/simpletypesystem/trunk

/*

Copyright (c) 2013-2014 Gustav Hartvigsson

Permission is hereby granted, free of charge, to any person obtaining a copy

of this software and associated documentation files (the "Software"), to deal

in the Software without restriction, including without limitation the rights

to use, copy, modify, merge, publish, distribute, sublicense, and/or sell

copies of the Software, and to permit persons to whom the Software is

furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in

all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR

IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,

FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE

AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER

LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,

OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN

THE SOFTWARE.

*/

/** @file */

#ifndef __H_ERROR__

#define __H_ERROR__

#include "baseobject.h"

#include "defs.h"

#include "utils.h"

#include "Func.h"

#include <limits.h>

#include <stdlib.h>

#include <stdio.h>

#include <string.h>

BEGIN_DECLS

/**

 * @defgroup SError SError

 * @addtogroup SError

 * @{

 * An opaque datastructure that holds the SError's private data.

 */

typedef struct SErrorPrivate SErrorPrivate;

typedef void (* SErrorDomainToString)(sint error_id, schar * error_msg);

/**

 * An SError is a data structure that inherits from SBaseObjectInstance.

 * 

 * An SError represents an error that can occur.

 */

typedef struct

SError {

  SObject parent;

  SErrorPrivate * priv; /** Pimple pointer to the private data. */

} SError;

/**

 * The error class is not changed, no extra methods are needed.

 */

typedef struct

SErrorClass {

  SObjectClass parent_class;

} SErrorClass;

#define S_ERROR(o) (SError *)(o)

#define S_ERROR_CLASS(k) (SErrorClass *)(k)

/** @brief

 * The different error types.

 */

typedef enum {

  S_ERROR_NONE = 0, /**< Not on error */

  S_ERROR_INPUT_OUTPUT,/**< An I/O error */

  S_ERROR_OVERFLOW, /**< An Overflow error */

  S_ERROR_TYPE_ERROR, /** Some generic type error. */

  S_ERROR_UNUSED_0,

  S_ERROR_UNUSED_1,

  S_ERROR_UNUSED_2,

  S_ERROR_OTHER, /**< Some unknowned error */

  S_ERROR_NULL /**< An NULL error */

} SErrorType;

/**

 * @var static char * SErrorTypeName[]

 * @breif

 * Holds the names of the string version of SErrorType.

 */

static schar * SErrorTypeName[] UNUSED = {

  "NONE",

  "INPUT/OUTPUT",

  "OVERFLOW",

  "TYPE_ERROR",

  "UNUSED_0/INVALID",

  "UNUSED_1/INVALID",

  "UNUSED_2/INVALID",

  "OTHER",

  "NULL",

  0x0,

  0x0

};

/**

 * @brief

 * Function to get the same of an error.

 * 

 * For use in bindings.

 *

 * @param k The key to look up.

 */

schar *

s_error_get_name (SErrorType k);

/**

 * the constructor for the an SError

 */

SError *

s_error_new (SErrorType error, schar * message);

/**

 * This function calles the deinitize method of the object.

 */

void

s_error_free (SError * self);

/**

 * This function returns the ErrorType of on object.

 */

SErrorType

s_error_get_error_type (SError * self);

/**

 * This function prints the current error to stdout.

 */

void

s_error_print_error (SError * self);

#if 0

TODO

##########

I have been thinking about Error Propagation in SSTS and GLib.

I noticed that errors in GLib does not rely propagate, you can do 

g_error_propagate () but that assumes that the error that is fed to the function

is an array. This assumption can never be made and makes it very complex to deal

with different ways of doing things in different parts of a program (GError * vs

GError **).

So what I am thinking of doing is a little more complex: have explicit 

registration of error domains (s_error_domain_register (schar * name,

FuncPointer to_string), s_error_domain_is_registered (schar *)).

The other thing I was thinking is that errors should have implicit propagation.

(IE: no s_error_new () only s_error_append ()) and have the propagation

information stored in an array. (ID and message).

The problem with this is that where do you call s_error_domain_register ()?

In the constructor of an SObject would be a great place to put it, if not for

the problem that we have to call s_error_domain_is_registered when a new

instance of an object is created.

In the main function is just tedious, Wrapping the main function in something

like an S_MAIN macro could work and have an internal

s_library_register_error_domains function somewhere and require the user to

implement an s_user_register_error_domains function.

This puts even more work on the shoulders of the developers that uses the

library.

It seems that no matter what I do (A true propagating error system or

something like GError from GLib) I put extra work in the lap of the developer

that uses the library.

Commets?

##########

void

s_error_domain_register (schar * domain, SErrorDomainToString func);

sboolean

s_error_domain_is_registerd (schar * domain);

#endif

/** @} */

END_DECLS

#endif