efl/src/lib/eina/eina_promise.h

1894 lines
68 KiB
C
Raw Normal View History

#ifndef _EINA_PROMISE_H_
#define _EINA_PROMISE_H_
#ifdef __cplusplus
extern "C" {
#endif
#include "eina_safety_checks.h"
#include "eina_types.h"
#include "eina_value.h"
/** @file */
/** @struct Eina_Promise
* @ingroup Eina_Promise
* An opaque structure representing a piece of data that will be available at a later point.
* @typedef struct _Eina_Promise Eina_Promise
* Convenience wrapper.
*/
typedef struct _Eina_Promise Eina_Promise;
/** @struct Eina_Future
* @ingroup Eina_Future
* An opaque structure representing a callback to be called when a promise is fulfilled.
* @typedef struct _Eina_Future Eina_Future
* Convenience wrapper.
*/
typedef struct _Eina_Future Eina_Future;
/** @typedef struct _Eina_Future_Desc Eina_Future_Desc
* @ingroup Eina_Future
* Convenience wrapper over #_Eina_Future_Desc. */
typedef struct _Eina_Future_Desc Eina_Future_Desc;
/** @typedef struct _Eina_Future_Race_Result Eina_Future_Race_Result
* @ingroup Eina_Future
* Convenience wrapper over #_Eina_Future_Race_Result. */
typedef struct _Eina_Future_Race_Result Eina_Future_Race_Result;
/**
* @defgroup Eina_Future_Callbacks Eina Future Callbacks
* @ingroup Eina_Future
* @brief Methods and structures dealing with #Eina_Future callbacks
* @{
*/
/** @typedef struct _Eina_Future_Cb_Easy_Desc Eina_Future_Cb_Easy_Desc
* Convenience wrapper over #_Eina_Future_Cb_Easy_Desc. */
typedef struct _Eina_Future_Cb_Easy_Desc Eina_Future_Cb_Easy_Desc;
/** @typedef struct _Eina_Future_Cb_Console_Desc Eina_Future_Cb_Console_Desc
* Convenience wrapper over _Eina_Future_Cb_Console_Desc. */
typedef struct _Eina_Future_Cb_Console_Desc Eina_Future_Cb_Console_Desc;
/** @typedef struct _Eina_Future_Scheduler Eina_Future_Scheduler
* Convenience wrapper over _Eina_Future_Scheduler. */
typedef struct _Eina_Future_Scheduler Eina_Future_Scheduler;
/** @typedef struct _Eina_Future_Schedule_Entry Eina_Future_Schedule_Entry
* Convenience wrapper over _Eina_Future_Schedule_Entry. */
typedef struct _Eina_Future_Schedule_Entry Eina_Future_Schedule_Entry;
/** @typedef struct _Eina_Future_Cb_Log_Desc Eina_Future_Cb_Log_Desc
* Convenience wrapper over _Eina_Future_Cb_Log_Desc. */
typedef struct _Eina_Future_Cb_Log_Desc Eina_Future_Cb_Log_Desc;
/**
* @typedef Eina_Future_Cb
* A callback used to inform that a future was resolved.
* Usually this callback is called from a clean context, that is, from the
* main loop or some platform defined safe context. However there are
* 2 exceptions:
*
* @li eina_future_cancel() was used, it's called immediately in the
* context that called cancel using `ECANCELED` as error.
*
* @li eina_future_then(), eina_future_then_from_desc(), eina_future_chain(), eina_future_chain_array()
* or similar failed due to invalid pointer or memory allocation. Then the callback is called from the
* failed context using `EINVAL` or `ENOMEM` as errors and @p dead_future will be @c NULL.
*
* @param[in] data The data provided by the user
*
* @param[in] value An #Eina_Value which contains the operation result. Before using
* the @p value, its type must be checked in order to avoid errors. This is needed because
* if an operation fails the #Eina_Value type will be #EINA_VALUE_TYPE_ERROR
* which is a different type than the expected operation result.
*
* @param[in] dead_future A pointer to the future that was completed.
*
* @return An #Eina_Value to pass to the next #Eina_Future in the chain (if any).
* If there is no need to convert the received value, it's @b recommended
* to passthrough the @p value argument. If you need to convert to a different type
* or generate a new value, use @c eina_value_setup() on @b another #Eina_Value
* and return it. By returning a promise #Eina_Value (eina_promise_as_value()) the
* whole chain will wait until the promise is resolved in
* order to continue its execution.
* Note that the value contents must survive this function scope,
* that is, do @b not use stack allocated blobs, arrays, structures or types that
* keep references to memory you give. Values will be automatically cleaned up
* using @c eina_value_flush() once they are unused (no more future or futures
* returned a new value).
*
* @note The returned value @b can be an #EINA_VALUE_TYPE_PROMISE! (see eina_promise_as_value() and
* eina_future_as_value()) In this case the future chain will wait until the promise is resolved.
*
* @see eina_future_cancel()
* @see eina_future_then()
* @see eina_future_then_from_desc()
* @see eina_future_then_easy()
* @see eina_future_chain()
* @see eina_future_chain_array()
* @see eina_future_as_value()
* @see eina_promise_as_value()
*/
typedef Eina_Value (*Eina_Future_Cb)(void *data, const Eina_Value value, const Eina_Future *dead_future);
/**
* @struct _Eina_Future_Schedule_Entry
*
* A struct that represents an scheduled event.
* This struct may be used by Eina to cancel
* a scheduled future.
*
* @see eina_promise_new()
*
* @see Eina_Future_Scheduler
* @see Eina_Future_Scheduler_Cb
*/
struct _Eina_Future_Schedule_Entry {
/**
* The scheduler used to create the entry.
* @note This must not be @c NULL.
*/
Eina_Future_Scheduler *scheduler;
};
/**
* @typedef Eina_Future_Scheduler_Cb
* A callback used by the #Eina_Future_Scheduler to deliver
* the future operation result.
*
* @param[out] f The delivered future.
* @param[in] value The future result
*
*
* @see eina_promise_new()
*
* @see Eina_Future_Schedule_Entry
* @see Eina_Future_Scheduler
*/
typedef void (*Eina_Future_Scheduler_Cb)(Eina_Future *f, Eina_Value value);
/**
* @struct _Eina_Future_Scheduler
* This struct is used as a bridge between Eina and the future scheduler.
* By using the functions provided by _Eina_Future_Scheduler Eina can
* schedule futures resolutions, rejections and cancellations to a safe context.
*
* @see eina_promise_new()
* @see Eina_Future_Schedule_Entry
* @see Eina_Future_Scheduler_Cb
*/
struct _Eina_Future_Scheduler {
/**
* Called by @p Eina_Future when a delivery must be scheduled to a safe context.
* i.e.: after @p eina_promise_resolve()
*
* @note Must not be @c NULL
*
* Must call back from a safe context using @p cb(f,value)
* @param[in,out] scheduler The scheduler to use.
* @param[in] cb The Eina_Future_Scheduler_Cb to be called and deliver the @p f and @p value.
* @param[in] f The future to be delivered to @p cb
* @param[in] value The value to be delivered to @p cb
* @return A scheduled entry or @c NULL on error
*/
Eina_Future_Schedule_Entry *(*schedule)(Eina_Future_Scheduler *scheduler, Eina_Future_Scheduler_Cb cb, Eina_Future *f, Eina_Value value);
/**
* Called by @p Eina_Future when a delivery must be canceled.
* i.e.: after @p eina_future_cancel()
*
* @note Must not be @c NULL.
*
* @param[in,out] entry The scheduled event to cancel
*/
void (*recall)(Eina_Future_Schedule_Entry *entry);
};
/**
* @typedef void (*Eina_Promise_Cancel_Cb) (void *data, const Eina_Promise *dead_promise)
*
* A callback used to inform that a promise was canceled.
*
* A promise may be canceled by the user calling `eina_future_cancel()`
* on any #Eina_Future that is part of the chain that uses an #Eina_Promise,
* that will cancel the whole chain and then the promise.
*
* It should stop all asynchronous operations or at least mark them to be
* discarded instead of resolved. Actually it can't be resolved once
* canceled since the given pointer @c dead_promise is now invalid.
*
* @note @li This callback is @b mandatory for a reason, do not provide an empty
* callback as it'll likely result in memory corruption and invalid access.
* If impossible to cancel an asynchronous task, then create an
* intermediate memory to hold #Eina_Promise and make it @c NULL,
* in this callback. Then prior to resolution check if the pointer is set.
*
* @note @li This callback is @b not called if eina_promise_resolve() or
* eina_promise_reject() are used. If any cleanup is needed, then
* call yourself. It's only meant as cancellation, not a general
* "free callback".
*
* @param[in] data The data provided by the user.
* @param[in] dead_promise The canceled promise.
* @see eina_promise_reject()
* @see eina_promise_resolve()
* @see eina_future_cancel()
*/
typedef void (*Eina_Promise_Cancel_Cb) (void *data, const Eina_Promise *dead_promise);
/**
* @typedef Eina_Future_Success_Cb
*
* A callback used to inform that the future completed with success.
*
* Unlike #Eina_Future_Cb this callback is only called if the future operation was successful, this is,
* no errors occurred (@p value type differs from #EINA_VALUE_TYPE_ERROR)
* and the @p value matches _Eina_Future_Cb_Easy_Desc::success_type (if given).
* In case _Eina_Future_Cb_Easy_Desc::success_type was not supplied (it's @c NULL) the @p value type
* must be checked before using it.
*
* @note This function is always called from a safe context (main loop or some platform defined safe context).
*
* @param[in] data The data provided by the user.
* @param[in] value The operation result
* @return An Eina_Value to pass to the next Eina_Future in the chain (if any).
* If there is no need to convert the received value, it's @b recommended
* to passthrough @p value argument. If you need to convert to a different type
* or generate a new value, use @c eina_value_setup() on @b another Eina_Value
* and return it. By returning a promise Eina_Value (eina_promise_as_value()) the
* whole chain will wait until the promise is resolved in
* order to continue its execution.
* Note that the value contents must survive this function scope,
* that is, do @b not use stack allocated blobs, arrays, structures or types that
* keep references to memory you give. Values will be automatically cleaned up
* using @c eina_value_flush() once they are unused (no more future or futures
* returned a new value).
* @see eina_future_cb_easy_from_desc()
* @see eina_future_cb_easy()
*/
typedef Eina_Value (*Eina_Future_Success_Cb)(void *data, const Eina_Value value);
/**
* @typedef Eina_Future_Error_Cb
*
* A callback used to inform that the future completed with failure.
*
* Unlike #Eina_Future_Success_Cb this function is only called when an error
* occurs during the future process or when _Eina_Future_Cb_Easy_Desc::success_type
* differs from the future result.
* On future creation errors and future cancellation this function will be called
* from the current context with the following errors: `EINVAL`, `ENOMEM` and `ECANCELED`.
* Otherwise this function is called from a safe context.
*
* If it was possible to recover from an error this function should return an empty value
* #EINA_VALUE_EMPTY or any other #Eina_Value type that differs from #EINA_VALUE_TYPE_ERROR.
* In this case the error will not be reported by the other futures in the chain (if any), otherwise
* if an #Eina_Value type #EINA_VALUE_TYPE_ERROR is returned the error will continue to be reported
* to the other futures in the chain.
*
* @param[in] data The data provided by the user.
* @param[in] error The operation error
* @return An #Eina_Value to pass to the next #Eina_Future in the chain (if any).
* If you need to convert to a different type or generate a new value,
* use eina_value_setup() on @b another Eina_Value
* and return it. By returning a promise Eina_Value (eina_promise_as_value()) the
* whole chain will wait until the promise is resolved in
* order to continue its execution.
* Note that the value contents must survive this function scope,
* that is, do @b not use stack allocated blobs, arrays, structures or types that
* keep references to memory you give. Values will be automatically cleaned up
* using eina_value_flush() once they are unused (no more future or futures
* returned a new value).
* @see eina_future_cb_easy_from_desc()
* @see eina_future_cb_easy()
*/
typedef Eina_Value (*Eina_Future_Error_Cb)(void *data, const Eina_Error error);
/**
* @typedef Eina_Future_Free_Cb
*
* A callback used to inform that the future was freed and the user should also @c free the @p data.
* This callback may be called from an unsafe context if the future was canceled or an error
* occurred.
*
* @note This callback is always called, even if #Eina_Future_Error_Cb and/or #Eina_Future_Success_Cb
* were not provided, which can also be used to monitor when a future ends.
*
* @param[in] data The data provided by the user.
* @param[in] dead_future The future that was freed.
*
* @see eina_future_cb_easy_from_desc()
* @see eina_future_cb_easy()
*/
typedef void (*Eina_Future_Free_Cb)(void *data, const Eina_Future *dead_future);
/**
* @struct _Eina_Future_Cb_Easy_Desc
*
* A struct with callbacks to be used by eina_future_cb_easy_from_desc() and eina_future_cb_easy()
*
* @see eina_future_cb_easy_from_desc()
* @see eina_future_cb_easy()
*/
struct _Eina_Future_Cb_Easy_Desc {
/**
* Called on success (value.type is not @c EINA_VALUE_TYPE_ERROR).
*
* if @c success_type is not NULL, then the value is guaranteed to be of that type,
* if it's not, then it will trigger @c error with @c EINVAL.
*
* After this function returns, @c free callback is called if provided.
*/
Eina_Future_Success_Cb success;
/**
* Called on error (value.type is @c EINA_VALUE_TYPE_ERROR).
*
* This function can return another error, propagating or converting it. However it
* may also return a non-error, in this case the next future in chain will receive a regular
* value, which may call its @c success.
*
* If this function is not provided, then it will passthrough the error to the next error handler.
*
* It may be called with @c EINVAL if @c success_type is provided and doesn't
* match the received type.
*
* It may be called with @c ECANCELED if future was canceled.
*
* It may be called with @c ENOMEM if memory allocation failed during callback creation.
*
* After this function returns, @c free callback is called if provided.
*/
Eina_Future_Error_Cb error;
/**
* Called on @b all situations to notify future destruction.
*
* This is called after @c success or @c error, as well as it's called if none of them are
* provided. Thus can be used as a "weak ref" mechanism.
*/
Eina_Future_Free_Cb free;
/**
* If provided, then @c success will only be called if the value type matches the given pointer.
*
* If provided and doesn't match, then @c error will be called with @c EINVAL. If no @c error,
* then it will be propagated to the next future in the chain.
*/
const Eina_Value_Type *success_type;
/**
* Context data given to every callback.
*
* This must be freed @b only by @c free callback as it's called from every case,
* otherwise it may lead to memory leaks.
*/
const void *data;
};
/**
* @struct _Eina_Future_Cb_Console_Desc
*
* A struct used to define the prefix and suffix to be printed
* along side the a future value. This struct is used by
* eina_future_cb_console_from_desc()
*
* @see eina_future_cb_console_from_desc()
*/
struct _Eina_Future_Cb_Console_Desc {
/**
* The prefix to be printed. If @c NULL an empty string ("") is used.
*/
const char *prefix;
/**
* The suffix the be printed. If @c NULL '\n' is used.
*/
const char *suffix;
};
/**
* @struct _Eina_Future_Cb_Log_Desc
*
* This struct is used by eina_future_cb_log_from_desc() and
* its contents will be routed to eina_log_print() along side
* the future value.
