+++ /dev/null
-/* xbt/datadesc.h - Describing the data you want to exchange */
-
-/* Copyright (c) 2004, 2005, 2006, 2007, 2009, 2010. The SimGrid Team.
- * All rights reserved. */
-
-/* This program is free software; you can redistribute it and/or modify it
- * under the terms of the license (GNU LGPL) which comes with this package. */
-
-#ifndef XBT_DATADESC_H
-#define XBT_DATADESC_H
-
-#include "xbt/misc.h" /* SG_BEGIN_DECL */
-#include "xbt/dynar.h" /* void_f_pvoid_t */
-
-SG_BEGIN_DECL()
-
-/** @addtogroup XBT_dd Data description
- * @brief Describing data to be exchanged
- *
- * Since XBT takes care of potential representation conversion when the platform is heterogeneous,
- * any data which transits on the network must be described beforehand.
- *
- * There are several possible interfaces for this, ranging from the really completely automatic parsing to
- * completely manual. Let's study each of them from the simplest to the more advanced:
- *
- * - Section \ref XBT_dd_basic presents how to retrieve and use an already described type.
- * - Section \ref XBT_dd_auto shows how to make XBT parse your type description automagically. This
- * is unfortunately not always possible (only works for some structures), but if it is for your data,
- * this is definitely the way to go.
- * - Section \ref XBT_dd_manual presents how to build a description manually. This is useful when you want
- * to describe an array or a pointer of pre-defined structures.
- * - You sometimes need to exchange informations between descriptions at send or receive time. This is
- * for example useful when your structure contains an array which size is given by another field of the
- * structure.
- * - Section \ref XBT_dd_cb_simple provides a simple interface to do so, allowing to share integers stored on a stack.
- * - Section \ref XBT_dd_cb_full provides a full featured interface to do so, but it may reveal somehow difficult to use.
- **/
-/** @defgroup XBT_dd_basic Basic operations on data descriptions
- * @ingroup XBT_dd
- * \htmlonly <!-- DOXYGEN_NAVBAR_LABEL="Basics" --> \endhtmlonly
- *
- * If you only want to send pre-existing types, simply retrieve the pre-defined description with
- * the \ref xbt_datadesc_by_name function. Existing types entail:
- * - char (both signed and unsigned)
- * - int (short, regular, long and long long, both signed and unsigned)
- * - float and double
- * - string (which is indeed a reference to a dynamically sized array of char, strlen being used to retrieve the size)
- *
- * Example:\verbatim xbt_datadesc_type_t i = xbt_datadesc_by_name("int");
- xbt_datadesc_type_t uc = xbt_datadesc_by_name("unsigned char");
- xbt_datadesc_type_t str = xbt_datadesc_by_name("string");\endverbatim
- *
- */
-/* @{ */
-/** @brief Opaque type describing a type description. */
-typedef struct s_xbt_datadesc_type *xbt_datadesc_type_t;
-
-/** \brief Search a type description from its name */
-XBT_PUBLIC(xbt_datadesc_type_t) xbt_datadesc_by_name(const char *name);
-XBT_PUBLIC(xbt_datadesc_type_t) xbt_datadesc_by_name_or_null(const char
- *name);
-
-/* @} */
-
-/** @defgroup XBT_dd_auto Automatic parsing of data descriptions
- * @ingroup XBT_dd
- * \htmlonly <!-- DOXYGEN_NAVBAR_LABEL="Automatic parsing" --> \endhtmlonly
- *
- * If you need to declare a new datatype, this is the simplest way to describe it to XBT. Simply
- * enclose its type definition into a \ref XBT_DEFINE_TYPE macro call, and you're set. Here is
- * an type declaration example: \verbatim XBT_DEFINE_TYPE(mytype,struct mytype {
- int myfirstfield;
- char mysecondfield;
- });\endverbatim
- * The type is then both copied verbatim into your source file and stored for further parsing. This allows
- * you to let XBT parse the exact version you are actually using in your program.
- * You can then retrieve the corresponding type description with \ref xbt_datadesc_by_symbol.
