (char *str = param; *str; ++str) { |
| // The value yielded is a pointer to a character in the string |
| domore = Generator_Yield(str); |
| if (!domore){ |
| break; |
| } |
| } |
|
| return (void *)domore; |
| } |
|
| The `bool` returned from `Generator_Yield()` indicates whether the generator function should yield more values. When it is `false` the `Generator` is being finalised - your generator function should close files, and release any other resources it has claimed, before exiting. |
|
| ##### void Generator_ctor(Generator *gen, size_t generator_stack_size, void *(*start)(void *), void *param) |
|
| gen |
| : The `Generator` to construct. |
|
| generator_stack_size |
| : The amount of stack given to the generator. |
|
| start |
| : The function which is the start/entry-point of the `Generator`. |
|
| param |
| : The value to pass to `start`. |
|
| Initialise a `Generator`. When you no longer need the `Generator`, use `Generator_dtor()` to destruct it. |
|
| #!C |
| Generator gen; |
| Generator_ctor(&gen, generator_stack_size, mystart, ¶ms); |
|
| // Generator is used |
|
| // ... later: |
| Generator_dtor(&gen); |
|
| ##### Generator *Generator_New(size_t generator_stack_size, void *(*start)(void *), void *param) |
|
| generator_stack_size |
| : The amount of stack to give to the generator. |
|
| start |
| : The function which is the start/entry-point of the `Generator`. |
|
| param |
| : The value to pass to `start`. |
|
| `new` a `Generator` - malloc, and initialise it. When you no longer need the `Generator` use `Generator_dtor` to finalise it. |
|
| #!C |
| Generator *gen = Generator_New(mystart, ¶ms); |
|
| // Generator is used |
|
| // ... later: |
| Generator_Delete(gen); |
|
| ##### void Generator_dtor(Generator *gen) |
|
| gen |
| : The `Generator` to destruct. |
|
| Finalise a `Generator`. Once a `Generator` is no longer needed, it must be finalised: |
|
| #!C |
| // earlier... |
| Generator gen; |
| Generator_ctor(&gen, generator_stack_size, mystart, ¶ms); |
|
| // Generator is used |
|
| // the Generator is no longer needed |
| Generator_dtor(&gen); |
|
|
| ##### void Generator_Delete(Generator *gen) |
|
| gen |
| : The `Generator` to delete. |
|
| Finalise then `free()` a `Generator`. Once a `new`ed `Generator` is no longer needed, it must be deleted: |
|
| #!C |
| // earlier... |
| Generator *gen = Generator_New(mystart, ¶ms); |
|
| // Generator is used |
|
| // the Generator is no longer needed |
| Generator_Delete(gen); |
|
|
| ##### bool Generator_Next(Generator *gen, void **value) |
|
| (returns) |
| : Whether there is a next value. `true` - there is a next value; `false` - the `Generator` has finished |
|
| gen |
| : The `Generator` to get the next value from. |
|
| value |
| : Where to store the next value. |
|
| Get the next value yielded by the `Generator`. |
|
| #!C |
| void *value; |
| while(Generator_Next(gen, &value)){ |
| // use value here |
| } |
|
| The `Generator` feeds values to its client using `Generator_Yield()` - it is these values which `Generator_Next()` sets, in the example, `value` to. |
|
| When a `Generator` is finished it returns from `start`. When you call `Generator_Yield()` on a finished `Generator` it returns `false` and `value` will be the return value from `start`. |
| |
| ##### bool Generator_Yield(void *value) |
|
| (returns) |
| : Whether the `Generator` should do more. |
|
| value |
| : The `Generator`'s next value. |
|
| Yield a value from a `Generator`. |
|
| #!C |
| bool domore = Generator_Yield(value); |
|
| `value` is then provided by `Generator_Next()` as the next value from the generator. |
|
| The `bool` returned by `Generator_Yield()` says whether more values should be provided by your generator function. `true` - provide more values if there are any. `false` - close files, free memory, free up any other resources and `return`. `false` is returned when the `Generator` is being finalised before it has finished, ie the client has exited its `for`-loop early. |