*
* @see eina_future_cb_log_from_desc()
*/
struct _Eina_Future_Cb_Log_Desc {
/**
* The prefix to be printed. If @c NULL an empty string ("") is used.
*/
const char *prefix;
/**
* The suffix the be printed. If @c NULL '\n' is used.
*/
const char *suffix;
/**
* The file name to be passed to eina_log_print(). if @c NULL "Unknown file" is used.
*/
const char *file;
/**
* The file name to be passed to eina_log_print(). if @c NULL "Unknown function" is used.
*/
const char *func;
/**
* The Eina_Log_Level to use.
*/
Eina_Log_Level level;
/**
* The log domain to use.
*/
int domain;
/**
* The line number to be passed to eina_log_print().
*/
int line;
};
/**
* @}
*/
/**
* @struct _Eina_Future_Desc
* @ingroup Eina_Future
* A struct used to define a callback and data for a future.
*
* This struct contains a future completion callback and a data to the future
* completion callback which is used by eina_future_then(), eina_future_chain()
* and friends to inform the user about the future result. The _Eina_Future_Desc::data
* variable should be freed when _Eina_Future_Desc::cb is called, otherwise it will leak.
*
* @note If eina_future_then(), eina_future_chain() and friends fails to return a valid future
* (in other words: @c NULL is returned) the _Eina_Future_Desc::cb will be called
* report an error like `EINVAL` or `ENOMEM` so _Eina_Future_Desc::data can be freed.
*
* @see eina_future_then()
* @see eina_future_chain_array()
* @see eina_future_cb_convert_to()
* @see eina_future_cb_console_from_desc()
* @see eina_future_cb_ignore_error()
* @see eina_future_cb_easy_from_desc()
* @see eina_future_cb_log_from_desc()
*/
struct _Eina_Future_Desc {
/**
* Called when the future is resolved or rejected.
*
* Once a future is resolved or rejected this function is called passing the future result
* to inform the user that the future operation has ended. Normally this
* function is called from a safe context (main loop or some platform defined safe context),
* however in case of a future cancellation (eina_future_cancel()) or if eina_future_then(),
* eina_future_chain() and friends fails to create a new future,
* this function is called from the current context.
*
* Use this function to free @p data if necessary.
* @see eina_future_chain()
* @see eina_future_then()
* @see eina_future_cancel()
*/
Eina_Future_Cb cb;
/**
* Context data to @p cb. The @p data should be freed inside @p cb, otherwise
* its memory will leak!
*/
const void *data;
/**
* The storage will be used by Eina to store a pointer to the
* created future. It can be @c NULL.
*/
Eina_Future **storage;
};
/**
* @defgroup Eina_Promise Eina Promises
* @ingroup Eina
*
* @brief Promises are a programming paradigm that simplifies synchronization when concurrent execution is present.
*
* Since C does not natively support Promises the Eina library provides the Eina_Promise and Eina_Future objects.
*
* In procedural languages like C if you need a value which is not yet available, for instance because
* it takes a long time to calculate or has to be fetched from a remote server, you typically have to wait.
*
* Other languages however can return a Promise and continue execution immediately. A promise acts as a placeholder
* for the requested value: the value is not available yet but will be at some point in the future.
*
* With a promise in hand you can attach callbacks to be triggered when the value becomes available (i.e. when
* the promise is fulfilled) and then continue your calculations. You can even pass the promise to other methods
* which will then be executed as values become available, forming complex chains.
*
* An Eina_Promise can be considered as an object with the sole purpose of emitting the "Promise Resolved" event.
* Eina_Future are callbacks attached to this object to be called when the event is emitted. The promised value is
* passed to the callbacks whenever it's available.
*
* Here's a typical example:
*
* @code
* #include <Ecore.h>
*
* static void
* _promise_cancel(void *data, Eina_Promise *p EINA_UNUSED)
* {
* Ctx *ctx = data;
* // In case the promise is canceled we must stop the timer!
* ecore_timer_del(ctx->timer);
* free(ctx);
* }
*
* static Eina_Bool
* _promise_resolve(void *data)
* {
* Ctx *ctx = data;
* Eina_Value v;
* eina_value_setup(&v, EINA_VALUE_TYPE_STRING);
* eina_value_set(&v, "Promise resolved");
* eina_promise_resolve(ctx->p, v);
* free(ctx);
* return EINA_FALSE;
* }
*
* Eina_Promise *
* promise_create(Eina_Future_Scheduler *scheduler)
* {
* Ctx *ctx = malloc(sizeof(Ctx));
* // A timer is scheduled in order to resolve the promise
* ctx->timer = ecore_timer_add(122, _promise_resolve, ctx);
* // The _promise_cancel() will be used to clean ctx if the promise is canceled.
* ctx->p = eina_promise_new(scheduler, _promise_cancel, ctx);
* return ctx->p;
* }
* @endcode
*
* @{
*/
/**
* Value type for #Eina_Value's containing an #Eina_Promise
*/
eina: Rename EAPI macro to EINA_API in Eina library Summary: Patch from a series of patches to rename EAPI symbols to specific library DSOs. EAPI was designed to be able to pass `__attribute__ ((visibility ("default")))` for symbols with GCC, which would mean that even if -fvisibility=hidden was used when compiling the library, the needed symbols would get exported. MSVC __almost__ works like GCC (or mingw) in which you can declare everything as export and it will just work (slower, but it will work). But there's a caveat: global variables will not work the same way for MSVC, but works for mingw and GCC. For global variables (as opposed to functions), MSVC requires correct DSO visibility for MSVC: instead of declaring a symbol as export for everything, you need to declare it as import when importing from another DSO and export when defining it locally. With current EAPI definitions, we get the following example working in mingw and MSVC (observe it doesn't define any global variables as exported symbols). Example 1: dll1: ``` EAPI void foo(void); EAPI void bar() { foo(); } ``` dll2: ``` EAPI void foo() { printf ("foo\n"); } ``` This works fine with API defined as __declspec(dllexport) in both cases and for gcc defining as `__atttribute__((visibility("default")))` However, the following: Example 2: dll1: ``` EAPI extern int foo; EAPI void foobar(void); EAPI void bar() { foo = 5; foobar(); } ``` dll2: ``` EAPI int foo = 0; EAPI void foobar() { printf ("foo %d\n", foo); } ``` This will work on mingw but will not work for MSVC. And that's why EAPI is the only solution that worked for MSVC. Co-authored-by: João Paulo Taylor Ienczak Zanette <jpaulotiz@gmail.com> Co-authored-by: Ricardo Campos <ricardo.campos@expertise.dev> Co-authored-by: Lucas Cavalcante de Sousa <lucks.sousa@gmail.com> Reviewers: jptiz, lucas, woohyun, vtorri, raster Reviewed By: jptiz, lucas, vtorri Subscribers: ProhtMeyhet, cedric, #reviewers, #committers Tags: #efl Differential Revision: https://phab.enlightenment.org/D12188
2020-11-25 04:35:48 -08:00
EINA_API extern const Eina_Value_Type EINA_VALUE_TYPE_PROMISE;
/**
* Creates a new promise.
*
* This function creates a new promise which can be used to create a future
* using eina_future_new(). Every time a promise is created an #Eina_Promise_Cancel_Cb
* must be provided which is used to free resources that were created.
*
* A promise may be canceled directly by calling:
* @code
* eina_future_cancel(eina_future_new(eina_promise_new(...)))
* @endcode
* That is, canceling any future that is chained to receive the results.
*
* However promises can be canceled indirectly by other entities.
* These other entities will call eina_future_cancel() themselves,
* however you may not be aware of that. Some common sources
* of indirect cancellations:
*
* @li A subsystem was shutdown, canceling all pending futures (i.e.: ecore_shutdown())
*
* @li An EO object was linked to the promise or future, then if the object dies (last reference
* is gone), then the pending promises and futures will be canceled.
*
* @li Some other entity (library provider or library user) chained and canceled his future,
* which will result in your future being canceled.
*
* Since a promise may be canceled indirectly (by code sections that goes beyond your scope)
* you should always provide a cancel callback, even if you think you'll not need it.
*
* If you already have a value and want to create a future that will
* resolve to it directly use the eina_future_resolved(), it has the
* same effect as creating a promise and immediately resolving it.
*
* @param[in,out] scheduler The scheduler.
* @param[in] cancel_cb A callback used to inform that the promise was canceled. Use
* this callback to @c free @p data. @p cancel_cb must not be @c NULL !
* @param[in] data Data to @p cancel_cb.
* @return A promise or @c NULL on error.
* @see eina_future_cancel()
* @see eina_future_new()
* @see eina_promise_continue_new()
* @see eina_promise_resolve()
* @see eina_promise_reject()
* @see eina_promise_as_value()
* @see Eina_Future_Scheduler
* @see Eina_Future_Schedule_Entry
* @see Eina_Future_Scheduler_Cb
*/
eina: Rename EAPI macro to EINA_API in Eina library Summary: Patch from a series of patches to rename EAPI symbols to specific library DSOs. EAPI was designed to be able to pass `__attribute__ ((visibility ("default")))` for symbols with GCC, which would mean that even if -fvisibility=hidden was used when compiling the library, the needed symbols would get exported. MSVC __almost__ works like GCC (or mingw) in which you can declare everything as export and it will just work (slower, but it will work). But there's a caveat: global variables will not work the same way for MSVC, but works for mingw and GCC. For global variables (as opposed to functions), MSVC requires correct DSO visibility for MSVC: instead of declaring a symbol as export for everything, you need to declare it as import when importing from another DSO and export when defining it locally. With current EAPI definitions, we get the following example working in mingw and MSVC (observe it doesn't define any global variables as exported symbols). Example 1: dll1: ``` EAPI void foo(void); EAPI void bar() { foo(); } ``` dll2: ``` EAPI void foo() { printf ("foo\n"); } ``` This works fine with API defined as __declspec(dllexport) in both cases and for gcc defining as `__atttribute__((visibility("default")))` However, the following: Example 2: dll1: ``` EAPI extern int foo; EAPI void foobar(void); EAPI void bar() { foo = 5; foobar(); } ``` dll2: ``` EAPI int foo = 0; EAPI void foobar() { printf ("foo %d\n", foo); } ``` This will work on mingw but will not work for MSVC. And that's why EAPI is the only solution that worked for MSVC. Co-authored-by: João Paulo Taylor Ienczak Zanette <jpaulotiz@gmail.com> Co-authored-by: Ricardo Campos <ricardo.campos@expertise.dev> Co-authored-by: Lucas Cavalcante de Sousa <lucks.sousa@gmail.com> Reviewers: jptiz, lucas, woohyun, vtorri, raster Reviewed By: jptiz, lucas, vtorri Subscribers: ProhtMeyhet, cedric, #reviewers, #committers Tags: #efl Differential Revision: https://phab.enlightenment.org/D12188
2020-11-25 04:35:48 -08:00
EINA_API Eina_Promise *eina_promise_new(Eina_Future_Scheduler *scheduler, Eina_Promise_Cancel_Cb cancel_cb, const void *data) EINA_ARG_NONNULL(1, 2) EINA_WARN_UNUSED_RESULT;
/**
* Creates a new promise from a dead_future.
*
* This function creates a new promise from a future currently being resolved which can be
* used to create a #Eina_Value with eina_promise_as_value(). Every time a promise is
* created an #Eina_Promise_Cancel_Cb must be provided which is used to free resources
* that were created.
*
* A promise may be canceled directly by calling:
* @code
* eina_future_cancel(eina_future_new(eina_promise_new(...)))
* @endcode
* That is, canceling any future that is chained to receive the results.
*
* However promises can be canceled indirectly by other entities.
* These other entities will call eina_future_cancel() themselves,
* however you may not be aware of that. Some common sources
* of indirect cancellations:
*
* @li A subsystem was shutdown, canceling all pending futures (i.e.: ecore_shutdown())
*
* @li An EO object was linked to the promise or future, then if the object dies (last reference
* is gone), then the pending promises and futures will be canceled.
*
* @li Some other entity (library provider or library user) chained and canceled his future,
* which will result in your future being canceled.
*
* Since a promise may be canceled indirectly (by code sections that goes beyond your scope)
* you should always provide a cancel callback, even if you think you'll not need it.
*
* Here's a typical example:
*
* @code
* Eina_Value
* _future_resolve(void *data, const Eina_Value v, const Eina_Future *dead_future)
* {
* Eina_Promise *p;
* p = eina_promise_continue_new(dead_future, _promise_cancel, NULL);
* return eina_promise_as_value(p);
* }
* @endcode
*
* If you already have a value and want to create a future that will
* resolve to it directly use the eina_future_resolved(), it has the
* same effect as creating a promise and immediately resolving it.
*
* @note This function is to be used solely inside of a future resolve callback with
* the #Eina_Value being returned from it.
*
* @param[in] dead_future The future being resolved to get a scheduler from.
* @param[in] cancel_cb A callback used to inform that the promise was canceled. Use
* this callback to @c free @p data. @p cancel_cb must not be @c NULL.
* @param[in] data Data to @p cancel_cb.
* @return A promise or @c NULL on error.
* @see eina_future_cancel()
* @see eina_future_new()
* @see eina_promise_new()
* @see eina_promise_resolve()
* @see eina_promise_reject()
* @see eina_promise_as_value()
* @see Eina_Future_Scheduler
* @see Eina_Future_Schedule_Entry
* @see Eina_Future_Scheduler_Cb
*/
eina: Rename EAPI macro to EINA_API in Eina library Summary: Patch from a series of patches to rename EAPI symbols to specific library DSOs. EAPI was designed to be able to pass `__attribute__ ((visibility ("default")))` for symbols with GCC, which would mean that even if -fvisibility=hidden was used when compiling the library, the needed symbols would get exported. MSVC __almost__ works like GCC (or mingw) in which you can declare everything as export and it will just work (slower, but it will work). But there's a caveat: global variables will not work the same way for MSVC, but works for mingw and GCC. For global variables (as opposed to functions), MSVC requires correct DSO visibility for MSVC: instead of declaring a symbol as export for everything, you need to declare it as import when importing from another DSO and export when defining it locally. With current EAPI definitions, we get the following example working in mingw and MSVC (observe it doesn't define any global variables as exported symbols). Example 1: dll1: ``` EAPI void foo(void); EAPI void bar() { foo(); } ``` dll2: ``` EAPI void foo() { printf ("foo\n"); } ``` This works fine with API defined as __declspec(dllexport) in both cases and for gcc defining as `__atttribute__((visibility("default")))` However, the following: Example 2: dll1: ``` EAPI extern int foo; EAPI void foobar(void); EAPI void bar() { foo = 5; foobar(); } ``` dll2: ``` EAPI int foo = 0; EAPI void foobar() { printf ("foo %d\n", foo); } ``` This will work on mingw but will not work for MSVC. And that's why EAPI is the only solution that worked for MSVC. Co-authored-by: João Paulo Taylor Ienczak Zanette <jpaulotiz@gmail.com> Co-authored-by: Ricardo Campos <ricardo.campos@expertise.dev> Co-authored-by: Lucas Cavalcante de Sousa <lucks.sousa@gmail.com> Reviewers: jptiz, lucas, woohyun, vtorri, raster Reviewed By: jptiz, lucas, vtorri Subscribers: ProhtMeyhet, cedric, #reviewers, #committers Tags: #efl Differential Revision: https://phab.enlightenment.org/D12188
2020-11-25 04:35:48 -08:00
EINA_API Eina_Promise *eina_promise_continue_new(const Eina_Future *dead_future, Eina_Promise_Cancel_Cb cancel_cb, const void *data) EINA_ARG_NONNULL(1, 2) EINA_WARN_UNUSED_RESULT;
/**
* Resolves a promise.