- * Don't worry too much for the performances, the type is only parsed once and a binary representation
- * is stored and used in any subsequent calls.
- *
- * If your structure contains any pointer, you have to explain XBT the size of the pointed array. This
- * can be 1 in the case of simple references, or more in the case of regular arrays. For that, use the
- * \ref XBT_ANNOTE macro within the type declaration you are passing to \ref XBT_DEFINE_TYPE. This macro
- * rewrites itself to nothing in the declaration (so they won't pollute the type definition copied verbatim
- * into your code), and give some information to XBT about your pointer.
-
- * XBT_ANNOTE takes two arguments being the key name and the key value. For now, the only accepted key name
- * is "size", to specify the length of the pointed array. It can either be:
- * - the string "1" (without the quote),
- * - the name of another field of the structure
- * - a sort of computed expression for multidimensional arrays (see below -- pay attention to the warnings below).
- *
- * Here is an example:\verbatim XBT_DEFINE_TYPE(s_clause,
- struct s_array {
- xbt_string_t name;
- struct s_array *father XBT_ANNOTE(size,1);
- int length;
- int *data XBT_ANNOTE(size,length);
- int rows;
- int cols;
- int *matrix XBT_ANNOTE(size,rows*cols);
- }
-;)\endverbatim
- * It specifies that the structure s_array contains six fields, that the \a name field is a classical null-terminated
- * char* string (#xbt_string_t is just an helper type defined exactly to help the parsing macro to specify the semantic of the pointer),
- * that \a father field is a simple reference, that the size of the array pointed by \a data is the \a length field, and that the
- * \a matrix field is an arraywhich size is the result of \a rows times \a cols.
- *
- * \warning Since XBT_DEFINE_TYPE is a macro, you shouldn't put any comma in your type definition
- * (comma separates macro args). For example, change \verbatim int a, b;\endverbatim to \verbatim int a;
-int b;\endverbatim
- *
- * \section xbt_dd_define \#define and fixed size array
- *
- * If you want to exchange arrays which size is given at compilation time by a
- * \#defined constant, you need to keep XBT informed. It would be done the
- * following way:
-
-\verbatim #define BLOCK_SIZE 32
-XBT_DEFINE_TYPE(s_toto,
-struct {
- double data[BLOCK_SIZE];
-} s_toto;)
-
-void register_messages() {
- xbt_datadesc_type_t toto_type;
-
- xbt_datadesc_set_const("BLOCK_SIZE",BLOCK_SIZE);
- toto_type = xbt_datadesc_by_symbol(s_toto);
-}\endverbatim
- *
- * The form <tt>xbt_datadesc_set_const("BLOCK_SIZE",BLOCK_SIZE);</tt> ensures
- * that when you change the definition of the constant, XBT keeps informed of
- * the right value. Passing the numerical value of the constant as second
- * argument would be a bad idea to that regard. Of course, the call to
- * xbt_datadesc_set_const() should come before any xbt_datadesc_by_symbol()
- * containing references to it.
- *
- * \section XBT_dd_multidim Defining multidimentional arrays
- *
- * The mecanism for multidimensional arrays is known to be fragile and cumbersome. If you want to use it,
- * you have to understand how it is implemented: the multiplication is performed using the sizes stack. In previous example,
- * a \ref xbt_datadesc_cb_push_int callback is added to the \a rows field and a \ref xbt_datadesc_cb_push_int_mult one is
- * added to \a cols. So, when the structure is sent, the \a rows field push its value onto the stack, then the \a cols field
- * retrieve this value from the stack, compute (and push) the multiplication value. The \a matrix field can then retrieve this
- * value by poping the array. There is several ways for this to go wrong:
- * - if the matrix field is placed before the sizes, the right value won't get pushed into the stack soon enough.
- * Reorder your structure fields if needed.
- * - if you write XBT_ANNOTE(size,cols*rows); in previous example (inverting rows and cols in annotation),
- * \a rows will be given a \ref xbt_datadesc_cb_push_int_mult. This cannot work since it will try to
- * pop the value which will be pushed by \a cols <i>afterward</i>.