|
| ## Coroutine |
|
| ##### Coroutine_Err |
|
| The enum of errors: |
|
| Coroutine_OK |
| : Everything is OK. This is 0 |
|
| Coroutine_Err_SystemNotRunning |
| : A `Coroutine` must be running to do this |
|
| Coroutine_Err_SystemRunning |
| : The `Coroutine` system must not be running to do this |
|
| Coroutine_Err_NoStack |
| : Not enough stack is available |
|
| Coroutine_Err_CoroutineFromWrongThread |
| : Trying to do something on one thread to a `Coroutine` from a different thread |
|
| Coroutine_Err_ACoroutineIsAlreadyRunning |
| : Trying `Coroutine_RunCoroutine` a `Coroutine` |
|
| Coroutine_Err_ExitWithRunningCoroutines |
| : All `Coroutine`s must be complete |
|
| Coroutine_Err_StackOverrun |
| : Stack overrun detected |
|
| Coroutine_Err_InternalInsistency |
| : Something didn't match inside the system |
|
| Coroutine_Err_CouldNotInitialiseSystem |
| : Something went wrong initialising (eg couldn't create a lock) |
|
| Coroutine_Err_WrongState |
| : It's in the wrong statem, eg trying to `Coroutine_Continue` a completed `Coroutine` |
|
| Coroutine_Err_Canceled |
| : It's canceled |
|
| ##### Coroutine_SetStackLimit(void *limit) |
|
| limit |
| : The location (low address) of the stack's end. |
|
| Set the limit of the stack. This is used to determine more accurately whether `Coroutine_CanStartCoroutine()` |
|
| ##### Coroutine_Report Coroutine_GetReport() |
|
| (returns) |
| : A report from this run of the Coroutine system. |
|
| #!C |
| typedef struct Coroutine_Report { |
| unsigned coroutines_created; |
| unsigned coroutines_pool_size; |
| unsigned lowest_headroom; |
| } Coroutine_Report; |
|
| coroutines_created |
| : How many coroutines were created |
|
| coroutines_pool_size |
| : The size of the coroutine pool (count of available, free `Coroutine` objects) when the system stopped. This is also the peak number of active coroutines. This will give you an idea of how much stack was needed for your coroutines. |
|
| lowest_headroom |
| : The lowest headroom (unused |
|
| largest_stack |
| : The largest stack requested for any `Coroutine`. |
|
| ##### Coroutine_CheckIntegrity() |
|
| (returns) |
| : `Coroutine_Err` for any problem |
|
| Check the integrity of the coroutine system, and `printf()` any problems. |
|
| ##### Coroutine_Start |
|
| #!C |
| void *(*)(void *param) |
|
| The entry function for a coroutine. The `param` is the value passed to `Coroutine_Continue`, and the `void *` return value can be accessed through the `Coroutine` object using `Coroutine_GetValue()`. |
|
| ##### Coroutine_SystemStart |
|
| #!C |
| Coroutine_Err (*)(void *) |
|
| The entry function for `Coroutine_RunSystem`. |
|
| ##### Coroutine_Err Coroutine_RunSystem(Coroutine_SystemStart start, void *value) |
|
| (returns) |
| : `Coroutine_OK` or an error. If the system starts, this will be the value returned by `start`. |
|
| start |
| : The function to call with the `Coroutine` system started. It is expected that this routine will |
| start a `Coroutine`. |
|
| value |
| : The value to pass to `start`. |
|
| ##### Coroutine *Coroutine_New(size_t size, Coroutine_Start start) |
|
| (returns) |
| A new Coroutine, or `NULL` if there was a failure, such as insufficient stack for the new Coroutine. |
|
| size |
| : The stack size given to this Coroutine. |
|
| start |
| : The routine called to start the Coroutine. |
|
| Create a new `Coroutine`. The `Coroutine` system must be started to create a `Coroutine`. The stack size available to the coroutine will be `COROUTINE_STACK_SIZE` defined in `coroutine.h`. When you have finished with your `Coroutine`, use `Coroutine_Delete()` to delete it. If there is not enough space for a new `Coroutine` on your stack, `NULL` will be returned. |
|
| ##### Coroutine_Err Coroutine_Run_Coroutine(Coroutine *cor, void *value) |
|
| (returns) |
| : any problem, or `Coroutine_OK` |
