Annotation of libwww/Library/src/SGML.html, revision 2.10
2.4 timbl 1: <HTML>
2: <HEAD>
3: <TITLE>SGML parse and stream definition for libwww</TITLE></HEAD>
2.1 timbl 4: <BODY>
5: <H1>SGML and Structured streams</H1>The SGML parser is a state machine.
6: It is called for every character<P>
7: of the input stream. The DTD data
8: structure contains pointers<P>
9: to functions which are called to
10: implement the actual effect of the<P>
11: text read. When these functions are
12: called, the attribute structures
13: pointed to by the DTD are valid,
14: and the function is passed a pointer
15: to the curent tag structure, and
16: an "element stack" which represents
17: the state of nesting within SGML
18: elements.<P>
19: The following aspects are from Dan
20: Connolly's suggestions: Binary search,
21: Strcutured object scheme basically,
22: SGML content enum type.<P>
23: (c) Copyright CERN 1991 - See Copyright.html
24: <PRE>#ifndef SGML_H
25: #define SGML_H
26:
27: #include "HTUtils.h"
28: #include "HTStream.h"
29:
30: </PRE>
31: <H2>SGML content types</H2>
32: <PRE>typedef enum _SGMLContent{
33: SGML_EMPTY, /* no content */
2.8 timbl 34: SGML_LITERAL, /* character data. Recognized exact close tag only.
2.1 timbl 35: Old www server compatibility only! Not SGML */
36: SGML_CDATA, /* character data. recognize </ only */
37: SGML_RCDATA, /* replaceable character data. recognize </ and &ref; */
38: SGML_MIXED, /* elements and parsed character data. recognize all markup */
39: SGML_ELEMENT /* any data found will be returned as an error*/
40: } SGMLContent;
41:
42:
43: typedef struct {
44: char * name; /* The (constant) name of the attribute */
45: /* Could put type info in here */
46: } attr;
47:
48:
49: /* A tag structure describes an SGML element.
50: ** -----------------------------------------
51: **
52: **
53: ** name is the string which comes after the tag opener "<".
54: **
55: ** attributes points to a zero-terminated array
56: ** of attribute names.
57: **
58: ** litteral determines how the SGML engine parses the charaters
59: ** within the element. If set, tag openers are ignored
60: ** except for that which opens a matching closing tag.
61: **
62: */
63: typedef struct _tag HTTag;
64: struct _tag{
65: char * name; /* The name of the tag */
66: attr * attributes; /* The list of acceptable attributes */
67: int number_of_attributes; /* Number of possible attributes */
68: SGMLContent contents; /* End only on end tag @@ */
69: };
70:
71:
72:
73:
74: /* DTD Information
75: ** ---------------
76: **
77: ** Not the whole DTD, but all this parser usues of it.
78: */
79: typedef struct {
2.2 timbl 80: HTTag * tags; /* Must be in strcmp order by name */
81: int number_of_tags;
82: CONST char ** entity_names; /* Must be in strcmp order by name */
83: int number_of_entities;
2.1 timbl 84: } SGML_dtd;
85:
2.9 frystyk 86: #define MAX_ATTRIBUTES 20 /* Max number of attributes per element */
2.1 timbl 87:
88: /* SGML context passed to parsers
89: */
90: typedef struct _HTSGMLContext *HTSGMLContext; /* Hidden */
91:
92:
93: /*__________________________________________________________________________
94: */
95:
2.6 timbl 96: </PRE>
97: <H2>Structured Object definition</H2>A structured object is something
98: which can reasonably be represented
99: in SGML. I'll rephrase that. A
100: structured object is am ordered tree-structured
101: arrangement of data which is representable
102: as text.The SGML parer outputs to
103: a Structured object. A Structured
104: object can output its contents to
105: another Structured Object. It's a
106: kind of typed stream. The architecure
107: is largely Dan Conolly's. Elements
108: and entities are passed to the sob
109: by number, implying a knowledge of
110: the DTD. Knowledge of the SGML syntax
111: is not here, though.<P>
112: Superclass: HTStream<P>
113: The creation methods will vary on
114: the type of Structured Object.Maybe
115: the callerData is enough info to
116: pass along.
117: <PRE>typedef struct _HTStructured HTStructured;
2.1 timbl 118:
119: typedef struct _HTStructuredClass{
120:
121: char* name; /* Just for diagnostics */
122:
123: void (*free) PARAMS((
124: HTStructured* me));
125:
2.4 timbl 126: void (*abort) PARAMS((
2.5 timbl 127: HTStructured* me,
128: HTError e));
2.1 timbl 129:
130: void (*put_character) PARAMS((
131: HTStructured* me,
132: char ch));
133:
134: void (*put_string) PARAMS((
135: HTStructured* me,
136: CONST char * str));
137:
138: void (*write) PARAMS((
2.2 timbl 139: HTStructured* me,
2.1 timbl 140: CONST char * str,
141: int len));
142:
143: void (*start_element) PARAMS((
144: HTStructured* me,
145: int element_number,
2.2 timbl 146: CONST BOOL* attribute_present,
147: CONST char** attribute_value));
2.1 timbl 148:
149: void (*end_element) PARAMS((
150: HTStructured* me,
151: int element_number));
152:
153: void (*put_entity) PARAMS((
154: HTStructured* me,
155: int entity_number));
156:
157: }HTStructuredClass;
158:
2.6 timbl 159: </PRE>
160: <H2>Find a Tag by Name</H2>Returns a pointer to the tag within
161: the DTD.
2.7 timbl 162: <PRE>extern HTTag * SGMLFindTag PARAMS((CONST SGML_dtd* dtd, CONST char * string));
2.1 timbl 163:
164:
2.6 timbl 165: </PRE>
2.10 ! timbl 166: <H2>Find a Attribute by Name</H2>Returns the number of the
! 167: atribute or -1 if failure.
! 168: <PRE>extern int SGMLFindAttribute PARAMS((HTTag* tag, CONST char * string));
! 169:
! 170:
! 171: </PRE>
2.6 timbl 172: <H2>Create an SGML parser</H2>
173: <PRE>/*
2.1 timbl 174: ** On entry,
175: ** dtd must point to a DTD structure as defined above
176: ** callbacks must point to user routines.
177: ** callData is returned in callbacks transparently.
178: ** On exit,
179: ** The default tag starter has been processed.
180: */
181:
182:
183: extern HTStream* SGML_new PARAMS((
184: CONST SGML_dtd * dtd,
185: HTStructured * target));
186:
2.2 timbl 187: extern CONST HTStreamClass SGMLParser;
2.1 timbl 188:
189:
190: #endif /* SGML_H */
2.7 timbl 191:
2.1 timbl 192:
2.3 timbl 193:
194:
2.4 timbl 195:
2.6 timbl 196:
197:
198:
199:
2.8 timbl 200:
201: </PRE></BODY>
2.4 timbl 202: </HTML>
Webmaster