- * - if you have more than one matrix in your structure, don't interleave the size. They are pushed/poped in the structure order.
- * - if some of the sizes are used in more than one matrix, you cannot use this mecanism -- sorry.
- *
- * If you cannot express your datadescs with this mechanism, you'll have to use the more advanced
- * (and somehow complex) one described in the \ref XBT_dd_cb_full.
- *
- * \section XBT_dd_multifile Projects spanning over multiple files
- *
- * XBT_DEFINE_TYPE declares some symbols to work, it needs some special
- * care when used in several files. In such case, you want the regular type
- * definition in all files, but the xbt specific symbol defined in only
- * one file. For example, consider the following xbt project sketch.
- *
-\verbatim #include <xbt/datadesc.h>
-
-XBT_DEFINE_TYPE(my_type,struct my_type {
- int a;
- int b;
- double c;
-});
-
-int client(int argc, char *argv[]) {
- ...
-}
-
-int server(int argc, char *argv[]) {
- ...
-}\endverbatim
- *
- * If you want to split this in two files (one for each kind of processes),
- * you need to put the XBT_DEFINE_TYPE block in a separate header (so that
- * each process kind see the associated C type definition). But
- * then you cannot include this right away in all files because the extra
- * symbols containing the XBT definition would be dupplicated.
- *
- * You thus have to decide in which C file the symbols will live. In that
- * file, include the header without restriction:
- *
-\verbatim #include "my_header.h"
-
-int client(int argc, char *argv[]) {
- ...
-}\endverbatim
-
- * And in the other files needing the C definitions without the extra XBT
- * symbols, declare the symbol XBT_DEFINE_TYPE_EXTERN before loading
- * xbt/datadesc.h:
- *
-\verbatim #define XBT_DEFINE_TYPE_EXTERN
-#include <xbt/datadesc.h>
-#include "my_header.h"
-
-int server(int argc, char *argv[]) {
- ...
-}\endverbatim
-
- *
- * Sometimes, the situation is even more complicated: There is some shared
- * messages that you want to see from every file, and some private messages
- * that you want to be defined only in one C file.
- * In that case, use the previous trick for common messages, and use
- * #XBT_DEFINE_TYPE_LOCAL for the private messages.
- *
- * For now, there is no way to have semi-private symbols (for example shared
- * in all files of a library), sorry. Use functions as interface to your
- * library instead of publishing directly the messages.
- *
- */
-/** @{ */
-
-
-/** @brief Automatically parse C code
- * @hideinitializer
- */
-#define XBT_DEFINE_TYPE(name,def) \
- const char * _xbt_this_type_symbol_does_not_exist__##name=#def; def
-
-#ifndef DOXYGEN_SKIP /* doxygen don't like macro fun too much */
-# ifdef XBT_DEFINE_TYPE_EXTERN
-# undef XBT_DEFINE_TYPE
-# define XBT_DEFINE_TYPE(name,def) def
-# undef XBT_DEFINE_TYPE_EXTERN
-# endif
-#endif
-
-/** @brief if this symbol is defined, the \a XBT_DEFINE_TYPE symbols live in another file.
- * @hideinitializer
- */
-#define XBT_DEFINE_TYPE_EXTERN 1
-/* leave the fun of declaring this to the user */
-#undef XBT_DEFINE_TYPE_EXTERN
-
-/** @brief Define a symbol to be automatically parsed, disregarding #XBT_DEFINE_TYPE_EXTERN
- * @hideinitializer
- *
- * Call this macro instead of #XBT_DEFINE_TYPE if you had to define #XBT_DEFINE_TYPE_EXTERN
- * to load some external symbols, but if you now want to automatically parse the content of
- * your private messages.