*
* This function schedules a resolve event in a
* safe context (main loop or some platform defined safe context),
* whenever possible the future callbacks will be dispatched.
*
* @param[in,out] p A promise to resolve.
* @param[in] value The value to be delivered.
*
* Note that the @p value contents must survive this function scope,
* that is, do @b not use stack allocated blobs, arrays, structures or types that
* keep references to memory you give. Values will be automatically cleaned up
* using @c eina_value_flush() once they are unused (no more future or futures
* returned a new value).
*
* @see eina_promise_new()
* @see eina_promise_reject()
* @see eina_promise_as_value()
*/
eina: Rename EAPI macro to EINA_API in Eina library Summary: Patch from a series of patches to rename EAPI symbols to specific library DSOs. EAPI was designed to be able to pass `__attribute__ ((visibility ("default")))` for symbols with GCC, which would mean that even if -fvisibility=hidden was used when compiling the library, the needed symbols would get exported. MSVC __almost__ works like GCC (or mingw) in which you can declare everything as export and it will just work (slower, but it will work). But there's a caveat: global variables will not work the same way for MSVC, but works for mingw and GCC. For global variables (as opposed to functions), MSVC requires correct DSO visibility for MSVC: instead of declaring a symbol as export for everything, you need to declare it as import when importing from another DSO and export when defining it locally. With current EAPI definitions, we get the following example working in mingw and MSVC (observe it doesn't define any global variables as exported symbols). Example 1: dll1: ``` EAPI void foo(void); EAPI void bar() { foo(); } ``` dll2: ``` EAPI void foo() { printf ("foo\n"); } ``` This works fine with API defined as __declspec(dllexport) in both cases and for gcc defining as `__atttribute__((visibility("default")))` However, the following: Example 2: dll1: ``` EAPI extern int foo; EAPI void foobar(void); EAPI void bar() { foo = 5; foobar(); } ``` dll2: ``` EAPI int foo = 0; EAPI void foobar() { printf ("foo %d\n", foo); } ``` This will work on mingw but will not work for MSVC. And that's why EAPI is the only solution that worked for MSVC. Co-authored-by: João Paulo Taylor Ienczak Zanette <jpaulotiz@gmail.com> Co-authored-by: Ricardo Campos <ricardo.campos@expertise.dev> Co-authored-by: Lucas Cavalcante de Sousa <lucks.sousa@gmail.com> Reviewers: jptiz, lucas, woohyun, vtorri, raster Reviewed By: jptiz, lucas, vtorri Subscribers: ProhtMeyhet, cedric, #reviewers, #committers Tags: #efl Differential Revision: https://phab.enlightenment.org/D12188
2020-11-25 04:35:48 -08:00
EINA_API void eina_promise_resolve(Eina_Promise *p, Eina_Value value) EINA_ARG_NONNULL(1);
/**
* Rejects a promise.
*
* This function schedules a reject event in a
* safe context (main loop or some platform defined safe context),
* whenever possible the future callbacks will be dispatched.
*
* @param[in,out] p A promise to reject.
* @param[in] err An Eina_Error value
* @note Internally this function will create an #Eina_Value with type #EINA_VALUE_TYPE_ERROR.
*
* @see eina_promise_new()
* @see eina_promise_resolve()
* @see eina_promise_as_value()
*/
eina: Rename EAPI macro to EINA_API in Eina library Summary: Patch from a series of patches to rename EAPI symbols to specific library DSOs. EAPI was designed to be able to pass `__attribute__ ((visibility ("default")))` for symbols with GCC, which would mean that even if -fvisibility=hidden was used when compiling the library, the needed symbols would get exported. MSVC __almost__ works like GCC (or mingw) in which you can declare everything as export and it will just work (slower, but it will work). But there's a caveat: global variables will not work the same way for MSVC, but works for mingw and GCC. For global variables (as opposed to functions), MSVC requires correct DSO visibility for MSVC: instead of declaring a symbol as export for everything, you need to declare it as import when importing from another DSO and export when defining it locally. With current EAPI definitions, we get the following example working in mingw and MSVC (observe it doesn't define any global variables as exported symbols). Example 1: dll1: ``` EAPI void foo(void); EAPI void bar() { foo(); } ``` dll2: ``` EAPI void foo() { printf ("foo\n"); } ``` This works fine with API defined as __declspec(dllexport) in both cases and for gcc defining as `__atttribute__((visibility("default")))` However, the following: Example 2: dll1: ``` EAPI extern int foo; EAPI void foobar(void); EAPI void bar() { foo = 5; foobar(); } ``` dll2: ``` EAPI int foo = 0; EAPI void foobar() { printf ("foo %d\n", foo); } ``` This will work on mingw but will not work for MSVC. And that's why EAPI is the only solution that worked for MSVC. Co-authored-by: João Paulo Taylor Ienczak Zanette <jpaulotiz@gmail.com> Co-authored-by: Ricardo Campos <ricardo.campos@expertise.dev> Co-authored-by: Lucas Cavalcante de Sousa <lucks.sousa@gmail.com> Reviewers: jptiz, lucas, woohyun, vtorri, raster Reviewed By: jptiz, lucas, vtorri Subscribers: ProhtMeyhet, cedric, #reviewers, #committers Tags: #efl Differential Revision: https://phab.enlightenment.org/D12188
2020-11-25 04:35:48 -08:00
EINA_API void eina_promise_reject(Eina_Promise *p, Eina_Error err) EINA_ARG_NONNULL(1);
/**
* @}
*/
/**
* @defgroup Eina_Future Eina Futures
* @ingroup Eina_Promise
* @brief Methods and structures dealing with #Eina_Future
* @{
*/
/**
* Cancels a future.
*
* This function will cancel the whole future chain immediately (it will not be schedule to the next mainloop pass)
* and it will also cancel the promise linked against it. The #Eina_Future_Cb will be called
* with an #Eina_Value typed as #EINA_VALUE_TYPE_ERROR, with its value set to @c ECANCELED
* @param[in,out] f The future to cancel.
*/
eina: Rename EAPI macro to EINA_API in Eina library Summary: Patch from a series of patches to rename EAPI symbols to specific library DSOs. EAPI was designed to be able to pass `__attribute__ ((visibility ("default")))` for symbols with GCC, which would mean that even if -fvisibility=hidden was used when compiling the library, the needed symbols would get exported. MSVC __almost__ works like GCC (or mingw) in which you can declare everything as export and it will just work (slower, but it will work). But there's a caveat: global variables will not work the same way for MSVC, but works for mingw and GCC. For global variables (as opposed to functions), MSVC requires correct DSO visibility for MSVC: instead of declaring a symbol as export for everything, you need to declare it as import when importing from another DSO and export when defining it locally. With current EAPI definitions, we get the following example working in mingw and MSVC (observe it doesn't define any global variables as exported symbols). Example 1: dll1: ``` EAPI void foo(void); EAPI void bar() { foo(); } ``` dll2: ``` EAPI void foo() { printf ("foo\n"); } ``` This works fine with API defined as __declspec(dllexport) in both cases and for gcc defining as `__atttribute__((visibility("default")))` However, the following: Example 2: dll1: ``` EAPI extern int foo; EAPI void foobar(void); EAPI void bar() { foo = 5; foobar(); } ``` dll2: ``` EAPI int foo = 0; EAPI void foobar() { printf ("foo %d\n", foo); } ``` This will work on mingw but will not work for MSVC. And that's why EAPI is the only solution that worked for MSVC. Co-authored-by: João Paulo Taylor Ienczak Zanette <jpaulotiz@gmail.com> Co-authored-by: Ricardo Campos <ricardo.campos@expertise.dev> Co-authored-by: Lucas Cavalcante de Sousa <lucks.sousa@gmail.com> Reviewers: jptiz, lucas, woohyun, vtorri, raster Reviewed By: jptiz, lucas, vtorri Subscribers: ProhtMeyhet, cedric, #reviewers, #committers Tags: #efl Differential Revision: https://phab.enlightenment.org/D12188
2020-11-25 04:35:48 -08:00
EINA_API void eina_future_cancel(Eina_Future *f) EINA_ARG_NONNULL(1);
/**
* Flushes an #Eina_Future_Desc
*
* This function is mainly used by bindings to flush an #Eina_Future_Desc.
* It will call the #Eina_Future_Cb with @c ENOMEM and zero the @p desc contents.
*
* @param[in,out] desc The #Eina_Future_Desc to flush, if @c NULL this is a noop.
*/
eina: Rename EAPI macro to EINA_API in Eina library Summary: Patch from a series of patches to rename EAPI symbols to specific library DSOs. EAPI was designed to be able to pass `__attribute__ ((visibility ("default")))` for symbols with GCC, which would mean that even if -fvisibility=hidden was used when compiling the library, the needed symbols would get exported. MSVC __almost__ works like GCC (or mingw) in which you can declare everything as export and it will just work (slower, but it will work). But there's a caveat: global variables will not work the same way for MSVC, but works for mingw and GCC. For global variables (as opposed to functions), MSVC requires correct DSO visibility for MSVC: instead of declaring a symbol as export for everything, you need to declare it as import when importing from another DSO and export when defining it locally. With current EAPI definitions, we get the following example working in mingw and MSVC (observe it doesn't define any global variables as exported symbols). Example 1: dll1: ``` EAPI void foo(void); EAPI void bar() { foo(); } ``` dll2: ``` EAPI void foo() { printf ("foo\n"); } ``` This works fine with API defined as __declspec(dllexport) in both cases and for gcc defining as `__atttribute__((visibility("default")))` However, the following: Example 2: dll1: ``` EAPI extern int foo; EAPI void foobar(void); EAPI void bar() { foo = 5; foobar(); } ``` dll2: ``` EAPI int foo = 0; EAPI void foobar() { printf ("foo %d\n", foo); } ``` This will work on mingw but will not work for MSVC. And that's why EAPI is the only solution that worked for MSVC. Co-authored-by: João Paulo Taylor Ienczak Zanette <jpaulotiz@gmail.com> Co-authored-by: Ricardo Campos <ricardo.campos@expertise.dev> Co-authored-by: Lucas Cavalcante de Sousa <lucks.sousa@gmail.com> Reviewers: jptiz, lucas, woohyun, vtorri, raster Reviewed By: jptiz, lucas, vtorri Subscribers: ProhtMeyhet, cedric, #reviewers, #committers Tags: #efl Differential Revision: https://phab.enlightenment.org/D12188
2020-11-25 04:35:48 -08:00
EINA_API void eina_future_desc_flush(Eina_Future_Desc *desc);
/**
* Flushes an #Eina_Future_Cb_Easy_Desc
*
* This function is mainly used by bindings to flush an #Eina_Future_Cb_Easy_Desc.
* It will call the #Eina_Future_Error_Cb with @c ENOMEM, the #Eina_Future_Free_Cb and
* zero the @p desc contents.
*
* @param[in,out] desc The #Eina_Future_Cb_Easy_Desc to flush, if @c NULL this is a noop.
*/
eina: Rename EAPI macro to EINA_API in Eina library Summary: Patch from a series of patches to rename EAPI symbols to specific library DSOs. EAPI was designed to be able to pass `__attribute__ ((visibility ("default")))` for symbols with GCC, which would mean that even if -fvisibility=hidden was used when compiling the library, the needed symbols would get exported. MSVC __almost__ works like GCC (or mingw) in which you can declare everything as export and it will just work (slower, but it will work). But there's a caveat: global variables will not work the same way for MSVC, but works for mingw and GCC. For global variables (as opposed to functions), MSVC requires correct DSO visibility for MSVC: instead of declaring a symbol as export for everything, you need to declare it as import when importing from another DSO and export when defining it locally. With current EAPI definitions, we get the following example working in mingw and MSVC (observe it doesn't define any global variables as exported symbols). Example 1: dll1: ``` EAPI void foo(void); EAPI void bar() { foo(); } ``` dll2: ``` EAPI void foo() { printf ("foo\n"); } ``` This works fine with API defined as __declspec(dllexport) in both cases and for gcc defining as `__atttribute__((visibility("default")))` However, the following: Example 2: dll1: ``` EAPI extern int foo; EAPI void foobar(void); EAPI void bar() { foo = 5; foobar(); } ``` dll2: ``` EAPI int foo = 0; EAPI void foobar() { printf ("foo %d\n", foo); } ``` This will work on mingw but will not work for MSVC. And that's why EAPI is the only solution that worked for MSVC. Co-authored-by: João Paulo Taylor Ienczak Zanette <jpaulotiz@gmail.com> Co-authored-by: Ricardo Campos <ricardo.campos@expertise.dev> Co-authored-by: Lucas Cavalcante de Sousa <lucks.sousa@gmail.com> Reviewers: jptiz, lucas, woohyun, vtorri, raster Reviewed By: jptiz, lucas, vtorri Subscribers: ProhtMeyhet, cedric, #reviewers, #committers Tags: #efl Differential Revision: https://phab.enlightenment.org/D12188
2020-11-25 04:35:48 -08:00
EINA_API void eina_future_cb_easy_desc_flush(Eina_Future_Cb_Easy_Desc *desc);
/**
* Creates a new #Eina_Value from a promise.
*
* This function creates a new #Eina_Value that will store an #Eina_Promise
* in it. This function is useful for dealing with promises inside
* an #Eina_Future_Cb. By returning a promise inside the #Eina_Value the
* whole chain will wait until the promise is resolved in
* order to continue its execution. Example:
*
* @code
* static Eina_Value
* _file_data_ready(const *data EINA_UNUSED, const Eina_Value v, const Eina_Future *dead EINA_UNUSED)
* {
* const char *file_data;
* Eina_Promise *p;
* // It was not possible to fetch the file size.
* if (v.type == EINA_VALUE_TYPE_ERROR)
* {
* Eina_Error err;
* eina_value_get(&v, &err);
* fprintf(stderr, "Could not get the file data. Reason: %s\n", eina_error_msg_get(err));
* ecore_main_loop_quit();
* return v;
* }
* else if (v.type != EINA_VALUE_TYPE_STRING)
* {
* fprintf(stderr, "Expecting type '%s' - received '%s'\n", EINA_VALUE_TYPE_STRING->name, v.type->name);
* ecore_main_loop_quit();
* return v;
* }
*
* eina_value_get(&v, &file_data);
* // count_words will count the words in the background and resolve the promise once it is over...
