/* Copyright 2008-2013 David Robillard Permission to use, copy, modify, and/or distribute this software for any purpose with or without fee is hereby granted, provided that the above copyright notice and this permission notice appear in all copies. THIS SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. */ /** @file util.h Helper functions for the LV2 Atom extension. Note these functions are all static inline, do not take their address. This header is non-normative, it is provided for convenience. */ #ifndef LV2_ATOM_UTIL_H #define LV2_ATOM_UTIL_H #include #include #include #include "atom.h" #ifdef __cplusplus extern "C" { #else # include #endif /** Pad a size to 64 bits. */ static inline uint32_t lv2_atom_pad_size(uint32_t size) { return (size + 7U) & (~7U); } /** Return the total size of @p atom, including the header. */ static inline uint32_t lv2_atom_total_size(const LV2_Atom* atom) { return (uint32_t)sizeof(LV2_Atom) + atom->size; } /** Return true iff @p atom is null. */ static inline bool lv2_atom_is_null(const LV2_Atom* atom) { return !atom || (atom->type == 0 && atom->size == 0); } /** Return true iff @p a is equal to @p b. */ static inline bool lv2_atom_equals(const LV2_Atom* a, const LV2_Atom* b) { return (a == b) || ((a->type == b->type) && (a->size == b->size) && !memcmp(a + 1, b + 1, a->size)); } /** @name Sequence Iterator @{ */ /** Get an iterator pointing to the first event in a Sequence body. */ static inline const LV2_Atom_Event* lv2_atom_sequence_begin(const LV2_Atom_Sequence_Body* body) { return (const LV2_Atom_Event*)(body + 1); } /** Get an iterator pointing to the end of a Sequence body. */ static inline LV2_Atom_Event* lv2_atom_sequence_end(LV2_Atom_Sequence_Body* body, uint32_t size) { return (LV2_Atom_Event*)((uint8_t*)body + lv2_atom_pad_size(size)); } /** Return true iff @p i has reached the end of @p body. */ static inline bool lv2_atom_sequence_is_end(const LV2_Atom_Sequence_Body* body, uint32_t size, const LV2_Atom_Event* i) { return (const uint8_t*)i >= ((const uint8_t*)body + size); } /** Return an iterator to the element following @p i. */ static inline const LV2_Atom_Event* lv2_atom_sequence_next(const LV2_Atom_Event* i) { return (const LV2_Atom_Event*)((const uint8_t*)i + sizeof(LV2_Atom_Event) + lv2_atom_pad_size(i->body.size)); } /** A macro for iterating over all events in a Sequence. @param seq The sequence to iterate over @param iter The name of the iterator This macro is used similarly to a for loop (which it expands to), e.g.: @code LV2_ATOM_SEQUENCE_FOREACH(sequence, ev) { // Do something with ev (an LV2_Atom_Event*) here... } @endcode */ #define LV2_ATOM_SEQUENCE_FOREACH(seq, iter) \ for (const LV2_Atom_Event* (iter) = lv2_atom_sequence_begin(&(seq)->body); \ !lv2_atom_sequence_is_end(&(seq)->body, (seq)->atom.size, (iter)); \ (iter) = lv2_atom_sequence_next(iter)) /** Like LV2_ATOM_SEQUENCE_FOREACH but for a headerless sequence body. */ #define LV2_ATOM_SEQUENCE_BODY_FOREACH(body, size, iter) \ for (const LV2_Atom_Event* (iter) = lv2_atom_sequence_begin(body); \ !lv2_atom_sequence_is_end(body, size, (iter)); \ (iter) = lv2_atom_sequence_next(iter)) /** @} @name Sequence Utilities @{ */ /** Clear all events from @p sequence. This simply resets the size field, the other fields are left untouched. */ static inline void lv2_atom_sequence_clear(LV2_Atom_Sequence* seq) { seq->atom.size = sizeof(LV2_Atom_Sequence_Body); } /** Append an event at the end of @p sequence. @param seq Sequence to append to. @param capacity Total capacity of the sequence atom (e.g. as set by the host for sequence output ports). @param event Event to write. @return A pointer to the newly written