- */
-#define XBT_DEFINE_TYPE_LOCAL(name, def) \
- const char * _xbt_this_type_symbol_does_not_exist__##name=#def; def
-
-/** @brief Retrieve a datadesc which was previously parsed
- * @hideinitializer
- */
-#define xbt_datadesc_by_symbol(name) \
- (xbt_datadesc_by_name_or_null(#name) ? \
- xbt_datadesc_by_name_or_null(#name) : \
- xbt_datadesc_parse(#name, \
- _xbt_this_type_symbol_does_not_exist__##name) \
- )
-
-/** @def XBT_ANNOTE
- * @brief Add an annotation to a type to be automatically parsed
- */
-#define XBT_ANNOTE(key,val)
-
-/** @brief Defines the value of a define to the datatype parsing infrastructure
- */
-XBT_PUBLIC(void) xbt_datadesc_set_const(const char *name, int value);
-
-/* @} */
-
-XBT_PUBLIC(xbt_datadesc_type_t)
- xbt_datadesc_parse(const char *name, const char *C_statement);
-
-/** @defgroup XBT_dd_manual Simple manual data description
- * @ingroup XBT_dd
- *
- * Here are the functions to use if you want to declare your description manually.
- * The function names should be self-explanatory in most cases.
- *
- * You can add callbacks to the datatypes doing any kind of action you may want. Usually,
- * pre-send callbacks are used to prepare the type expedition while post-receive callbacks
- * are used to fix any issue after the receive.
- *
- * If your types are dynamic, you'll need to add some extra callback. For example, there is a
- * specific callback for the string type which is in charge of computing the length of the char
- * array. This is done with the cbps mechanism, explained in next section.
- *
- * If your types may contain pointer cycle, you must specify it to XBT using the @ref xbt_datadesc_cycle_set.
- *
- * Example:\verbatim
- typedef struct {
- unsigned char c1;
- unsigned long int l1;
- unsigned char c2;
- unsigned long int l2;
- } mystruct;
- [...]
- my_type=xbt_datadesc_struct("mystruct");
- xbt_datadesc_struct_append(my_type,"c1", xbt_datadesc_by_name("unsigned char"));
- xbt_datadesc_struct_append(my_type,"l1", xbt_datadesc_by_name("unsigned long"));
- xbt_datadesc_struct_append(my_type,"c2", xbt_datadesc_by_name("unsigned char"));
- xbt_datadesc_struct_append(my_type,"l2", xbt_datadesc_by_name("unsigned long int"));
- xbt_datadesc_struct_close(my_type);
-
- my_type=xbt_datadesc_ref("mystruct*", xbt_datadesc_by_name("mystruct"));
-
- [Use my_type to send pointers to mystruct data]\endverbatim
- */
-/* @{ */
-
-
-/** \brief Opaque type describing a type description callback persistent state. */
-typedef struct s_xbt_cbps *xbt_cbps_t;
-
-/* callbacks prototypes */
-/** \brief Prototype of type callbacks returning nothing. */
-typedef void (*xbt_datadesc_type_cb_void_t) (xbt_datadesc_type_t
- typedesc, xbt_cbps_t vars,
- void *data);
-/** \brief Prototype of type callbacks returning an int. */
-typedef int (*xbt_datadesc_type_cb_int_t) (xbt_datadesc_type_t
- typedesc, xbt_cbps_t vars,
- void *data);
-/** \brief Prototype of type callbacks selecting a type. */
-typedef
-
-
-
-
-
-xbt_datadesc_type_t(*xbt_datadesc_selector_t) (xbt_datadesc_type_t
- typedesc,
- xbt_cbps_t vars,
- void *data);
-
-
-/******************************************
- **** Declare datadescription yourself ****
- ******************************************/
-
-XBT_PUBLIC(xbt_datadesc_type_t) xbt_datadesc_struct(const char *name);
-XBT_PUBLIC(void) xbt_datadesc_struct_append(xbt_datadesc_type_t
- struct_type, const char *name,
- xbt_datadesc_type_t
- field_type);
-XBT_PUBLIC(void) xbt_datadesc_struct_close(xbt_datadesc_type_t
- struct_type);
-
-