|
| cor |
| : The Coroutine to run. |
|
| value |
| : The value to pass to the Coroutine's `start` routine. |
|
| Run the `Coroutine` and return when it returns. This is how to start coroutines running in the coroutine system. It is an error for the run coroutine to return before all other coroutines have completed, and the coroutine system must be started to call this. |
|
| ##### Coroutine_Err Coroutine_Run(size_t size, Coroutine_Start start, void *value, void **result) |
|
| (returns) |
| : `Coroutine_OK` or any problem |
|
| size |
| : The stack size to give to the Corotuine. |
|
| start |
| : The routine to start the Coroutine. |
|
| value |
| : The value to pass to `start()`. |
|
| result |
| : Where to store the return value from `start(value)`. This may be `NULL`. |
|
| `start(value)` is called from within a coroutine and its value returned in `*result`. |
| If this completes without any failure, `false` is returned, otherwise, typically |
| because `Coroutine_New()` returned `NULL`, `true` is returned. `result` may be `NULL` if you don't |
| need the resturn value from `start()`. |
| When the coroutine system is active - you are already running in a coroutine - `start(value)` |
| is simply called and its result returned in `*result`. When the Coroutine system is not running, |
| `Coroutine_Run()` starts it, creates a `Coroutine` and runs that Coroutine to call `start(calue)` |
| and return value is returned in `*result`, then stops the Coroutine system. If you need to force |
| a new Coroutine to be created, with a particular stack size to call `start(value)`, then use |
| `Coroutine_Chain()` instead. |
|
| The total stack allowed for all coroutines running on any thread is the size of the call stack on that thread. |
|
| ##### void Coroutine_Delete(Coroutine *cor) |
|
| cor |
| : The Coroutine to delete. |
|
| Use `Coroutine_Delete()` to delete a coroutine when it is no longer needed. It is an error to attempt to delete a coroutine which is running. |
|
| ##### Coroutine_Err Coroutine_Continue(Coroutine *cor, void *value, bool early) |
|
| (returns) |
| : `Coroutine_OK` or any error. |
|
| cor |
| : The Coroutine to continue. |
|
| value |
| : The value to return from `cor`'s yield function. |
|
| early |
| : Whether to continue `cor` early (`true`), or late (`false`). Early means before other Coroutines which are waiting |
| to be called, whereas late means after them. |
|
| Continue the given `Coroutine`. `value` is passed to the coroutine, as `param` to the `start` function, or as the return value from `Coroutine_Yield`. `early` determines whether the continued coroutine will be run next, or after all the other, currently runnable, coroutines. If the `Coroutine` is already runnable, nothing is done, and `false` is returned. If the `Coroutine` is free, or complete, nothing is done and `true` is returned to show there was a problem. |
|
| ##### void *Coroutine_Yield(void *value, Coroutine_YieldCallback on_yield, void *this) |
|
| value |
| : The value to yield fropm the coroutine. |
|
| on_yield |
| : A callback to be called once this Coroutine has yielded, but before another one has been continued. |
|
| this |
| : The parameter to pass to `on_yield`. |
|
| Yield `value` from the current coroutine; this coroutine is moved to the list of coroutines waiting to be continued. The next runable coroutine is run - either by its start routine being called with `value` as its `param`, or by `value`being returned from its `Coroutine_Yield()`. |
|
| ##### void *Coroutine_GetValue(Coroutine *cor) |
|
| (returns) |
| : The Coroutine's value - the last yielded or returned value. |
|
| cor |
| : The Coroutine to query. |
|
| Return the `Coroutine`'s value - the value last yielded, or returned by its `start` routine. |
|
| ##### Coroutine *Coroutine_GetActive() |
|
| (returns) |
| : The currently active Coroutine. |
|
| Return whihc coroutine is currently running, ie the caller's `Coroutine`. |
