NAME

SQL::SyntaxModel::API_C - Describe the core C API for SQL::SyntaxModel

COPYRIGHT AND LICENSE

This file is part of the SQL::SyntaxModel library (libSQLSM).

SQL::SyntaxModel is Copyright (c) 1999-2003, Darren R. Duncan. All rights reserved. Address comments, suggestions, and bug reports to perl@DarrenDuncan.net, or visit "http://www.DarrenDuncan.net" for more information.

SQL::SyntaxModel is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License (GPL) version 2 as published by the Free Software Foundation (http://www.fsf.org/). You should have received a copy of the GPL as part of the SQL::SyntaxModel distribution, in the file named "LICENSE"; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA.

Any versions of SQL::SyntaxModel that you modify and distribute must carry prominent notices stating that you changed the files and the date of any changes, in addition to preserving this original copyright notice and other credits. SQL::SyntaxModel is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GPL for more details.

Linking SQL::SyntaxModel statically or dynamically with other modules is making a combined work based on SQL::SyntaxModel. Thus, the terms and conditions of the GPL cover the whole combination.

As a special exception, the copyright holders of SQL::SyntaxModel give you permission to link SQL::SyntaxModel with independent modules that are interfaces to or implementations of databases, regardless of the license terms of these independent modules, and to copy and distribute the resulting combined work under terms of your choice, provided that every copy of the combined work is accompanied by a complete copy of the source code of SQL::SyntaxModel (the version of SQL::SyntaxModel used to produce the combined work), being distributed under the terms of the GPL plus this exception. An independent module is a module which is not derived from or based on SQL::SyntaxModel, and which is fully useable when not linked to SQL::SyntaxModel in any form.

Note that people who make modified versions of SQL::SyntaxModel are not obligated to grant this special exception for their modified versions; it is their choice whether to do so. The GPL gives permission to release a modified version without this exception; this exception also makes it possible to release a modified version which carries forward this exception.

While it is by no means required, the copyright holders of SQL::SyntaxModel would appreciate being informed any time you create a modified version of SQL::SyntaxModel that you are willing to distribute, because that is a practical way of suggesting improvements to the standard version.

DESCRIPTION

The first releases of SQL::SyntaxModel are being implemented in pure Perl 5 thanks mainly to the heritage of original intention for using it with Perl programs, and otherwise to the fact that Perl is great for rapid prototyping, and has a large existing support for talking to databases, via the mature DBI framework, upon which to build. The longer term goal, however, is that SQL::SyntaxModel will be a pure ANSI C library which has optional bindings for multiple other languages, such as C++, Objective-C, Perl 5 (most importantly), Parrot and Perl 6, Python, PHP 4+, Java, and whatever else users want.

This has practical benefits such as compatability with databases for which there is no existing Perl support (all database drivers already have C support), and the ability to use it with any favorite programming language (which can bring in developers from those communities aka Parrot). Moreover, and most easily guessed, C is a lot more efficient in both CPU and memory considering the huge amount of data processing that SQL::SyntaxModel has to do. There are good reasons for things like LIBXML, GD, and DBI to be written in C which are applicable to SQL::SyntaxModel also. To that end, SQL::SyntaxModel is being designed from the ground up to be easily implemented in either Perl 5 or C, in both cases using an object-oriented conceptual process.

This document is a reference for the SQL::SyntaxModel C API that is being planned, consisting mainly of the C header files (struct definitions and function declarations) that the actual SQL::SyntaxModel core will end up being. After all, SQL::SyntaxModel is actually more of an interface or protocol definition than an actual implementation (although it includes those parts too). This is intended to give programmers an idea why the design is going the way that it is and why details that pure Perl programs don't necessarily have to know about are being considered. While pure C is not object oriented at all, the SQL::SyntaxModel API is designed to be as object-like as possible, such that each C struct for storing data will have a corresponding set of functions for interfacing with it, and the details for memory allocation and deallocation are handled by those also.

Note that the core libraries are intended to not have any external dependencies, except for data type handling libraries such as for string or sparse list handling. These classes do not do any I/O and they do not talk to the operating system, so they should be fully portable. All such things are relegated to extensions like the drivers.

Note also that the following should be considered almost-C rather than perfect C. The reasoning is that my experience in writing C is rusty at the moment and I haven't necessarily tried to compile these yet. For the moment they are just a type of documentation. They will be improved.

CONTENT OF libSQLSM.h

#ifndef _libSQLSM_h
#define _libSQLSM_h

/************************************************************************/
/* define some simple data types here */

typedef unsigned char sqlsm_dt_flag; /* stores only two values: 0 (false) or 1 (true) */
typedef unsigned char sqlsm_dt_octet; /* a generic eight-bit value */
typedef unsigned int  sqlsm_dt_scale; /* a precise integer between 0 and +4,294,967,295 */
typedef struct {
	sqlsm_dt_octet* string_data; /* not null terminated */
	sqlsm_dt_scale  used_octets;
	sqlsm_dt_scale  buff_octets;
} sqlsm_dt_label; /* an arbitrary sized UTF-8 character string, up to 4 gigabytes */