* p = count_words(file_data);
* return eina_promise_as_value(p);
* }
*
* static Eina_Value
* _word_count_ready(const *data EINA_UNUSED, const Eina_Value v, const Eina_Future *dead EINA_UNUSED)
* {
* // The _word_count_ready will only be called once count_words() resolves/rejects the promise returned by _file_data_ready()
* int count;
* if (v.type == EINA_VALUE_TYPE_ERROR)
* {
* Eina_Error err;
* eina_value_get(&v, &err);
* fprintf(stderr, "Could not count the file words. Reason: %s\n", eina_error_msg_get(err));
* ecore_main_loop_quit();
* return v;
* }
* else if (v.type != EINA_VALUE_TYPE_INT)
* {
* fprintf(stderr, "Expecting type '%s' - received '%s'\n", EINA_VALUE_TYPE_INT->name, v.type->name);
* ecore_main_loop_quit();
* return v;
* }
* eina_value_get(&v, &count);
* printf("File word count %d\n", count);
* return v;
* }
*
* void
* file_operation(void)
* {
* Eina_Future *f = get_file_data("/MyFile.txt");
* eina_future_chain(f, {.cb = _file_data_ready, .data = NULL},
* {.cb = _word_count_ready, .data = NULL});
* }
* @endcode
*
* @param[in,out] p The promise object to wrap in a value.
* @return An #Eina_Value. On error the value's type will be @c NULL.
*
* @note If an error happens the promise will be CANCELED.
* @see eina_future_as_value()
* @see eina_promise_reject()
* @see eina_promise_resolve()
*/
eina: Rename EAPI macro to EINA_API in Eina library Summary: Patch from a series of patches to rename EAPI symbols to specific library DSOs. EAPI was designed to be able to pass `__attribute__ ((visibility ("default")))` for symbols with GCC, which would mean that even if -fvisibility=hidden was used when compiling the library, the needed symbols would get exported. MSVC __almost__ works like GCC (or mingw) in which you can declare everything as export and it will just work (slower, but it will work). But there's a caveat: global variables will not work the same way for MSVC, but works for mingw and GCC. For global variables (as opposed to functions), MSVC requires correct DSO visibility for MSVC: instead of declaring a symbol as export for everything, you need to declare it as import when importing from another DSO and export when defining it locally. With current EAPI definitions, we get the following example working in mingw and MSVC (observe it doesn't define any global variables as exported symbols). Example 1: dll1: ``` EAPI void foo(void); EAPI void bar() { foo(); } ``` dll2: ``` EAPI void foo() { printf ("foo\n"); } ``` This works fine with API defined as __declspec(dllexport) in both cases and for gcc defining as `__atttribute__((visibility("default")))` However, the following: Example 2: dll1: ``` EAPI extern int foo; EAPI void foobar(void); EAPI void bar() { foo = 5; foobar(); } ``` dll2: ``` EAPI int foo = 0; EAPI void foobar() { printf ("foo %d\n", foo); } ``` This will work on mingw but will not work for MSVC. And that's why EAPI is the only solution that worked for MSVC. Co-authored-by: João Paulo Taylor Ienczak Zanette <jpaulotiz@gmail.com> Co-authored-by: Ricardo Campos <ricardo.campos@expertise.dev> Co-authored-by: Lucas Cavalcante de Sousa <lucks.sousa@gmail.com> Reviewers: jptiz, lucas, woohyun, vtorri, raster Reviewed By: jptiz, lucas, vtorri Subscribers: ProhtMeyhet, cedric, #reviewers, #committers Tags: #efl Differential Revision: https://phab.enlightenment.org/D12188
2020-11-25 04:35:48 -08:00
EINA_API Eina_Value eina_promise_as_value(Eina_Promise *p) EINA_ARG_NONNULL(1) EINA_WARN_UNUSED_RESULT;
/**
* Creates an #Eina_Value from a future.
*
* This function is used for the same purposes as eina_promise_as_value(),
* but receives an #Eina_Future instead.
*
* @param[in,out] f A future to create a Eina_Value from.
* @return An Eina_Value. On error the value's type will be @c NULL.
* @note If an error happens the future @p f will be CANCELED
* @see eina_promise_as_value()
*/
eina: Rename EAPI macro to EINA_API in Eina library Summary: Patch from a series of patches to rename EAPI symbols to specific library DSOs. EAPI was designed to be able to pass `__attribute__ ((visibility ("default")))` for symbols with GCC, which would mean that even if -fvisibility=hidden was used when compiling the library, the needed symbols would get exported. MSVC __almost__ works like GCC (or mingw) in which you can declare everything as export and it will just work (slower, but it will work). But there's a caveat: global variables will not work the same way for MSVC, but works for mingw and GCC. For global variables (as opposed to functions), MSVC requires correct DSO visibility for MSVC: instead of declaring a symbol as export for everything, you need to declare it as import when importing from another DSO and export when defining it locally. With current EAPI definitions, we get the following example working in mingw and MSVC (observe it doesn't define any global variables as exported symbols). Example 1: dll1: ``` EAPI void foo(void); EAPI void bar() { foo(); } ``` dll2: ``` EAPI void foo() { printf ("foo\n"); } ``` This works fine with API defined as __declspec(dllexport) in both cases and for gcc defining as `__atttribute__((visibility("default")))` However, the following: Example 2: dll1: ``` EAPI extern int foo; EAPI void foobar(void); EAPI void bar() { foo = 5; foobar(); } ``` dll2: ``` EAPI int foo = 0; EAPI void foobar() { printf ("foo %d\n", foo); } ``` This will work on mingw but will not work for MSVC. And that's why EAPI is the only solution that worked for MSVC. Co-authored-by: João Paulo Taylor Ienczak Zanette <jpaulotiz@gmail.com> Co-authored-by: Ricardo Campos <ricardo.campos@expertise.dev> Co-authored-by: Lucas Cavalcante de Sousa <lucks.sousa@gmail.com> Reviewers: jptiz, lucas, woohyun, vtorri, raster Reviewed By: jptiz, lucas, vtorri Subscribers: ProhtMeyhet, cedric, #reviewers, #committers Tags: #efl Differential Revision: https://phab.enlightenment.org/D12188
2020-11-25 04:35:48 -08:00
EINA_API Eina_Value eina_future_as_value(Eina_Future *f)EINA_ARG_NONNULL(1) EINA_WARN_UNUSED_RESULT;
/**
* Creates a new future.
*
* This function creates a new future and can be used to report
* that an operation has succeeded or failed using
* eina_promise_resolve() or eina_promise_reject().
*
* Futures can also be canceled using eina_future_cancel(), which
* will cause the whole chain to be canceled alongside with any pending promise.
*
* @note A promise can only have one future attached to it, calling
* eina_future_new() on the same promise twice will
* result in an error (@c NULL is returned) and the future
* attached to the promise will be canceled!
*
* @param[in,out] p A promise used to attach a future. May not be @c NULL.
* @return The future or @c NULL on error.
* @note If an error happens this function will CANCEL the promise.
* @see eina_promise_new()
* @see eina_promise_reject()
* @see eina_promise_resolve()
* @see eina_future_cancel()
*/
eina: Rename EAPI macro to EINA_API in Eina library Summary: Patch from a series of patches to rename EAPI symbols to specific library DSOs. EAPI was designed to be able to pass `__attribute__ ((visibility ("default")))` for symbols with GCC, which would mean that even if -fvisibility=hidden was used when compiling the library, the needed symbols would get exported. MSVC __almost__ works like GCC (or mingw) in which you can declare everything as export and it will just work (slower, but it will work). But there's a caveat: global variables will not work the same way for MSVC, but works for mingw and GCC. For global variables (as opposed to functions), MSVC requires correct DSO visibility for MSVC: instead of declaring a symbol as export for everything, you need to declare it as import when importing from another DSO and export when defining it locally. With current EAPI definitions, we get the following example working in mingw and MSVC (observe it doesn't define any global variables as exported symbols). Example 1: dll1: ``` EAPI void foo(void); EAPI void bar() { foo(); } ``` dll2: ``` EAPI void foo() { printf ("foo\n"); } ``` This works fine with API defined as __declspec(dllexport) in both cases and for gcc defining as `__atttribute__((visibility("default")))` However, the following: Example 2: dll1: ``` EAPI extern int foo; EAPI void foobar(void); EAPI void bar() { foo = 5; foobar(); } ``` dll2: ``` EAPI int foo = 0; EAPI void foobar() { printf ("foo %d\n", foo); } ``` This will work on mingw but will not work for MSVC. And that's why EAPI is the only solution that worked for MSVC. Co-authored-by: João Paulo Taylor Ienczak Zanette <jpaulotiz@gmail.com> Co-authored-by: Ricardo Campos <ricardo.campos@expertise.dev> Co-authored-by: Lucas Cavalcante de Sousa <lucks.sousa@gmail.com> Reviewers: jptiz, lucas, woohyun, vtorri, raster Reviewed By: jptiz, lucas, vtorri Subscribers: ProhtMeyhet, cedric, #reviewers, #committers Tags: #efl Differential Revision: https://phab.enlightenment.org/D12188
2020-11-25 04:35:48 -08:00
EINA_API Eina_Future *eina_future_new(Eina_Promise *p) EINA_ARG_NONNULL(1) EINA_WARN_UNUSED_RESULT;
/**
* Creates a new future that is already resolved to a value.
*
* This function creates a new future with an already known value,
* that will be resolved and dispatched by the given @p scheduler as
* usual.
*
* This is a helper that behaves the same as eina_promise_new()
* followed by eina_future_new() and then eina_promise_resolve().
*
* Futures can also be canceled using eina_future_cancel(), which will
* cause the whole chain to be canceled alongside with any pending
* promise.
*
* @param[in,out] scheduler The scheduler to use.
* @param[in] value The value to be delivered. Note that the value
* contents must survive this function scope, that is, do @b not use
* stack allocated blobs, arrays, structures or types that keep
* references to memory you give. Values will be automatically cleaned
* up using @c eina_value_flush() once they are unused (no more future
* or futures returned a new value).
*
* @return The future or @c NULL on error.
*
* @see eina_promise_new()
* @see eina_future_new()
* @see eina_promise_reject()
* @see eina_promise_resolve()
* @see eina_future_cancel()
*/
eina: Rename EAPI macro to EINA_API in Eina library Summary: Patch from a series of patches to rename EAPI symbols to specific library DSOs. EAPI was designed to be able to pass `__attribute__ ((visibility ("default")))` for symbols with GCC, which would mean that even if -fvisibility=hidden was used when compiling the library, the needed symbols would get exported. MSVC __almost__ works like GCC (or mingw) in which you can declare everything as export and it will just work (slower, but it will work). But there's a caveat: global variables will not work the same way for MSVC, but works for mingw and GCC. For global variables (as opposed to functions), MSVC requires correct DSO visibility for MSVC: instead of declaring a symbol as export for everything, you need to declare it as import when importing from another DSO and export when defining it locally. With current EAPI definitions, we get the following example working in mingw and MSVC (observe it doesn't define any global variables as exported symbols). Example 1: dll1: ``` EAPI void foo(void); EAPI void bar() { foo(); } ``` dll2: ``` EAPI void foo() { printf ("foo\n"); } ``` This works fine with API defined as __declspec(dllexport) in both cases and for gcc defining as `__atttribute__((visibility("default")))` However, the following: Example 2: dll1: ``` EAPI extern int foo; EAPI void foobar(void); EAPI void bar() { foo = 5; foobar(); } ``` dll2: ``` EAPI int foo = 0; EAPI void foobar() { printf ("foo %d\n", foo); } ``` This will work on mingw but will not work for MSVC. And that's why EAPI is the only solution that worked for MSVC. Co-authored-by: João Paulo Taylor Ienczak Zanette <jpaulotiz@gmail.com> Co-authored-by: Ricardo Campos <ricardo.campos@expertise.dev> Co-authored-by: Lucas Cavalcante de Sousa <lucks.sousa@gmail.com> Reviewers: jptiz, lucas, woohyun, vtorri, raster Reviewed By: jptiz, lucas, vtorri Subscribers: ProhtMeyhet, cedric, #reviewers, #committers Tags: #efl Differential Revision: https://phab.enlightenment.org/D12188
2020-11-25 04:35:48 -08:00
EINA_API Eina_Future *eina_future_resolved(Eina_Future_Scheduler *scheduler, Eina_Value value) EINA_ARG_NONNULL(1);
/**
* Creates a new future that is already rejected to a specified error.
*
* This function creates a new future with an already known error,
* that will be resolved and dispatched by the given @p scheduler as
* usual.
*
* This is a helper that behaves the same as eina_promise_new()
* followed by eina_future_new() and then eina_promise_reject().
*
* Futures can also be canceled using eina_future_cancel(), which will
* cause the whole chain to be canceled alongside with any pending
* promise.
*
* @param[in,out] scheduler The scheduler to use.
* @param[in] err An #Eina_Error value
*
* @return The future or @c NULL on error.
*
* @see eina_promise_new()
* @see eina_future_new()
* @see eina_promise_reject()
* @see eina_promise_resolve()
* @see eina_future_cancel()
*/
eina: Rename EAPI macro to EINA_API in Eina library Summary: Patch from a series of patches to rename EAPI symbols to specific library DSOs. EAPI was designed to be able to pass `__attribute__ ((visibility ("default")))` for symbols with GCC, which would mean that even if -fvisibility=hidden was used when compiling the library, the needed symbols would get exported. MSVC __almost__ works like GCC (or mingw) in which you can declare everything as export and it will just work (slower, but it will work). But there's a caveat: global variables will not work the same way for MSVC, but works for mingw and GCC. For global variables (as opposed to functions), MSVC requires correct DSO visibility for MSVC: instead of declaring a symbol as export for everything, you need to declare it as import when importing from another DSO and export when defining it locally. With current EAPI definitions, we get the following example working in mingw and MSVC (observe it doesn't define any global variables as exported symbols). Example 1: dll1: ``` EAPI void foo(void); EAPI void bar() { foo(); } ``` dll2: ``` EAPI void foo() { printf ("foo\n"); } ``` This works fine with API defined as __declspec(dllexport) in both cases and for gcc defining as `__atttribute__((visibility("default")))` However, the following: Example 2: dll1: ``` EAPI extern int foo; EAPI void foobar(void); EAPI void bar() { foo = 5; foobar(); } ``` dll2: ``` EAPI int foo = 0; EAPI void foobar() { printf ("foo %d\n", foo); } ``` This will work on mingw but will not work for MSVC. And that's why EAPI is the only solution that worked for MSVC. Co-authored-by: João Paulo Taylor Ienczak Zanette <jpaulotiz@gmail.com> Co-authored-by: Ricardo Campos <ricardo.campos@expertise.dev> Co-authored-by: Lucas Cavalcante de Sousa <lucks.sousa@gmail.com> Reviewers: jptiz, lucas, woohyun, vtorri, raster Reviewed By: jptiz, lucas, vtorri Subscribers: ProhtMeyhet, cedric, #reviewers, #committers Tags: #efl Differential Revision: https://phab.enlightenment.org/D12188
2020-11-25 04:35:48 -08:00
EINA_API Eina_Future *eina_future_rejected(Eina_Future_Scheduler *scheduler, Eina_Error err);
/**
* Register an #Eina_Future_Desc to be used when the future is resolved/rejected.
*
* With this function a callback and data is attached to the future and then
* once it's resolved or rejected the callback will be informed.
*
* If during the future creation an error happens this function will return @c NULL,
* and the Eina_Future_Desc::cb will be called reporting an error (@c EINVAL or @c ENOMEM),
* so the user has a chance to free Eina_Future_Desc::data.
*
* In case a future in the chain is canceled, the whole chain will be canceled immediately
* and the error @c ECANCELED will be reported.
*
* Here's a simple usage of this function.
*
* @code
* static Eina_Value
* _file_ready(const *data, const Eina_Value v, const Eina_Future *dead EINA_UNUSED)
* {
* Ctx *ctx = data;
* // It was not possible to fetch the file size.