event in @p seq, or NULL on failure (insufficient space). */ static inline LV2_Atom_Event* lv2_atom_sequence_append_event(LV2_Atom_Sequence* seq, uint32_t capacity, const LV2_Atom_Event* event) { const uint32_t total_size = (uint32_t)sizeof(*event) + event->body.size; if (capacity - seq->atom.size < total_size) { return NULL; } LV2_Atom_Event* e = lv2_atom_sequence_end(&seq->body, seq->atom.size); memcpy(e, event, total_size); seq->atom.size += lv2_atom_pad_size(total_size); return e; } /** @} @name Tuple Iterator @{ */ /** Get an iterator pointing to the first element in @p tup. */ static inline const LV2_Atom* lv2_atom_tuple_begin(const LV2_Atom_Tuple* tup) { return (const LV2_Atom*)(LV2_ATOM_BODY_CONST(tup)); } /** Return true iff @p i has reached the end of @p body. */ static inline bool lv2_atom_tuple_is_end(const void* body, uint32_t size, const LV2_Atom* i) { return (const uint8_t*)i >= ((const uint8_t*)body + size); } /** Return an iterator to the element following @p i. */ static inline const LV2_Atom* lv2_atom_tuple_next(const LV2_Atom* i) { return (const LV2_Atom*)( (const uint8_t*)i + sizeof(LV2_Atom) + lv2_atom_pad_size(i->size)); } /** A macro for iterating over all properties of a Tuple. @param tuple The tuple to iterate over @param iter The name of the iterator This macro is used similarly to a for loop (which it expands to), e.g.: @code LV2_ATOMO_TUPLE_FOREACH(tuple, elem) { // Do something with elem (an LV2_Atom*) here... } @endcode */ #define LV2_ATOM_TUPLE_FOREACH(tuple, iter) \ for (const LV2_Atom* (iter) = lv2_atom_tuple_begin(tuple); \ !lv2_atom_tuple_is_end(LV2_ATOM_BODY_CONST(tuple), (tuple)->size, (iter)); \ (iter) = lv2_atom_tuple_next(iter)) /** Like LV2_ATOM_TUPLE_FOREACH but for a headerless tuple body. */ #define LV2_ATOM_TUPLE_BODY_FOREACH(body, size, iter) \ for (const LV2_Atom* (iter) = (const LV2_Atom*)body; \ !lv2_atom_tuple_is_end(body, size, (iter)); \ (iter) = lv2_atom_tuple_next(iter)) /** @} @name Object Iterator @{ */ /** Return a pointer to the first property in @p body. */ static inline const LV2_Atom_Property_Body* lv2_atom_object_begin(const LV2_Atom_Object_Body* body) { return (const LV2_Atom_Property_Body*)(body + 1); } /** Return true iff @p i has reached the end of @p obj. */ static inline bool lv2_atom_object_is_end(const LV2_Atom_Object_Body* body, uint32_t size, const LV2_Atom_Property_Body* i) { return (const uint8_t*)i >= ((const uint8_t*)body + size); } /** Return an iterator to the property following @p i. */ static inline const LV2_Atom_Property_Body* lv2_atom_object_next(const LV2_Atom_Property_Body* i) { const LV2_Atom* const value = (const LV2_Atom*)( (const uint8_t*)i + 2 * sizeof(uint32_t)); return (const LV2_Atom_Property_Body*)( (const uint8_t*)i + lv2_atom_pad_size( (uint32_t)sizeof(LV2_Atom_Property_Body) + value->size)); } /** A macro for iterating over all properties of an Object. @param obj The object to iterate over @param iter The name of the iterator This macro is used similarly to a for loop (which it expands to), e.g.: @code LV2_ATOM_OBJECT_FOREACH(object, i) { // Do something with prop (an LV2_Atom_Property_Body*) here... } @endcode */ #define LV2_ATOM_OBJECT_FOREACH(obj, iter) \ for (const LV2_Atom_Property_Body* (iter) = lv2_atom_object_begin(&(obj)->body); \ !lv2_atom_object_is_end(&(obj)->body, (obj)->atom.size, (iter)); \ (iter) = lv2_atom_object_next(iter)) /** Like LV2_ATOM_OBJECT_FOREACH but for a headerless object body. */ #define LV2_ATOM_OBJECT_BODY_FOREACH(body, size, iter) \ for (const LV2_Atom_Property_Body* (iter) = lv2_atom_object_begin(body); \ !lv2_atom_object_is_end(body, size, (iter)); \ (iter) = lv2_atom_object_next(iter)) /** @} @name