-XBT_PUBLIC(xbt_datadesc_type_t) xbt_datadesc_union(const char *name,
- xbt_datadesc_type_cb_int_t
- selector);
-XBT_PUBLIC(void) xbt_datadesc_union_append(xbt_datadesc_type_t
- union_type, const char *name,
- xbt_datadesc_type_t
- field_type);
-XBT_PUBLIC(void) xbt_datadesc_union_close(xbt_datadesc_type_t
- union_type);
-
-
-XBT_PUBLIC(xbt_datadesc_type_t)
- xbt_datadesc_ref(const char *name, xbt_datadesc_type_t referenced_type);
-XBT_PUBLIC(xbt_datadesc_type_t)
- xbt_datadesc_copy(const char *name, xbt_datadesc_type_t copied_type);
-XBT_PUBLIC(xbt_datadesc_type_t)
- xbt_datadesc_ref_generic(const char *name,
- xbt_datadesc_selector_t selector);
-
-XBT_PUBLIC(xbt_datadesc_type_t)
- xbt_datadesc_array_fixed(const char *name,
- xbt_datadesc_type_t element_type,
- long int fixed_size);
-XBT_PUBLIC(xbt_datadesc_type_t)
- xbt_datadesc_array_dyn(const char *name,
- xbt_datadesc_type_t element_type,
- xbt_datadesc_type_cb_int_t dynamic_size);
-XBT_PUBLIC(xbt_datadesc_type_t)
- xbt_datadesc_ref_pop_arr(xbt_datadesc_type_t element_type);
-
-XBT_PUBLIC(xbt_datadesc_type_t)
- xbt_datadesc_dynar(xbt_datadesc_type_t elm_t, void_f_pvoid_t free_func);
-XBT_PUBLIC(xbt_datadesc_type_t)
- xbt_datadesc_matrix(xbt_datadesc_type_t elm_t,
- void_f_pvoid_t const free_f);
-
-/*********************************
- * Change stuff within datadescs *
- *********************************/
-
-/** \brief Specify that this type may contain cycles */
-XBT_PUBLIC(void) xbt_datadesc_cycle_set(xbt_datadesc_type_t type);
-/** \brief Specify that this type do not contain any cycles (default) */
-XBT_PUBLIC(void) xbt_datadesc_cycle_unset(xbt_datadesc_type_t type);
-/** \brief Add a pre-send callback to this datadesc. */
-XBT_PUBLIC(void) xbt_datadesc_cb_send(xbt_datadesc_type_t type,
- xbt_datadesc_type_cb_void_t pre);
-/** \brief Add a post-receive callback to this datadesc.*/
-XBT_PUBLIC(void) xbt_datadesc_cb_recv(xbt_datadesc_type_t type,
- xbt_datadesc_type_cb_void_t post);
-/** \brief Add a pre-send callback to the given field of the datadesc */
-XBT_PUBLIC(void) xbt_datadesc_cb_field_send(xbt_datadesc_type_t type,
- const char *field_name,
- xbt_datadesc_type_cb_void_t
- pre);
-/** \brief Add a post-receive callback to the given field of the datadesc */
-XBT_PUBLIC(void) xbt_datadesc_cb_field_recv(xbt_datadesc_type_t type,
- const char *field_name,
- xbt_datadesc_type_cb_void_t
- post);
-/** \brief Add a pre-send callback to the given field resulting in its value to be pushed */
-XBT_PUBLIC(void) xbt_datadesc_cb_field_push(xbt_datadesc_type_t type,
- const char *field_name);
-/** \brief Add a pre-send callback to the given field resulting in its value multiplied to any previously pushed value and then pushed back */
-XBT_PUBLIC(void)
-xbt_datadesc_cb_field_push_multiplier(xbt_datadesc_type_t type,
- const char
- *field_name);
-
-/******************************
- * Get stuff within datadescs *
- ******************************/
-/** \brief Returns the name of a datadescription */
-XBT_PUBLIC(const char *) xbt_datadesc_get_name(xbt_datadesc_type_t ddt);
-/** \brief Returns the identifier of a datadescription */
-XBT_PUBLIC(int) xbt_datadesc_get_id(xbt_datadesc_type_t ddt);
-
-/* @} */
-
-/** @defgroup XBT_dd_cb_simple Data description with Callback Persistant State: Simple push/pop mechanism
- * @ingroup XBT_dd
- *
- * Sometimes, one of the callbacks need to leave information for the next ones. If this is a simple integer (such as
- * an array size), you can use the functions described here. If not, you'll have to play with the complete cbps interface.