* if (v.type == EINA_VALUE_TYPE_ERROR)
* {
* Eina_Error err;
* eina_value_get(&v, &err);
* fprintf(stderr, "Could not read the file size. Reason: %s\n", eina_error_msg_get(err));
* ecore_main_loop_quit();
* return v;
* }
* else if (v.type != EINA_VALUE_TYPE_INT)
* {
* fprintf(stderr, "Expecting type '%s' - received '%s'\n", EINA_VALUE_TYPE_INT->name, v.type->name);
* ecore_main_loop_quit();
* return v;
* }
* eina_value_get(&v, &ctx->size);
* printf("File size is %d bytes\n", ctx->size);
* return v;
* }
*
* void
* file_operation(void)
* {
* Eina_Future *f = get_file_size_async("/MyFile.txt");
* eina_future_then_from_desc(f, (const Eina_Future_Desc){.cb = _size_ready, .data = NULL});
* // There's a helper macro called eina_future_then() which simplifies the usage.
* // The code below has the same effect.
* // eina_future_then(f, _size_ready, NULL);
* }
* @endcode
*
* Although the code presented at `_size_ready()` is very simple, most of it
* is just used to check the #Eina_Value type. In order
* to avoid this type of code the function eina_future_cb_easy_from_desc()
* was created. Please, check its documentation for more information.
*
* This function can also be used to create a future chain, making
* it possible to execute the future result in a cascade order. When
* using a future chain the #Eina_Value returned by a Eina_Future_Desc::cb
* will be delivered to the next Eina_Future_Desc::cb in the chain.
*
* Here's an example:
*
* @code
* static Eina_Value
* _future_cb1(const *data EINA_UNUSED, const Eina_Value v)
* {
* Eina_Value new_v;
* int i;
*
* // There's no need to check the Eina_Value type since we're using eina_future_cb_easy()
* eina_value_get(&v, &i);
* printf("File size as int: %d\n", i);
* eina_value_setup(&new_v, EINA_VALUE_TYPE_STRING);
* // Convert the file size to string
* eina_value_convert(&v, &new_v);
* return new_v;
* }
*
* static Eina_Value
* _future_cb2(const *data EINA_UNUSED, const Eina_Value v)
* {
* Eina_Value new_v;
* const char *file_size_str;
*
* // There's no need to check the Eina_Value type since we're using eina_future_cb_easy()
* eina_value_get(&v, &file_size_str);
* printf("File size as string: %s\n", file_size_str);
* eina_value_setup(&new_v, EINA_VALUE_TYPE_DOUBLE);
* eina_value_convert(&v, &new_v);
* return new_v;
* }
*
* static Eina_Value
* _future_cb3(const *data EINA_UNUSED, const Eina_Value v)
* {
* double d;
*
* // There's no need to check the Eina_Value type since we're using eina_future_cb_easy()
* eina_value_get(&v, &d);
* printf("File size as double: %g\n", d);
* return v;
* }
*
* static Eina_Value
* _future_err(void *data EINA_UNUSED, Eina_Error err)
* {
* // This function is called if future result type does not match or another error occurred
* Eina_Value new_v;
* eina_value_setup(&new_v, EINA_VALUE_TYPE_ERROR);
* eina_value_set(&new_v, err);
* fprintf(stderr, "Error during future process. Reason: %s\n", eina_error_msg_get(err));
* // Pass the error to the next future in the chain..
* return new_v;
* }
*
* void chain(void)
* {
* Eina_Future *f = get_file_size_async("/MyFile.txt");
* f = eina_future_then_easy(f, .success = _future_cb1, .success_type = EINA_VALUE_TYPE_INT);
* // _future_cb2 will be executed after _future_cb1()
* f = eina_future_then_easy(f, .success = _future_cb2, .success_type = EINA_VALUE_TYPE_STRING);
2017-12-22 03:31:09 -08:00
* // _future_cb3 will be executed after _future_cb2()
* f = eina_future_then_easy(f, .success = _future_cb3, .success_type = EINA_VALUE_TYPE_DOUBLE);
* // If an error happens _future_err will be called
* eina_future_then_easy(f, .error = _future_err);
* }
* @endcode
*
* Although it's possible to create a future chain using eina_future_then() and eina_future_then_from_desc()
* there's a function that makes this task much easier, please check eina_future_chain_array() for more
* information.
* @note This example does manual conversion and print, however Eina offers
* eina_future_cb_convert_to() and eina_future_cb_console_from_desc() to make those common case easier.
*
* @param[in,out] prev A future to link against
* @param[in] desc A description struct containing the callback and data.
* @return A new future or @c NULL on error.
* @note If an error happens the whole future chain will CANCELED and
* desc.cb will be called in order to free desc.data.
* @see eina_future_new()
* @see eina_future_then()
* @see Eina_Future_Desc
* @see eina_future_chain_array()
* @see eina_future_chain()
* @see eina_future_cb_console_from_desc()
* @see eina_future_cb_easy_from_desc()
* @see eina_future_cb_easy()
* @see eina_future_cb_convert_to()
* @see eina_future_cancel()
* @see eina_future_then_easy()
* @see eina_future_cb_log_from_desc()
*/
eina: Rename EAPI macro to EINA_API in Eina library Summary: Patch from a series of patches to rename EAPI symbols to specific library DSOs. EAPI was designed to be able to pass `__attribute__ ((visibility ("default")))` for symbols with GCC, which would mean that even if -fvisibility=hidden was used when compiling the library, the needed symbols would get exported. MSVC __almost__ works like GCC (or mingw) in which you can declare everything as export and it will just work (slower, but it will work). But there's a caveat: global variables will not work the same way for MSVC, but works for mingw and GCC. For global variables (as opposed to functions), MSVC requires correct DSO visibility for MSVC: instead of declaring a symbol as export for everything, you need to declare it as import when importing from another DSO and export when defining it locally. With current EAPI definitions, we get the following example working in mingw and MSVC (observe it doesn't define any global variables as exported symbols). Example 1: dll1: ``` EAPI void foo(void); EAPI void bar() { foo(); } ``` dll2: ``` EAPI void foo() { printf ("foo\n"); } ``` This works fine with API defined as __declspec(dllexport) in both cases and for gcc defining as `__atttribute__((visibility("default")))` However, the following: Example 2: dll1: ``` EAPI extern int foo; EAPI void foobar(void); EAPI void bar() { foo = 5; foobar(); } ``` dll2: ``` EAPI int foo = 0; EAPI void foobar() { printf ("foo %d\n", foo); } ``` This will work on mingw but will not work for MSVC. And that's why EAPI is the only solution that worked for MSVC. Co-authored-by: João Paulo Taylor Ienczak Zanette <jpaulotiz@gmail.com> Co-authored-by: Ricardo Campos <ricardo.campos@expertise.dev> Co-authored-by: Lucas Cavalcante de Sousa <lucks.sousa@gmail.com> Reviewers: jptiz, lucas, woohyun, vtorri, raster Reviewed By: jptiz, lucas, vtorri Subscribers: ProhtMeyhet, cedric, #reviewers, #committers Tags: #efl Differential Revision: https://phab.enlightenment.org/D12188
2020-11-25 04:35:48 -08:00
EINA_API Eina_Future *eina_future_then_from_desc(Eina_Future *prev, const Eina_Future_Desc desc) EINA_ARG_NONNULL(1);
/**
* Creates an Eina_Future_Desc that will log the previous future resolved value.
*
* This function can be used to quickly log the value that an Eina_Future_Desc::cb
* is returning. The returned value will be passed to the next future in the chain without
* modifications.
*
* There are some helper macros like eina_future_cb_log_dbg() which will automatically
* fill the following fields:
*
* @li Eina_Future_Cb_Log_Desc::file: The `__FILE__` function will be used.
* @li Eina_Future_Cb_Log_Desc::func: The `__func__` macro will be used.
* @li Eina_Future_Cb_Log_Desc::level: `EINA_LOG_LEVEL_DBG` will be used.
* @li Eina_Future_Cb_Log_Desc::domain: `EINA_LOG_DOMAIN_DEFAULT` will be used.
* @li Eina_Future_Cb_Log_Desc::line: The `__LINE__` macro will be used.
*
* @param[in] desc The description data to be used.
* @return An #Eina_Future_Desc
*
* @see Eina_Future_Cb_Log_Desc
* @see eina_future_then()
* @see eina_future_chain()
* @see eina_future_cb_log_dbg()
* @see eina_future_cb_log_crit()
* @see eina_future_cb_log_err()
* @see eina_future_cb_log_info()
* @see eina_future_cb_log_warn()
* @see eina_future_cb_console_from_desc()
*/
eina: Rename EAPI macro to EINA_API in Eina library Summary: Patch from a series of patches to rename EAPI symbols to specific library DSOs. EAPI was designed to be able to pass `__attribute__ ((visibility ("default")))` for symbols with GCC, which would mean that even if -fvisibility=hidden was used when compiling the library, the needed symbols would get exported. MSVC __almost__ works like GCC (or mingw) in which you can declare everything as export and it will just work (slower, but it will work). But there's a caveat: global variables will not work the same way for MSVC, but works for mingw and GCC. For global variables (as opposed to functions), MSVC requires correct DSO visibility for MSVC: instead of declaring a symbol as export for everything, you need to declare it as import when importing from another DSO and export when defining it locally. With current EAPI definitions, we get the following example working in mingw and MSVC (observe it doesn't define any global variables as exported symbols). Example 1: dll1: ``` EAPI void foo(void); EAPI void bar() { foo(); } ``` dll2: ``` EAPI void foo() { printf ("foo\n"); } ``` This works fine with API defined as __declspec(dllexport) in both cases and for gcc defining as `__atttribute__((visibility("default")))` However, the following: Example 2: dll1: ``` EAPI extern int foo; EAPI void foobar(void); EAPI void bar() { foo = 5; foobar(); } ``` dll2: ``` EAPI int foo = 0; EAPI void foobar() { printf ("foo %d\n", foo); } ``` This will work on mingw but will not work for MSVC. And that's why EAPI is the only solution that worked for MSVC. Co-authored-by: João Paulo Taylor Ienczak Zanette <jpaulotiz@gmail.com> Co-authored-by: Ricardo Campos <ricardo.campos@expertise.dev> Co-authored-by: Lucas Cavalcante de Sousa <lucks.sousa@gmail.com> Reviewers: jptiz, lucas, woohyun, vtorri, raster Reviewed By: jptiz, lucas, vtorri Subscribers: ProhtMeyhet, cedric, #reviewers, #committers Tags: #efl Differential Revision: https://phab.enlightenment.org/D12188
2020-11-25 04:35:48 -08:00
EINA_API Eina_Future_Desc eina_future_cb_log_from_desc(const Eina_Future_Cb_Log_Desc desc) EINA_WARN_UNUSED_RESULT;
/**
* Creates a future chain.
*
* This behaves exactly like eina_future_then_from_desc(), but makes it easier to create future chains.
*
* If during the future chain creation an error happens this function will return @c NULL,
* and the Eina_Future_Desc::cb from the @p descs array will be called reporting an error (`EINVAL` or `ENOMEM`),
* so the user has a chance to free Eina_Future_Desc::data.
*
* Just like eina_future_then_from_desc(), the value returned by a Eina_Future_Desc::cb callback will
* be delivered to the next Eina_Future_Desc::cb in the chain.
*
* In case a future in the chain is canceled, the whole chain will be canceled immediately
* and the error `ECANCELED` will be reported.
*
* Here's an example:
*
* @code
* // callbacks code....
*
* Eina_Future* chain(void)
* {
* Eina_Future *f = get_file_size_async("/MyFile.txt");
* return eina_future_chain(f, eina_future_cb_easy(_future_cb1, _future_err, NULL, EINA_VALUE_TYPE_INT, NULL),
* eina_future_cb_easy(_future_cb2, _future_err, NULL, EINA_VALUE_TYPE_STRING, NULL),
* eina_future_cb_easy(_future_cb3, _future_err, NULL, EINA_VALUE_TYPE_DOUBLE, NULL),
* {.cb = _future_cb4, .data = NULL });
* }
* @endcode
*
* @param[in,out] prev The previous future
* @param[in] descs An array of descriptions.
The last element of the array must have the Eina_Future_Desc::cb set to @c NULL
* @return A future or @c NULL on error.
* @note If an error happens the whole future chain will CANCELED and
* desc.cb will be called in order to free desc.data.
* @see eina_future_new()
* @see eina_future_then()
* @see Eina_Future_Desc
* @see eina_future_chain()
* @see eina_future_cb_ignore_error()
* @see eina_future_cb_console_from_desc()
* @see eina_future_cb_log_from_desc()
* @see eina_future_cb_easy_from_desc()
* @see eina_future_cb_easy()
* @see eina_future_then_from_desc()
* @see eina_future_then_easy()
* @see eina_future_cb_convert_to()
*/
eina: Rename EAPI macro to EINA_API in Eina library Summary: Patch from a series of patches to rename EAPI symbols to specific library DSOs. EAPI was designed to be able to pass `__attribute__ ((visibility ("default")))` for symbols with GCC, which would mean that even if -fvisibility=hidden was used when compiling the library, the needed symbols would get exported. MSVC __almost__ works like GCC (or mingw) in which you can declare everything as export and it will just work (slower, but it will work). But there's a caveat: global variables will not work the same way for MSVC, but works for mingw and GCC. For global variables (as opposed to functions), MSVC requires correct DSO visibility for MSVC: instead of declaring a symbol as export for everything, you need to declare it as import when importing from another DSO and export when defining it locally. With current EAPI definitions, we get the following example working in mingw and MSVC (observe it doesn't define any global variables as exported symbols). Example 1: dll1: ``` EAPI void foo(void); EAPI void bar() { foo(); } ``` dll2: ``` EAPI void foo() { printf ("foo\n"); } ``` This works fine with API defined as __declspec(dllexport) in both cases and for gcc defining as `__atttribute__((visibility("default")))` However, the following: Example 2: dll1: ``` EAPI extern int foo; EAPI void foobar(void); EAPI void bar() { foo = 5; foobar(); } ``` dll2: ``` EAPI int foo = 0; EAPI void foobar() { printf ("foo %d\n", foo); } ``` This will work on mingw but will not work for MSVC. And that's why EAPI is the only solution that worked for MSVC. Co-authored-by: João Paulo Taylor Ienczak Zanette <jpaulotiz@gmail.com> Co-authored-by: Ricardo Campos <ricardo.campos@expertise.dev> Co-authored-by: Lucas Cavalcante de Sousa <lucks.sousa@gmail.com> Reviewers: jptiz, lucas, woohyun, vtorri, raster Reviewed By: jptiz, lucas, vtorri Subscribers: ProhtMeyhet, cedric, #reviewers, #committers Tags: #efl Differential Revision: https://phab.enlightenment.org/D12188
2020-11-25 04:35:48 -08:00
EINA_API Eina_Future *eina_future_chain_array(Eina_Future *prev, const Eina_Future_Desc descs[]) EINA_ARG_NONNULL(1, 2);
/**
* Wrapper around eina_future_chain_array() and eina_future_cb_easy_from_desc()
*
* This functions makes it easier to use eina_future_chain_array() with eina_future_cb_easy_from_desc(),
* check the macro eina_future_chain_easy() for a syntax sugar.