Object Query @{ */ /** A single entry in an Object query. */ typedef struct { uint32_t key; /**< Key to query (input set by user) */ const LV2_Atom** value; /**< Found value (output set by query function) */ } LV2_Atom_Object_Query; static const LV2_Atom_Object_Query LV2_ATOM_OBJECT_QUERY_END = { 0, NULL }; /** Get an object's values for various keys. The value pointer of each item in @p query will be set to the location of the corresponding value in @p object. Every value pointer in @p query MUST be initialised to NULL. This function reads @p object in a single linear sweep. By allocating @p query on the stack, objects can be "queried" quickly without allocating any memory. This function is realtime safe. This function can only do "flat" queries, it is not smart enough to match variables in nested objects. For example: @code const LV2_Atom* name = NULL; const LV2_Atom* age = NULL; LV2_Atom_Object_Query q[] = { { urids.eg_name, &name }, { urids.eg_age, &age }, LV2_ATOM_OBJECT_QUERY_END }; lv2_atom_object_query(obj, q); // name and age are now set to the appropriate values in obj, or NULL. @endcode */ static inline int lv2_atom_object_query(const LV2_Atom_Object* object, LV2_Atom_Object_Query* query) { int matches = 0; int n_queries = 0; /* Count number of query keys so we can short-circuit when done */ for (LV2_Atom_Object_Query* q = query; q->key; ++q) { ++n_queries; } LV2_ATOM_OBJECT_FOREACH(object, prop) { for (LV2_Atom_Object_Query* q = query; q->key; ++q) { if (q->key == prop->key && !*q->value) { *q->value = &prop->value; if (++matches == n_queries) { return matches; } break; } } } return matches; } /** Body only version of lv2_atom_object_get(). */ static inline int lv2_atom_object_body_get(uint32_t size, const LV2_Atom_Object_Body* body, ...) { int matches = 0; int n_queries = 0; /* Count number of keys so we can short-circuit when done */ va_list args; va_start(args, body); for (n_queries = 0; va_arg(args, uint32_t); ++n_queries) { if (!va_arg(args, const LV2_Atom**)) { return -1; } } va_end(args); LV2_ATOM_OBJECT_BODY_FOREACH(body, size, prop) { va_start(args, body); for (int i = 0; i < n_queries; ++i) { uint32_t qkey = va_arg(args, uint32_t); const LV2_Atom** qval = va_arg(args, const LV2_Atom**); if (qkey == prop->key && !*qval) { *qval = &prop->value; if (++matches == n_queries) { return matches; } break; } } va_end(args); } return matches; } /** Variable argument version of lv2_atom_object_query(). This is nicer-looking in code, but a bit more error-prone since it is not type safe and the argument list must be terminated. The arguments should be a series of uint32_t key and const LV2_Atom** value pairs, terminated by a zero key. The value pointers MUST be initialized to NULL. For example: @code const LV2_Atom* name = NULL; const LV2_Atom* age = NULL; lv2_atom_object_get(obj, uris.name_key, &name, uris.age_key, &age, 0); @endcode */ static inline int lv2_atom_object_get(const LV2_Atom_Object* object, ...) { int matches = 0; int n_queries = 0; /* Count number of keys so we can short-circuit when done */ va_list args; va_start(args, object); for (n_queries = 0; va_arg(args, uint32_t); ++n_queries) { if (!va_arg(args, const LV2_Atom**)) { return -1; } } va_end(args); LV2_ATOM_OBJECT_FOREACH(object, prop) { va_start(args, object); for (int i = 0; i < n_queries; ++i) { uint32_t qkey = va_arg(args, uint32_t); const LV2_Atom** qval = va_arg(args, const LV2_Atom**); if (qkey == prop->key && !*qval) { *qval = &prop->value; if (++matches == n_queries) { return matches; } break; } } va_end(args); } return matches; } /** @} */ #ifdef __cplusplus } /* extern "C" */ #endif #endif /* LV2_ATOM_UTIL_H */