- *
- * \htmlonly <!-- DOXYGEN_NAVBAR_LABEL="Simple push/pop Callback State" -->\endhtmlonly
- *
- * Here is an example:\verbatim
-struct s_array {
- int length;
- int *data;
-}
-[...]
-my_type=xbt_datadesc_struct("s_array");
-xbt_datadesc_struct_append(my_type,"length", xbt_datadesc_by_name("int"));
-xbt_datadesc_cb_field_send (my_type, "length", xbt_datadesc_cb_push_int);
-
-xbt_datadesc_struct_append(my_type,"data",
- xbt_datadesc_array_dyn ("s_array::data",xbt_datadesc_by_name("int"), xbt_datadesc_cb_pop));
-xbt_datadesc_struct_close(my_type);
-\endverbatim
-
- *
- * The *_mult versions are intended for multi-dimensional arrays: They multiply their value to the previously pushed one
- * (by another field callback) and push the result of the multiplication back. An example of use follows. Please note
- * that the first field needs a regular push callback, not a multiplier one. Think of it as a stacked calculator (man dc(1)).\verbatim
-struct s_matrix {
- int row;
- int col;
- int *data;
-}
-[...]
-my_type=xbt_datadesc_struct("s_matrix");
-xbt_datadesc_struct_append(my_type,"row", xbt_datadesc_by_name("int"));
-xbt_datadesc_cb_field_send (my_type, "length", xbt_datadesc_cb_push_int);
-xbt_datadesc_struct_append(my_type,"col", xbt_datadesc_by_name("int"));
-xbt_datadesc_cb_field_send (my_type, "length", xbt_datadesc_cb_push_int_mult);
-
-xbt_datadesc_struct_append(my_type,"data",
- xbt_datadesc_array_dyn ("s_matrix::data",xbt_datadesc_by_name("int"), xbt_datadesc_cb_pop));
-xbt_datadesc_struct_close(my_type);
-\endverbatim
-
- */
-/* @{ */
-
-XBT_PUBLIC(void) xbt_cbps_i_push(xbt_cbps_t ps, int val);
-XBT_PUBLIC(int) xbt_cbps_i_pop(xbt_cbps_t ps);
-
-XBT_PUBLIC(int) xbt_datadesc_cb_pop(xbt_datadesc_type_t typedesc,
- xbt_cbps_t vars, void *data);
-
-XBT_PUBLIC(void) xbt_datadesc_cb_push_int(xbt_datadesc_type_t typedesc,
- xbt_cbps_t vars, void *data);
-XBT_PUBLIC(void) xbt_datadesc_cb_push_uint(xbt_datadesc_type_t typedesc,
- xbt_cbps_t vars, void *data);
-XBT_PUBLIC(void) xbt_datadesc_cb_push_lint(xbt_datadesc_type_t typedesc,
- xbt_cbps_t vars, void *data);
-XBT_PUBLIC(void) xbt_datadesc_cb_push_ulint(xbt_datadesc_type_t typedesc,
- xbt_cbps_t vars, void *data);
-
-XBT_PUBLIC(void) xbt_datadesc_cb_push_int_mult(xbt_datadesc_type_t
- typedesc, xbt_cbps_t vars,
- void *data);
-XBT_PUBLIC(void) xbt_datadesc_cb_push_uint_mult(xbt_datadesc_type_t
- typedesc,
- xbt_cbps_t vars,
- void *data);
-XBT_PUBLIC(void) xbt_datadesc_cb_push_lint_mult(xbt_datadesc_type_t
- typedesc,
- xbt_cbps_t vars,
- void *data);
-XBT_PUBLIC(void) xbt_datadesc_cb_push_ulint_mult(xbt_datadesc_type_t
- typedesc,
- xbt_cbps_t vars,
- void *data);
-
-
-/* @} */
-
-/** @defgroup XBT_dd_cb_full Data description with Callback Persistant State: Full featured interface
- * @ingroup XBT_dd
- *
- * Sometimes, one of the callbacks need to leave information for the next
- * ones. If the simple push/pop mechanism introduced in previous section
- * isn't enough, you can always use this full featured one. The bad point is
- * that it is quite badly documented...