*
*
* @param[in,out] prev The previous future
* @param[in] descs An array of descriptions. The last element of the array must have the Eina_Future_Desc::cb set to @c NULL
* @return A future or @c NULL on error.
* @note If an error happens the whole future chain will CANCELED and
* desc.cb will be called in order to free desc.data.
* @see eina_future_chain_easy()
* @see eina_future_chain()
* @see eina_future_cb_easy()
* @see eina_future_chain_array()
* @see eina_future_cb_easy_from_desc()
*/
eina: Rename EAPI macro to EINA_API in Eina library Summary: Patch from a series of patches to rename EAPI symbols to specific library DSOs. EAPI was designed to be able to pass `__attribute__ ((visibility ("default")))` for symbols with GCC, which would mean that even if -fvisibility=hidden was used when compiling the library, the needed symbols would get exported. MSVC __almost__ works like GCC (or mingw) in which you can declare everything as export and it will just work (slower, but it will work). But there's a caveat: global variables will not work the same way for MSVC, but works for mingw and GCC. For global variables (as opposed to functions), MSVC requires correct DSO visibility for MSVC: instead of declaring a symbol as export for everything, you need to declare it as import when importing from another DSO and export when defining it locally. With current EAPI definitions, we get the following example working in mingw and MSVC (observe it doesn't define any global variables as exported symbols). Example 1: dll1: ``` EAPI void foo(void); EAPI void bar() { foo(); } ``` dll2: ``` EAPI void foo() { printf ("foo\n"); } ``` This works fine with API defined as __declspec(dllexport) in both cases and for gcc defining as `__atttribute__((visibility("default")))` However, the following: Example 2: dll1: ``` EAPI extern int foo; EAPI void foobar(void); EAPI void bar() { foo = 5; foobar(); } ``` dll2: ``` EAPI int foo = 0; EAPI void foobar() { printf ("foo %d\n", foo); } ``` This will work on mingw but will not work for MSVC. And that's why EAPI is the only solution that worked for MSVC. Co-authored-by: João Paulo Taylor Ienczak Zanette <jpaulotiz@gmail.com> Co-authored-by: Ricardo Campos <ricardo.campos@expertise.dev> Co-authored-by: Lucas Cavalcante de Sousa <lucks.sousa@gmail.com> Reviewers: jptiz, lucas, woohyun, vtorri, raster Reviewed By: jptiz, lucas, vtorri Subscribers: ProhtMeyhet, cedric, #reviewers, #committers Tags: #efl Differential Revision: https://phab.enlightenment.org/D12188
2020-11-25 04:35:48 -08:00
EINA_API Eina_Future *eina_future_chain_easy_array(Eina_Future *prev, const Eina_Future_Cb_Easy_Desc descs[]) EINA_ARG_NONNULL(1, 2);
/**
* Creates an #Eina_Future_Desc that will print the previous future's resolved value.
*
* This function can be used to quickly inspect the value that an Eina_Future_Desc::cb
* is returning. The returned value will be passed to the next future in the chain without
* modifications.
*
* There's also a helper macro called eina_future_cb_console() which makes this
* function easier to use.
*
* Example:
*
* @code
*
* eina_future_chain(a_future, {.cb = cb1, .data = NULL},
* //Inspect the cb1 value and pass to cb2 without modifications
* eina_future_cb_console("cb1 value:"),
* {.cb = cb2, .data = NULL},
* //Inspect the cb2 value
* eina_future_cb_console("cb2 value:", " cb2 value suffix\n"))
* @endcode
*
* @param[in] desc Description object with contextual information.
* @return An #Eina_Future_Desc
*
* The description object, @p desc, can (optionally) include a prefix, suffix,
* filename and function name. If these are @c NULL, empty strings ("") are used
* instead. If @p desc->suffix is provided, the '\\n' should be provided to ensure
* the printed string ends with a line feed.
*
* @see eina_future_then()
* @see Eina_Future_Desc
* @see eina_future_chain()
* @see eina_future_cb_easy_from_desc()
* @see eina_future_cb_easy()
* @see eina_future_then_from_desc()
* @see eina_future_then_easy()
* @see eina_future_cb_console()
* @see eina_future_cb_ignore_error()
* @see eina_future_cb_log_from_desc()
*/
eina: Rename EAPI macro to EINA_API in Eina library Summary: Patch from a series of patches to rename EAPI symbols to specific library DSOs. EAPI was designed to be able to pass `__attribute__ ((visibility ("default")))` for symbols with GCC, which would mean that even if -fvisibility=hidden was used when compiling the library, the needed symbols would get exported. MSVC __almost__ works like GCC (or mingw) in which you can declare everything as export and it will just work (slower, but it will work). But there's a caveat: global variables will not work the same way for MSVC, but works for mingw and GCC. For global variables (as opposed to functions), MSVC requires correct DSO visibility for MSVC: instead of declaring a symbol as export for everything, you need to declare it as import when importing from another DSO and export when defining it locally. With current EAPI definitions, we get the following example working in mingw and MSVC (observe it doesn't define any global variables as exported symbols). Example 1: dll1: ``` EAPI void foo(void); EAPI void bar() { foo(); } ``` dll2: ``` EAPI void foo() { printf ("foo\n"); } ``` This works fine with API defined as __declspec(dllexport) in both cases and for gcc defining as `__atttribute__((visibility("default")))` However, the following: Example 2: dll1: ``` EAPI extern int foo; EAPI void foobar(void); EAPI void bar() { foo = 5; foobar(); } ``` dll2: ``` EAPI int foo = 0; EAPI void foobar() { printf ("foo %d\n", foo); } ``` This will work on mingw but will not work for MSVC. And that's why EAPI is the only solution that worked for MSVC. Co-authored-by: João Paulo Taylor Ienczak Zanette <jpaulotiz@gmail.com> Co-authored-by: Ricardo Campos <ricardo.campos@expertise.dev> Co-authored-by: Lucas Cavalcante de Sousa <lucks.sousa@gmail.com> Reviewers: jptiz, lucas, woohyun, vtorri, raster Reviewed By: jptiz, lucas, vtorri Subscribers: ProhtMeyhet, cedric, #reviewers, #committers Tags: #efl Differential Revision: https://phab.enlightenment.org/D12188
2020-11-25 04:35:48 -08:00
EINA_API Eina_Future_Desc eina_future_cb_console_from_desc(const Eina_Future_Cb_Console_Desc desc) EINA_WARN_UNUSED_RESULT;
/**
* Returns an #Eina_Future_Desc that ignores an error.
*
* This function may be used if one wants to ignore an error. If the
* specified error happens an #EINA_VALUE_EMPTY will be delivered to the
* next future in the chain.
*
* @param[in] err The error to be ignored.
* @return A future descriptor to be used with eina_future_then() or eina_future_chain()
*/
eina: Rename EAPI macro to EINA_API in Eina library Summary: Patch from a series of patches to rename EAPI symbols to specific library DSOs. EAPI was designed to be able to pass `__attribute__ ((visibility ("default")))` for symbols with GCC, which would mean that even if -fvisibility=hidden was used when compiling the library, the needed symbols would get exported. MSVC __almost__ works like GCC (or mingw) in which you can declare everything as export and it will just work (slower, but it will work). But there's a caveat: global variables will not work the same way for MSVC, but works for mingw and GCC. For global variables (as opposed to functions), MSVC requires correct DSO visibility for MSVC: instead of declaring a symbol as export for everything, you need to declare it as import when importing from another DSO and export when defining it locally. With current EAPI definitions, we get the following example working in mingw and MSVC (observe it doesn't define any global variables as exported symbols). Example 1: dll1: ``` EAPI void foo(void); EAPI void bar() { foo(); } ``` dll2: ``` EAPI void foo() { printf ("foo\n"); } ``` This works fine with API defined as __declspec(dllexport) in both cases and for gcc defining as `__atttribute__((visibility("default")))` However, the following: Example 2: dll1: ``` EAPI extern int foo; EAPI void foobar(void); EAPI void bar() { foo = 5; foobar(); } ``` dll2: ``` EAPI int foo = 0; EAPI void foobar() { printf ("foo %d\n", foo); } ``` This will work on mingw but will not work for MSVC. And that's why EAPI is the only solution that worked for MSVC. Co-authored-by: João Paulo Taylor Ienczak Zanette <jpaulotiz@gmail.com> Co-authored-by: Ricardo Campos <ricardo.campos@expertise.dev> Co-authored-by: Lucas Cavalcante de Sousa <lucks.sousa@gmail.com> Reviewers: jptiz, lucas, woohyun, vtorri, raster Reviewed By: jptiz, lucas, vtorri Subscribers: ProhtMeyhet, cedric, #reviewers, #committers Tags: #efl Differential Revision: https://phab.enlightenment.org/D12188
2020-11-25 04:35:48 -08:00
EINA_API Eina_Future_Desc eina_future_cb_ignore_error(Eina_Error err);
/**
* Creates an #Eina_Future_Desc which will convert the received eina value to a given type.
*
* @param[in] type The #Eina_Value_Type to convert to.
* @return An #Eina_Future_Desc
* @see eina_future_then()
* @see Eina_Future_Desc
* @see eina_future_chain()
* @see eina_future_cb_easy_from_desc()
* @see eina_future_cb_easy()
* @see eina_future_then_from_desc()
* @see eina_future_then_easy()
*/
eina: Rename EAPI macro to EINA_API in Eina library Summary: Patch from a series of patches to rename EAPI symbols to specific library DSOs. EAPI was designed to be able to pass `__attribute__ ((visibility ("default")))` for symbols with GCC, which would mean that even if -fvisibility=hidden was used when compiling the library, the needed symbols would get exported. MSVC __almost__ works like GCC (or mingw) in which you can declare everything as export and it will just work (slower, but it will work). But there's a caveat: global variables will not work the same way for MSVC, but works for mingw and GCC. For global variables (as opposed to functions), MSVC requires correct DSO visibility for MSVC: instead of declaring a symbol as export for everything, you need to declare it as import when importing from another DSO and export when defining it locally. With current EAPI definitions, we get the following example working in mingw and MSVC (observe it doesn't define any global variables as exported symbols). Example 1: dll1: ``` EAPI void foo(void); EAPI void bar() { foo(); } ``` dll2: ``` EAPI void foo() { printf ("foo\n"); } ``` This works fine with API defined as __declspec(dllexport) in both cases and for gcc defining as `__atttribute__((visibility("default")))` However, the following: Example 2: dll1: ``` EAPI extern int foo; EAPI void foobar(void); EAPI void bar() { foo = 5; foobar(); } ``` dll2: ``` EAPI int foo = 0; EAPI void foobar() { printf ("foo %d\n", foo); } ``` This will work on mingw but will not work for MSVC. And that's why EAPI is the only solution that worked for MSVC. Co-authored-by: João Paulo Taylor Ienczak Zanette <jpaulotiz@gmail.com> Co-authored-by: Ricardo Campos <ricardo.campos@expertise.dev> Co-authored-by: Lucas Cavalcante de Sousa <lucks.sousa@gmail.com> Reviewers: jptiz, lucas, woohyun, vtorri, raster Reviewed By: jptiz, lucas, vtorri Subscribers: ProhtMeyhet, cedric, #reviewers, #committers Tags: #efl Differential Revision: https://phab.enlightenment.org/D12188
2020-11-25 04:35:48 -08:00
EINA_API Eina_Future_Desc eina_future_cb_convert_to(const Eina_Value_Type *type);
/**
* Creates an #Eina_Future_Desc based on an #Eina_Future_Cb_Easy_Desc.
*
* This function aims to be used in conjunction with eina_future_chain(),
* eina_future_then_from_desc() and friends and its main objective is to simplify
* error handling and #Eina_Value type checks.
* It uses three callbacks to inform the user about the future's
* result and life cycle. They are:
*
* @li Eina_Future_Cb_Easy_Desc::success: This callback is called when
* the future execution was successful, that is, no errors occurred and
* the result type matches Eina_Future_Cb_Easy_Desc::success_type. In case
* Eina_Future_Cb_Easy_Desc::success_type is @c NULL, this function will
* only be called if the future did not report an error. The value returned
* by this function will be propagated to the next future in the chain (if any).
*
* @li Eina_Future_Cb_Easy_Desc::error: This callback is called when
* the future result is an error or Eina_Future_Cb_Easy_Desc::success_type
* does not match the future result type. The value returned
* by this function will be propagated to the next future in the chain (if any).
*
* @li Eina_Future_Cb_Easy_Desc::free: Called after the future was freed and any resources
* allocated must be freed at this point. This callback is always called.
*
* Check the example below for a sample usage:
*
* @code
* static Eina_Value
* _file_size_ok(void *data, Eina_Value v)
* {
* Ctx *ctx = data;
* // Since an Eina_Future_Cb_Easy_Desc::success_type was provided, there's no need to check the value type
* int s;
* eina_value_get(&v, &s);
* printf("File size is %d bytes\n", s);
* ctx->file_size = s;
* return v;
* }
*
* static Eina_Value
* _file_size_err(void *data, Eina_Error err)
* {
* fprintf(stderr, "Could not read the file size. Reason: %s\n", eina_error_msg_get(err));
* // Stop propagating the error.
* return EINA_VALUE_EMPTY;
* }
*
* static void
* _future_freed(void *data, const Eina_Future dead)
* {
* Ctx *ctx = data;
* printf("Future %p deleted\n", dead);
* ctx->file_size_handler_cb(ctx->file_size);
* free(ctx);
* }
*
* void do_work(File_Size_Handler_Cb cb)
* {
* Ctx *ctx = malloc(sizeof(Ctx));
* ctx->f = get_file_size("/tmp/todo.txt");
* ctx->file_size = -1;
* ctx->file_size_handler_cb = cb;
* eina_future_then_easy(f, _file_size_ok, _file_size_err, _future_freed, EINA_VALUE_TYPE_INT, ctx);
* }
* @endcode
*
* @param[in] desc The easy callback's description.
* @return An #Eina_Future_Desc
*
* @see eina_future_chain()
* @see eina_future_then()
* @see eina_future_cb_easy()
*/
eina: Rename EAPI macro to EINA_API in Eina library Summary: Patch from a series of patches to rename EAPI symbols to specific library DSOs. EAPI was designed to be able to pass `__attribute__ ((visibility ("default")))` for symbols with GCC, which would mean that even if -fvisibility=hidden was used when compiling the library, the needed symbols would get exported. MSVC __almost__ works like GCC (or mingw) in which you can declare everything as export and it will just work (slower, but it will work). But there's a caveat: global variables will not work the same way for MSVC, but works for mingw and GCC. For global variables (as opposed to functions), MSVC requires correct DSO visibility for MSVC: instead of declaring a symbol as export for everything, you need to declare it as import when importing from another DSO and export when defining it locally. With current EAPI definitions, we get the following example working in mingw and MSVC (observe it doesn't define any global variables as exported symbols). Example 1: dll1: ``` EAPI void foo(void); EAPI void bar() { foo(); } ``` dll2: ``` EAPI void foo() { printf ("foo\n"); } ``` This works fine with API defined as __declspec(dllexport) in both cases and for gcc defining as `__atttribute__((visibility("default")))` However, the following: Example 2: dll1: ``` EAPI extern int foo; EAPI void foobar(void); EAPI void bar() { foo = 5; foobar(); } ``` dll2: ``` EAPI int foo = 0; EAPI void foobar() { printf ("foo %d\n", foo); } ``` This will work on mingw but will not work for MSVC. And that's why EAPI is the only solution that worked for MSVC. Co-authored-by: João Paulo Taylor Ienczak Zanette <jpaulotiz@gmail.com> Co-authored-by: Ricardo Campos <ricardo.campos@expertise.dev> Co-authored-by: Lucas Cavalcante de Sousa <lucks.sousa@gmail.com> Reviewers: jptiz, lucas, woohyun, vtorri, raster Reviewed By: jptiz, lucas, vtorri Subscribers: ProhtMeyhet, cedric, #reviewers, #committers Tags: #efl Differential Revision: https://phab.enlightenment.org/D12188
2020-11-25 04:35:48 -08:00
EINA_API Eina_Future_Desc eina_future_cb_easy_from_desc(const Eina_Future_Cb_Easy_Desc desc) EINA_WARN_UNUSED_RESULT;
/**
* Creates an all promise.