typedef enum { 
	SQLSM_LITP_NUM_INT , /* an exact integral or whole number */
	SQLSM_LITP_NUM_EXA , /* an exact decimal or fractional number */
	SQLSM_LITP_NUM_APR , /* an approximately represented number (eg: floating point) */
	SQLSM_LITP_STR_BIT , /* a string of bits or generic binary data */
	SQLSM_LITP_STR_CHAR, /* a string of characters (encoding spec separately) */
	SQLSM_LITP_BOOLEAN , /* two values: false or true */
	SQLSM_LITP_DATETIME, /* a date or time or both */
	SQLSM_LITP_INTERVAL  /* the amount of a time difference between two events */
} sqlsm_dt_literal_type; 

typedef enum { 
	SQLSM_CHENTP_UTF8 , /* unicode UTF-8  */
	SQLSM_CHENTP_UTF16, /* unicode UTF-16 */
	SQLSM_CHENTP_UTF32, /* unicode UTF-32 */
	SQLSM_CHENTP_ASC  , /* 8-bit ascii    */
	SQLSM_CHENTP_EBS    /* 8-bit ebsdic   */
	/* others will be added as appropriate */
} sqlsm_dt_char_enc_type; 

typedef struct {
	sqlsm_dt_literal_type  literal_type ; /* says how to interpret other struct properties */
	sqlsm_dt_scale         fixed_octets ; /* when literal is NUM_INT, NUM_APR; usu 1,2,4,8 */
	sqlsm_dt_char_enc_type char_enc_type; /* when literal is a STR_CHAR, how it is encoded */
	sqlsm_dt_octet*        literal_data ; /* not null terminated (or null part of trail junk) */
		/* above never ref'd by mult structs; struct itself can be ref'd by multiple vars */
	sqlsm_dt_scale         used_octets  ; /* doesn't include any trailing null with strings */
	sqlsm_dt_scale         buff_octets  ; /* does include or surpass any trailing null */
	/* other metadata may be added as appropriate; we may embed some fixed width type data */
	sqlsm_dt_scale         malcvar_refs ; /* count of refs to this struct from malloc'd vars */
} sqlsm_dt_literal; 

/************************************************************************/
/* define some simple data types here */

typedef enum {
	SQLSM_ENST_FOO, /* replace with semi-long actual list of generic enumerated value sets */
	SQLSM_ENST_BAR
} sqlsm_dt_enum_set;

typedef enum {
	SQLSM_ENIT_FOO, /* replace with long actual list of generic enumerated value set members */
	SQLSM_ENIT_BAR  /* this enum may need repl w const list or ml en union, if sev key can't have same int */
} sqlsm_dt_enum_item;

typedef enum {
	SQLSM_NDTP_FOO, /* replace with semi-short actual list of Node Types */
	SQLSM_NDTP_BAR
} sqlsm_dt_node_type;

typedef enum {
	SQLSM_NDATNM_FOO, /* replace with actual list of Node Attribute Names */
	SQLSM_NDATNM_BAR  /* this enum may need repl w const list or ml en union, if sev key can't have same int */
} sqlsm_dt_node_atnm;

typedef struct {
	sqlsm_dt_node_type  node_type  ; /* what type of Node this is */
	sqlsm_dt_scale      node_id    ; /* uniq ident for Node within Container+Type */
	sqlsm_dt_container* container  ; /* (ref) what Container this Node lives in, if any */
	sqlsm_dt_literal*   at_literals; /* (array) attrs of Node which are literals */
	sqlsm_dt_enum_item* at_enums   ; /* (array) attrs of Node which are enumerated values */
	union { /* when 'container' is null, use 'id'; when it isn't null, use 'ref' */
		sqlsm_dt_scale id ; /* id that has to be resolved to point to a Node */
		sqlsm_dt_node* ref; /* pointer to a Node */
	} id_or_ref*        at_nodes   ; /* (array of id_or_ref) attrs of Node which are Nodes */
		/* Note: if it would be better, we may combine all at* into one longer union list */
	sqlsm_dt_node_atnm  p_node_atnm; /* name of at_nodes attribute which is our primary parent */
	sqlsm_dt_flag       is_linked  ; /* true when actual refs in at_nodes are recipricated */
	sqlsm_dt_node_dll*  child_nodes; /* (list of refs) recipr to Nodes whose at_nodes ref to us */
} sqlsm_dt_node;

typedef struct {
	sqlsm_dt_node*     item; /* the actual Node represented by this list item */
	sqlsm_dt_node_dll* next; /* next Node in list, if any */
	sqlsm_dt_node_dll* prev; /* previous Node in list, if any */
} sqlsm_dt_node_dll; /* doubly-linked-list; use when number of list items not known in advance */

typedef struct {
	sqlsm_dt_node_dll** all_nodes; /* (array of list of refs) all Nodes in this Container */
		/* first dimension has one array element per allowed Node Type */
		/* second dimension is a doubly-linked-list of all Nodes having that Node Type */
		/* that list will probably get replaced with something more efficient at */
		/* making sure each Node Id is only used once; and list order doesn't matter here */
} sqlsm_dt_container;

/************************************************************************/
/* define some accessor functions here */

/************************************************************************/

#endif /* _libSQLSM_h */

SEE ALSO

perl(1), SQL::SyntaxModel, SQL::SyntaxModel::DataDictionary, SQL::SyntaxModel::XMLSchema, Rosetta, Rosetta::Framework.