- *
- * \htmlonly <!-- DOXYGEN_NAVBAR_LABEL="Full featured Callback State" -->\endhtmlonly
- *
- */
-
-/* @{ */
-
-XBT_PUBLIC(void) xbt_cbps_v_pop(xbt_cbps_t ps, const char *name,
- /* OUT */ xbt_datadesc_type_t * ddt,
- /* OUT */ void **res);
-XBT_PUBLIC(void) xbt_cbps_v_push(xbt_cbps_t ps,
- const char *name,
- void *data, xbt_datadesc_type_t ddt);
-XBT_PUBLIC(void) xbt_cbps_v_set(xbt_cbps_t ps,
- const char *name,
- void *data, xbt_datadesc_type_t ddt);
-
-XBT_PUBLIC(void *) xbt_cbps_v_get(xbt_cbps_t ps, const char *name,
- /* OUT */ xbt_datadesc_type_t * ddt);
-
-XBT_PUBLIC(void) xbt_cbps_block_begin(xbt_cbps_t ps);
-XBT_PUBLIC(void) xbt_cbps_block_end(xbt_cbps_t ps);
-
-/* @} */
-/* @} */
-
-
-/*******************************
- **** About data convertion ****
- *******************************/
-XBT_PUBLIC(int) xbt_arch_selfid(void); /* ID of this arch */
-
-
-/*****************************
- **** NWS datadescription * FIXME: obsolete?
- *****************************/
-
-/**
- * Basic types we can embeed in DataDescriptors.
- */
-typedef enum
- { CHAR_TYPE, DOUBLE_TYPE, FLOAT_TYPE, INT_TYPE, LONG_TYPE, SHORT_TYPE,
- UNSIGNED_INT_TYPE, UNSIGNED_LONG_TYPE, UNSIGNED_SHORT_TYPE, STRUCT_TYPE
-} DataTypes;
-#define SIMPLE_TYPE_COUNT 9
-
-/** \brief Describe a collection of data.
- *
-** A description of a collection of \a type data. \a repetitions is used only
-** for arrays; it contains the number of elements. \a offset is used only for
-** struct members in host format; it contains the offset of the member from the
-** beginning of the struct, taking into account internal padding added by the
-** compiler for alignment purposes. \a members, \a length, and \a tailPadding are
-** used only for STRUCT_TYPE data; the \a length -long array \a members describes
-** the members of the nested struct, and \a tailPadding indicates how many
-** padding bytes the compiler adds to the end of the structure.
-*/
-
-typedef struct DataDescriptorStruct {
- DataTypes type;
- size_t repetitions;
- size_t offset;
- /*@null@ */ struct DataDescriptorStruct *members;
- size_t length;
- size_t tailPadding;
-} DataDescriptor;
-/** DataDescriptor for an array */
-#define SIMPLE_DATA(type,repetitions) \
- {type, repetitions, 0, NULL, 0, 0}
-/** DataDescriptor for an structure member */
-#define SIMPLE_MEMBER(type,repetitions,offset) \
- {type, repetitions, offset, NULL, 0, 0}
-/** DataDescriptor for padding bytes */
-#define PAD_BYTES(structType,lastMember,memberType,repetitions) \
- sizeof(structType) - offsetof(structType, lastMember) - \
- sizeof(memberType) * repetitions
-
-XBT_PUBLIC(xbt_datadesc_type_t)
- xbt_datadesc_import_nws(const char *name,
- const DataDescriptor * desc,
- unsigned long howmany);
-
-
-SG_END_DECL()
-#endif /* XBT_DATADESC_H */