*
* Creates a promise that is resolved once all the futures
* from the @p array are resolved.
* The promise is resolved with an #Eina_Value type array which
* contains #EINA_VALUE_TYPE_VALUE elements. The result array is
* ordered according to the @p array argument. Example:
*
* @code
* static const char *
* _get_operation_name_by_index(int idx)
* {
* switch (idx)
* {
* case 0: return "Get file data";
* case 1: return "Get file size";
* default: return "sum";
* }
* }
*
* static Eina_Value
* _all_cb(const void *data EINA_UNUSED, const Eina_Value array, const Eina_Future *dead EINA_UNUSED)
* {
* Eina_Error err;
* unsigned int i, len;
*
* if (array.type == EINA_VALUE_TYPE_ERROR)
* {
* eina_value_get(&array, &err);
* fprintf(stderr, "Could not complete all operations. Reason: %s\n", eina_error_msg_get(err));
* return array;
* }
* len = eina_value_array_count(&array);
* for (i = 0; i < len; i++)
* {
* Eina_Value v;
* eina_value_array_get(&array, i, &v);
* if (v.type == EINA_VALUE_TYPE_ERROR)
* {
* eina_value_get(&v, &err);
* fprintf(stderr, "Could not complete operation '%s'. Reason: %s\n", _get_operation_name_by_index(i), eina_error_msg_get(err));
* continue;
* }
* if (!i)
* {
* const char *msg;
* if (v.type != EINA_VALUE_TYPE_STRING)
* {
* fprintf(stderr, "Operation %s expects '%s' - received '%s'\n", _get_operation_name_by_index(i), EINA_VALUE_TYPE_STRING->name, v.type->name);
* continue;
* }
* eina_value_get(&v, &msg);
* printf("File content:%s\n", msg);
* }
* else if (i == 1)
* {
* int i;
* if (v.type != EINA_VALUE_TYPE_INT)
* {
* fprintf(stderr, "Operation %s expects '%s' - received '%s'\n", _get_operation_name_by_index(i), EINA_VALUE_TYPE_INT->name, v.type->name);
* continue;
* }
* eina_value_get(&v, &i);
* printf("File size: %d\n", i);
* }
* else
* {
* double p;
* if (v.type != EINA_VALUE_TYPE_DOUBLE)
* {
* fprintf(stderr, "Operation %s expects '%s' - received '%s'\n", _get_operation_name_by_index(i), EINA_VALUE_TYPE_DOUBLE->name, v.type->name);
* continue;
* }
* eina_value_get(&v, &p);
* printf("50 places of PI: %f\n", p);
* }
* }
* return array;
* }
*
* void func(void)
* {
* Eina_Future *f1, *f2, *f3, f_all;
*
* f1 = read_file("/tmp/todo.txt");
* f2 = get_file_size("/tmp/file.txt");
* // calculates 50 places of PI
* f3 = calc_pi(50);
* f_all = eina_future_all(f1, f2, f3);
* eina_future_then(f_all, _all_cb, NULL);
* }
* @endcode
*
* @param[in,out] array An array of futures, terminated with #EINA_FUTURE_SENTINEL.
* @return A promise or @c NULL on error.
* @note On error all the futures will be CANCELED.
* @see eina_future_all_array()
*/
eina: Rename EAPI macro to EINA_API in Eina library Summary: Patch from a series of patches to rename EAPI symbols to specific library DSOs. EAPI was designed to be able to pass `__attribute__ ((visibility ("default")))` for symbols with GCC, which would mean that even if -fvisibility=hidden was used when compiling the library, the needed symbols would get exported. MSVC __almost__ works like GCC (or mingw) in which you can declare everything as export and it will just work (slower, but it will work). But there's a caveat: global variables will not work the same way for MSVC, but works for mingw and GCC. For global variables (as opposed to functions), MSVC requires correct DSO visibility for MSVC: instead of declaring a symbol as export for everything, you need to declare it as import when importing from another DSO and export when defining it locally. With current EAPI definitions, we get the following example working in mingw and MSVC (observe it doesn't define any global variables as exported symbols). Example 1: dll1: ``` EAPI void foo(void); EAPI void bar() { foo(); } ``` dll2: ``` EAPI void foo() { printf ("foo\n"); } ``` This works fine with API defined as __declspec(dllexport) in both cases and for gcc defining as `__atttribute__((visibility("default")))` However, the following: Example 2: dll1: ``` EAPI extern int foo; EAPI void foobar(void); EAPI void bar() { foo = 5; foobar(); } ``` dll2: ``` EAPI int foo = 0; EAPI void foobar() { printf ("foo %d\n", foo); } ``` This will work on mingw but will not work for MSVC. And that's why EAPI is the only solution that worked for MSVC. Co-authored-by: João Paulo Taylor Ienczak Zanette <jpaulotiz@gmail.com> Co-authored-by: Ricardo Campos <ricardo.campos@expertise.dev> Co-authored-by: Lucas Cavalcante de Sousa <lucks.sousa@gmail.com> Reviewers: jptiz, lucas, woohyun, vtorri, raster Reviewed By: jptiz, lucas, vtorri Subscribers: ProhtMeyhet, cedric, #reviewers, #committers Tags: #efl Differential Revision: https://phab.enlightenment.org/D12188
2020-11-25 04:35:48 -08:00
EINA_API Eina_Promise *eina_promise_all_array(Eina_Future *array[]) EINA_ARG_NONNULL(1) EINA_WARN_UNUSED_RESULT;
/**
* Creates an all promise from an iterator.
*
* Creates a promise that is resolved once all the futures
* from the @p iterator are resolved.
* The promise is resolved with an #Eina_Value type array which
* contains #EINA_VALUE_TYPE_VALUE elements. The result array is
* ordered according to the @p iterator order.
*
* @param[in] iterator An iterator of futures. Will be destroyed after the call.
* @return A promise or @c NULL on error.
* @note On error all the futures will be CANCELED.
* @see eina_future_all_iterator()
*/
eina: Rename EAPI macro to EINA_API in Eina library Summary: Patch from a series of patches to rename EAPI symbols to specific library DSOs. EAPI was designed to be able to pass `__attribute__ ((visibility ("default")))` for symbols with GCC, which would mean that even if -fvisibility=hidden was used when compiling the library, the needed symbols would get exported. MSVC __almost__ works like GCC (or mingw) in which you can declare everything as export and it will just work (slower, but it will work). But there's a caveat: global variables will not work the same way for MSVC, but works for mingw and GCC. For global variables (as opposed to functions), MSVC requires correct DSO visibility for MSVC: instead of declaring a symbol as export for everything, you need to declare it as import when importing from another DSO and export when defining it locally. With current EAPI definitions, we get the following example working in mingw and MSVC (observe it doesn't define any global variables as exported symbols). Example 1: dll1: ``` EAPI void foo(void); EAPI void bar() { foo(); } ``` dll2: ``` EAPI void foo() { printf ("foo\n"); } ``` This works fine with API defined as __declspec(dllexport) in both cases and for gcc defining as `__atttribute__((visibility("default")))` However, the following: Example 2: dll1: ``` EAPI extern int foo; EAPI void foobar(void); EAPI void bar() { foo = 5; foobar(); } ``` dll2: ``` EAPI int foo = 0; EAPI void foobar() { printf ("foo %d\n", foo); } ``` This will work on mingw but will not work for MSVC. And that's why EAPI is the only solution that worked for MSVC. Co-authored-by: João Paulo Taylor Ienczak Zanette <jpaulotiz@gmail.com> Co-authored-by: Ricardo Campos <ricardo.campos@expertise.dev> Co-authored-by: Lucas Cavalcante de Sousa <lucks.sousa@gmail.com> Reviewers: jptiz, lucas, woohyun, vtorri, raster Reviewed By: jptiz, lucas, vtorri Subscribers: ProhtMeyhet, cedric, #reviewers, #committers Tags: #efl Differential Revision: https://phab.enlightenment.org/D12188
2020-11-25 04:35:48 -08:00
EINA_API Eina_Promise *eina_promise_all_iterator(Eina_Iterator *iterator) EINA_ARG_NONNULL(1) EINA_WARN_UNUSED_RESULT;
/**
* Creates a race promise.
*
* Creates a promise that resolves when a future from the @p array
* is completed. The remaining futures will be canceled with the
* error code `ECANCELED`
*
* The resulting value is an #EINA_VALUE_TYPE_STRUCT with two fields:
*
* @li A field named @c value containing an #Eina_Value with the result itself.
* @li A field named @c index containing the index of the completed
* function relative to the @p array.
*
* Example:
*
* @code
* static const char *
* _get_operation_name_by_index(int idx)
* {
* switch (idx)
* {
* case 0: return "Get file data";
* case 1: return "Get file size";
* default: return "sum";
* }
* }
*
* static Eina_Value
* _race_cb(const void *data EINA_UNUSED, const Eina_Value v)
* {
* unsigned int i;
* Eina_Value result;
* Eina_Error err;
* Eina_Value_Struct *st;
*
* // No need to check for the 'v' type. eina_future_cb_easy() did that for us.
* // However we should check if the struct has the correct description
* st = eina_value_memory_get(&v);
* if (st->desc != EINA_PROMISE_RACE_STRUCT_DESC)
* {
* fprintf(stderr, "Eina_Value is not a race struct\n");
* return v;
* }
* eina_value_struct_get(&v, "index", &i);
* // Get the operation result
* eina_value_struct_get(&v, "value", &result);
* if (!i)
* {
* const char *msg;
* if (result.type != EINA_VALUE_TYPE_STRING)
* fprintf(stderr, "Operation %s expects '%s' - received '%s'\n", _get_operation_name_by_index(i), EINA_VALUE_TYPE_STRING->name, result.type->name);
* else
* {
* eina_value_get(&result, &msg);
* printf("File content:%s\n", msg);
* }
* }
* else if (i == 1)
* {
* int i;
* if (result.type != EINA_VALUE_TYPE_INT)
* fprintf(stderr, "Operation %s expects '%s' - received '%s'\n", _get_operation_name_by_index(i), EINA_VALUE_TYPE_INT->name, v.type->name);
* else
* {
* eina_value_get(&result, &i);
* printf("File size: %d\n", i);
* }
* }
* else
* {
* double p;
* if (result.type != EINA_VALUE_TYPE_DOUBLE)
* fprintf(stderr, "Operation %s expects '%s' - received '%s'\n", _get_operation_name_by_index(i), EINA_VALUE_TYPE_DOUBLE->name, result.type->name);
* else
* {
* eina_value_get(&result, &p);
* printf("50 places of PI: %f\n", p);
* }
* }
* eina_value_flush(&result);
* return v;
* }
*
* static Eina_Value
* _race_err(void *data, Eina_Error err)
* {
* fprintf(stderr, "Could not complete the race future. Reason: %s\n", eina_error_msg_get(err));
* return EINA_VALUE_EMPTY;
* }
*
* void func(void)
* {
* static const *Eina_Future[] = {NULL, NULL, NULL, NULL};
* Eina_List *l = NULL;
*
* futures[0] = read_file("/tmp/todo.txt");
* futures[1] = get_file_size("/tmp/file.txt");
* // calculates 50 places of PI
* futures[2] = calc_pi(50);
* f_race = eina_future_race_array(futures);
* eina_future_then_easy(f_race, _race_cb, _race_err, NULL, EINA_VALUE_TYPE_STRUCT, NULL);
* }
* @endcode
*
* @param[in,out] array An array of futures, terminated by #EINA_FUTURE_SENTINEL.
* @return A promise or @c NULL on error.
* @note On error all the futures will be CANCELED.
* @see eina_future_race_array()
* @see _Eina_Future_Race_Result
*/
eina: Rename EAPI macro to EINA_API in Eina library Summary: Patch from a series of patches to rename EAPI symbols to specific library DSOs. EAPI was designed to be able to pass `__attribute__ ((visibility ("default")))` for symbols with GCC, which would mean that even if -fvisibility=hidden was used when compiling the library, the needed symbols would get exported. MSVC __almost__ works like GCC (or mingw) in which you can declare everything as export and it will just work (slower, but it will work). But there's a caveat: global variables will not work the same way for MSVC, but works for mingw and GCC. For global variables (as opposed to functions), MSVC requires correct DSO visibility for MSVC: instead of declaring a symbol as export for everything, you need to declare it as import when importing from another DSO and export when defining it locally. With current EAPI definitions, we get the following example working in mingw and MSVC (observe it doesn't define any global variables as exported symbols). Example 1: dll1: ``` EAPI void foo(void); EAPI void bar() { foo(); } ``` dll2: ``` EAPI void foo() { printf ("foo\n"); } ``` This works fine with API defined as __declspec(dllexport) in both cases and for gcc defining as `__atttribute__((visibility("default")))` However, the following: Example 2: dll1: ``` EAPI extern int foo; EAPI void foobar(void); EAPI void bar() { foo = 5; foobar(); } ``` dll2: ``` EAPI int foo = 0; EAPI void foobar() { printf ("foo %d\n", foo); } ``` This will work on mingw but will not work for MSVC. And that's why EAPI is the only solution that worked for MSVC. Co-authored-by: João Paulo Taylor Ienczak Zanette <jpaulotiz@gmail.com> Co-authored-by: Ricardo Campos <ricardo.campos@expertise.dev> Co-authored-by: Lucas Cavalcante de Sousa <lucks.sousa@gmail.com> Reviewers: jptiz, lucas, woohyun, vtorri, raster Reviewed By: jptiz, lucas, vtorri Subscribers: ProhtMeyhet, cedric, #reviewers, #committers Tags: #efl Differential Revision: https://phab.enlightenment.org/D12188
2020-11-25 04:35:48 -08:00
EINA_API Eina_Promise *eina_promise_race_array(Eina_Future *array[]) EINA_ARG_NONNULL(1) EINA_WARN_UNUSED_RESULT;
/**
* @struct _Eina_Future_Race_Result
* The struct that is used to store the race result.
*
* When using eina_promise_race_array() and friends, the future result
* will be reported as a struct. The values can be obtained using
* eina_value_struct_get() or one could access the struct directly
* such as this example:
*
* @code
* static Eina_Value
* _race_cb(const void *data EINA_UNUSED, const Eina_Value v)
* {
* //code...