|
| ##### bool Coroutine_IsRunning(Coroutine *cor) |
|
| (returns) |
| : Whether `cor` is running - it's the active coroutine or waiting to be continued. |
|
| cor |
| : The Coroutine to query. |
|
| Return whether the given coroutine is still running - it may be running, ready to run, or waiting to be continued, but won't have returned from its `start` function. |
|
| ##### bool Coroutine_IsComplete(Coroutine *cor) |
|
| (returns) |
| : Whether `cor` is complete, ie has returned from `start()`./ |
|
| cor |
| : The Coroutine to query. |
|
| Return whether the given coroutine is complete - is has returned from its `start` function. |
|
| ##### intptr_t Coroutine_GetStackHeadroom() |
|
| (returns) |
| : The amount of stack headroom. |
|
| Return the headroom available in the current coroutine's stack. This can be used to detect when your coroutine is nearing its stack limit, and then use `Coroutine_Chain()` to continue in a new chunk of coroutine stack. |
|
| ##### bool Coroutine_CanStartCoroutine(size_t size) |
|
| (returns) |
| : Whether a Coroutine with the given amount of stack could be created. |
|
| size |
| : The amount of stack in the Coroutine we might want to create. |
|
| Return whether the coroutine system can start a new coroutine. This check can only be done with the coroutine system active (currently running |
| a coroutine). If there's a free coroutine, or enough space on the stack for a new one, then this will return `true`. To set the limit of the |
| stack use `Coroutine_SetStackLimit()` |
|
| ##### void *Coroutine_GetStackHWM(void) |
|
| (returns) |
| : The lowest address where the active Coroutine's stack has grown to ever. |
|
| Find out where this coroutine's guard patterns end. This is intended as a part of the tools to measure how much stack something is using: |
|
| #!C |
| Coroutine_ClearStackForHWM(); |
| char *before = (char *)Coroutine_GetStackHWM(); |
| // do the thing you want to measure here |
| char *after = (char *)Coroutine_GetStackHWM(); |
| intptr_t amount_used = before - after; |
|
| ##### void Coroutine_ClearStackForHWM(void) |
|
| Fill the unused stack in this coroutine with a guard pattern. This is intended as a part of the tools to measure how much stack something is using: |
|
| #!C |
| Coroutine_ClearStackForHWM(); |
| char *before = (char *)Coroutine_GetStackHWM(); |
| // do the thing you want to measure here |
| char *after = (char *)Coroutine_GetStackHWM(); |
| intptr_t amount_used = before - after; |
|
| ##### void *Coroutine_GetCStackTop() |
|
| (returns) |
| : Where the Coroutine system has reached in the C stack. |
|
| Return an address which is near to the top of used C stack. |
|
| ##### Coroutine_Err Coroutine_Chain(size_t size, Coroutine_Start start, void *value, void **result) |
|
| (returns) |
| : Whether there was a problem. `Coroutine_OK` - `start(value)` was run; an error - there was a problem. |
|
| size |
| : The amount of stack to give the chained Coroutine. |
|
| start |
| : The entry point ot the chained Coroutine. |
|
| value |
| : The value to pass to `start()` |
|
| result |
| : Where to store the return value from `start(value)`. This may be `NULL`. |
|
| Run `start` with `value` on a new coroutine, and return its return value in `*result`. `result` |
| may be NULL. `Coroutine_Run()` returns `false` if nothing fails, and `true` if something went wrong, |
| usually when `Coroutine_New()` ran out of stack. `stack_size` is the amount of stack made available |
| to the chained Coroutine. |
| It is expected that `Coroutine_Chain()` will be used when your coroutine is running short |
| of stack - it is not an alternative to `Coroutine_Run()`. |
|
| ##### void Coroutine_Dump_() |
|
| *Do not use this function in production code* |
|
| This prints the current state of the Coroutine system. It is used for development, and is not part of the official interface. |
|
jonroach