* Eina_Value_Struct st;
* Eina_Future_Race_Result *rr;
* eina_value_get(v, &st);
* rr = st.memory;
* printf("Winning future index: %u\n", rr->index);
* //more code..
* return v;
* }
* @endcode
*
* @see eina_promise_race_array()
* @see eina_future_race_array()
* @see eina_promise_race()
* @see eina_future_race()
* @see EINA_PROMISE_RACE_STRUCT_DESC
*/
struct _Eina_Future_Race_Result {
/**
* The race result.
*/
Eina_Value value;
/**
* The future index that won the race.
*/
unsigned int index;
};
/**
* @var const Eina_Value_Struct_Desc *EINA_PROMISE_RACE_STRUCT_DESC
*
* A pointer to the race struct description, which
* is used by eina_promise_race_array().
*
* This struct contains two members:
* @li @c value: An #EINA_VALUE_TYPE_VALUE that contains the future result that won the race.
* @li @c index: An #EINA_VALUE_TYPE_UINT that contains the future index that won the race.
*
* @see eina_promise_race_array()
* @see _Eina_Future_Race_Result
*/
eina: Rename EAPI macro to EINA_API in Eina library Summary: Patch from a series of patches to rename EAPI symbols to specific library DSOs. EAPI was designed to be able to pass `__attribute__ ((visibility ("default")))` for symbols with GCC, which would mean that even if -fvisibility=hidden was used when compiling the library, the needed symbols would get exported. MSVC __almost__ works like GCC (or mingw) in which you can declare everything as export and it will just work (slower, but it will work). But there's a caveat: global variables will not work the same way for MSVC, but works for mingw and GCC. For global variables (as opposed to functions), MSVC requires correct DSO visibility for MSVC: instead of declaring a symbol as export for everything, you need to declare it as import when importing from another DSO and export when defining it locally. With current EAPI definitions, we get the following example working in mingw and MSVC (observe it doesn't define any global variables as exported symbols). Example 1: dll1: ``` EAPI void foo(void); EAPI void bar() { foo(); } ``` dll2: ``` EAPI void foo() { printf ("foo\n"); } ``` This works fine with API defined as __declspec(dllexport) in both cases and for gcc defining as `__atttribute__((visibility("default")))` However, the following: Example 2: dll1: ``` EAPI extern int foo; EAPI void foobar(void); EAPI void bar() { foo = 5; foobar(); } ``` dll2: ``` EAPI int foo = 0; EAPI void foobar() { printf ("foo %d\n", foo); } ``` This will work on mingw but will not work for MSVC. And that's why EAPI is the only solution that worked for MSVC. Co-authored-by: João Paulo Taylor Ienczak Zanette <jpaulotiz@gmail.com> Co-authored-by: Ricardo Campos <ricardo.campos@expertise.dev> Co-authored-by: Lucas Cavalcante de Sousa <lucks.sousa@gmail.com> Reviewers: jptiz, lucas, woohyun, vtorri, raster Reviewed By: jptiz, lucas, vtorri Subscribers: ProhtMeyhet, cedric, #reviewers, #committers Tags: #efl Differential Revision: https://phab.enlightenment.org/D12188
2020-11-25 04:35:48 -08:00
EINA_API extern const Eina_Value_Struct_Desc *EINA_PROMISE_RACE_STRUCT_DESC;
/**
* Creates a future that will be resolved once all futures from @p array is resolved.
* This is a helper over eina_promise_all_array()
*
* @param[in,out] array A future array, must be terminated with #EINA_FUTURE_SENTINEL.
* @return A future.
* @see eina_promise_all_array()
* @see EINA_FUTURE_SENTINEL
*/
static inline Eina_Future *
eina_future_all_array(Eina_Future *array[])
{
Eina_Promise *p = eina_promise_all_array(array);
if (!p) return NULL;
return eina_future_new(p);
}
/**
* Creates a future that will be resolved once all futures from @p iterator are resolved.
* This is a helper over eina_promise_all_iterator()
*
* @param[in] iterator An iterator containing futures. Will be destroyed on exit of the function.
* @return A future.
* @see eina_promise_all_iterator()
*/
static inline Eina_Future *
eina_future_all_iterator(Eina_Iterator *iterator)
{
Eina_Promise *p = eina_promise_all_iterator(iterator);
if (!p) return NULL;
return eina_future_new(p);
}
/**
* Creates a future that will be resolved once a future @p array is resolved.
* This is a helper over eina_promise_race_array()
*
* @param[in,out] array A future array, must be terminated with #EINA_FUTURE_SENTINEL
* @return A future.
* @see eina_promise_race_array()
* @see EINA_FUTURE_SENTINEL
*/
static inline Eina_Future *
eina_future_race_array(Eina_Future *array[])
{
Eina_Promise *p = eina_promise_race_array(array);
if (!p) return NULL;
return eina_future_new(p);
}
/**
* Used by eina_promise_race_array() and eina_promise_all_array() and
* friends to flag the end of the array.
*
* @see eina_promise_race_array()
* @see eina_promise_all_array()
*/
#define EINA_FUTURE_SENTINEL ((void *)(unsigned long)-1)
/**
* A syntactic sugar over eina_promise_race_array().
* Usage:
* @code
* promise = eina_promise_race(future1, future2, future3, future4);
* @endcode
* @see eina_promise_race_array()
*/
#define eina_promise_race(...) eina_promise_race_array((Eina_Future *[]){__VA_ARGS__, EINA_FUTURE_SENTINEL})
/**
* A syntactic sugar over eina_future_race_array().
* Usage:
* @code
* future = eina_future_race(future1, future2, future3, future4);
* @endcode
* @see eina_future_race_array()
*/
#define eina_future_race(...) eina_future_race_array((Eina_Future *[]){__VA_ARGS__, EINA_FUTURE_SENTINEL})
/**
* A syntactic sugar over eina_future_all_array().
* Usage:
* @code
* future = eina_future_all(future1, future2, future3, future4);
* @endcode
* @see eina_future_all_array()
*/
#define eina_future_all(...) eina_future_all_array((Eina_Future *[]){__VA_ARGS__, EINA_FUTURE_SENTINEL})
/**
* A syntactic sugar over eina_promise_all_array().
* Usage:
* @code
* promise = eina_promise_all(future1, future2, future3, future4);
* @endcode
* @see eina_promise_all_array()
*/
#define eina_promise_all(...) eina_promise_all_array((Eina_Future *[]){__VA_ARGS__, EINA_FUTURE_SENTINEL})
/**
* A syntactic sugar over eina_future_cb_easy_from_desc().
* Usage:
* @code
* future_desc = eina_future_cb_easy(_success_cb, _error_cb, _free_cb, EINA_VALUE_TYPE_INT, my_data);
* @endcode
* @see eina_future_cb_easy_from_desc()
*/
#define eina_future_cb_easy(...) eina_future_cb_easy_from_desc((Eina_Future_Cb_Easy_Desc){__VA_ARGS__})
/**
* A syntactic sugar over eina_future_chain_array().
* Usage:
* @code
* future = eina_future_chain(future, {.cb = _my_cb, .data = my_data}, {.cb = _my_another_cb, .data = NULL});
* @endcode
* @see eina_future_chain_array()
*/
#define eina_future_chain(_prev, ...) eina_future_chain_array(_prev, (Eina_Future_Desc[]){__VA_ARGS__, {.cb = NULL, .data = NULL}})
/**
* A syntactic sugar over eina_future_then_from_desc().
* Usage:
* @code
* future = eina_future_then(future, _my_cb, my_data);
* @endcode
* @see eina_future_then_from_desc()
* @see eina_future_then_easy()
*/
#define eina_future_then(_prev, ...) eina_future_then_from_desc(_prev, (Eina_Future_Desc){__VA_ARGS__})
/**
* A syntactic sugar over eina_future_cb_console_from_desc().
* Usage:
* @code
* desc = eina_future_cb_console(.prefix = "prefix", .suffix = "suffix");
* @endcode
* @see eina_future_cb_console_from_desc()
*/
#define eina_future_cb_console(...) eina_future_cb_console_from_desc((Eina_Future_Cb_Console_Desc){__VA_ARGS__})
/**
* A syntactic sugar over eina_future_cb_log_from_desc().
*
* This macro will set the following fields of the #Eina_Future_Cb_Log_Desc:
*
* @li Eina_Future_Cb_Log_Desc::file: The `__FILE__` function will be used.
* @li Eina_Future_Cb_Log_Desc::func: The `__func__` macro will be used.
* @li Eina_Future_Cb_Log_Desc::level: `EINA_LOG_LEVEL_DBG` will be used.
* @li Eina_Future_Cb_Log_Desc::domain: `EINA_LOG_DOMAIN_DEFAULT` will be used.
* @li Eina_Future_Cb_Log_Desc::line: The `__LINE__` macro will be used.
*
* Usage:
* @code
* desc = eina_future_cb_log_dbg(.prefix = "prefix", .suffix = "suffix");
* @endcode
* @see eina_future_cb_log_from_desc()
*/
#define eina_future_cb_log_dbg(_prefix, _suffix) \
eina_future_cb_log_from_desc((Eina_Future_Cb_Log_Desc){_prefix, _suffix, __FILE__, \
__func__, EINA_LOG_LEVEL_DBG, EINA_LOG_DOMAIN_DEFAULT, __LINE__})
/**
* A syntactic sugar over eina_future_cb_log_from_desc().
*
* This macro will set the following fields of the Eina_Future_Cb_Log_Desc:
*
* @li Eina_Future_Cb_Log_Desc::file: The __FILE__ function will be used.
* @li Eina_Future_Cb_Log_Desc::func: The __func__ macro will be used.
* @li Eina_Future_Cb_Log_Desc::level: EINA_LOG_LEVEL_CRITICAL will be used.
* @li Eina_Future_Cb_Log_Desc::domain: EINA_LOG_DOMAIN_DEFAULT will be used.
* @li Eina_Future_Cb_Log_Desc::line: The __LINE__ macro will be used.
*
* Usage:
* @code
* desc = eina_future_cb_log_crit(.prefix = "prefix", .suffix = "suffix");
* @endcode
* @see eina_future_cb_log_from_desc()
*/
#define eina_future_cb_log_crit(_prefix, _suffix) \
eina_future_cb_log_from_desc((Eina_Future_Cb_Log_Desc){_prefix, _suffix, __FILE__, \
__func__, EINA_LOG_LEVEL_CRITICAL, EINA_LOG_DOMAIN_DEFAULT, __LINE__})
/**
* A syntactic sugar over eina_future_cb_log_from_desc().
*
* This macro will set the following fields of the Eina_Future_Cb_Log_Desc:
*
* @li Eina_Future_Cb_Log_Desc::file: The __FILE__ function will be used.
* @li Eina_Future_Cb_Log_Desc::func: The __func__ macro will be used.
* @li Eina_Future_Cb_Log_Desc::level: EINA_LOG_LEVEL_ERR will be used.
* @li Eina_Future_Cb_Log_Desc::domain: EINA_LOG_DOMAIN_DEFAULT will be used.
* @li Eina_Future_Cb_Log_Desc::line: The __LINE__ macro will be used.
*
* Usage:
* @code
* desc = eina_future_cb_log_err(.prefix = "prefix", .suffix = "suffix");
* @endcode
* @see eina_future_cb_log_from_desc()
*/
#define eina_future_cb_log_err(_prefix, _suffix) \
eina_future_cb_log_from_desc((Eina_Future_Cb_Log_Desc){_prefix, _suffix, __FILE__, \
__func__, EINA_LOG_LEVEL_ERR, EINA_LOG_DOMAIN_DEFAULT, __LINE__})
/**
* A syntactic sugar over eina_future_cb_log_from_desc().
*
* This macro will set the following fields of the Eina_Future_Cb_Log_Desc:
*
* @li Eina_Future_Cb_Log_Desc::file: The __FILE__ function will be used.
* @li Eina_Future_Cb_Log_Desc::func: The __func__ macro will be used.
* @li Eina_Future_Cb_Log_Desc::level: EINA_LOG_LEVEL_INFO will be used.
* @li Eina_Future_Cb_Log_Desc::domain: EINA_LOG_DOMAIN_DEFAULT will be used.
* @li Eina_Future_Cb_Log_Desc::line: The __LINE__ macro will be used.
*
* Usage:
* @code
* desc = eina_future_cb_log_info(.prefix = "prefix", .suffix = "suffix");
* @endcode
* @see eina_future_cb_log_from_desc()
*/
#define eina_future_cb_log_info(_prefix, _suffix) \
eina_future_cb_log_from_desc((Eina_Future_Cb_Log_Desc){_prefix, _suffix, __FILE__, \
__func__, EINA_LOG_LEVEL_INFO, EINA_LOG_DOMAIN_DEFAULT, __LINE__})
/**
* A syntactic sugar over eina_future_cb_log_from_desc().
*
* This macro will set the following fields of the Eina_Future_Cb_Log_Desc:
*
* @li Eina_Future_Cb_Log_Desc::file: The __FILE__ function will be used.
* @li Eina_Future_Cb_Log_Desc::func: The __func__ macro will be used.
* @li Eina_Future_Cb_Log_Desc::level: EINA_LOG_LEVEL_WARN will be used.
* @li Eina_Future_Cb_Log_Desc::domain: EINA_LOG_DOMAIN_DEFAULT will be used.
* @li Eina_Future_Cb_Log_Desc::line: The __LINE__ macro will be used.
*
* Usage:
* @code
* desc = eina_future_cb_log_warn(.prefix = "prefix", .suffix = "suffix");
* @endcode
* @param _prefix PREFIX
* @param _suffix SUFFIX
* @returns SOMETHING
* @see eina_future_cb_log_from_desc()
*/
#define eina_future_cb_log_warn(_prefix, _suffix) \
eina_future_cb_log_from_desc((Eina_Future_Cb_Log_Desc){_prefix, _suffix, __FILE__, \
__func__, EINA_LOG_LEVEL_WARN, EINA_LOG_DOMAIN_DEFAULT, __LINE__})
/**
* A syntactic sugar over eina_future_then() and eina_future_cb_easy().
*
* Usage:
* @code
* f = eina_future_then_easy(f, .success = _success_cb, .success_type = EINA_VALUE_TYPE_DOUBLE, .data = NULL, );
* @endcode
* @see eina_future_then_from_desc()
* @see eina_future_easy()
* @see eina_future_then()
* @see eina_future_cb_easy_from_desc()
*/
#define eina_future_then_easy(_prev, ...) eina_future_then_from_desc(_prev, eina_future_cb_easy(__VA_ARGS__))
/**
* A syntactic sugar over eina_future_chain() and eina_future_cb_easy().
*
* Usage:
* @code
* f = eina_future_chain_easy(f, {.success = _success_cb, .success_type = EINA_VALUE_TYPE_DOUBLE, .data = NULL},
* { .success = _success2_cb }, {.error = error_cb});
* @endcode
* @see eina_future_chain_array()
* @see eina_future_easy()
* @see eina_future_chain_easy_array()
* @see eina_future_cb_easy_from_desc()
*/
#define eina_future_chain_easy(_prev, ...) eina_future_chain_easy_array(_prev, (Eina_Future_Cb_Easy_Desc[]) {__VA_ARGS__, {NULL, NULL, NULL, NULL, NULL}})
/**
* @}
*/
#ifdef __cplusplus
}
#endif
#endif