| 1 |
/* |
|---|
| 2 |
www.sourceforge.net/projects/tinyxpath |
|---|
| 3 |
Copyright (c) 2002-2004 Yves Berquin (yvesb@users.sourceforge.net) |
|---|
| 4 |
|
|---|
| 5 |
This software is provided 'as-is', without any express or implied |
|---|
| 6 |
warranty. In no event will the authors be held liable for any |
|---|
| 7 |
damages arising from the use of this software. |
|---|
| 8 |
|
|---|
| 9 |
Permission is granted to anyone to use this software for any |
|---|
| 10 |
purpose, including commercial applications, and to alter it and |
|---|
| 11 |
redistribute it freely, subject to the following restrictions: |
|---|
| 12 |
|
|---|
| 13 |
1. The origin of this software must not be misrepresented; you must |
|---|
| 14 |
not claim that you wrote the original software. If you use this |
|---|
| 15 |
software in a product, an acknowledgment in the product documentation |
|---|
| 16 |
would be appreciated but is not required. |
|---|
| 17 |
|
|---|
| 18 |
2. Altered source versions must be plainly marked as such, and |
|---|
| 19 |
must not be misrepresented as being the original software. |
|---|
| 20 |
|
|---|
| 21 |
3. This notice may not be removed or altered from any source |
|---|
| 22 |
distribution. |
|---|
| 23 |
*/ |
|---|
| 24 |
/** |
|---|
| 25 |
\file xpath_stream.h |
|---|
| 26 |
\author Yves Berquin |
|---|
| 27 |
Specialized byte stream for the TinyXPath project |
|---|
| 28 |
*/ |
|---|
| 29 |
|
|---|
| 30 |
|
|---|
| 31 |
|
|---|
| 32 |
|
|---|
| 33 |
private import xpath.byte_stream; |
|---|
| 34 |
private import xpath.xpath_syntax; |
|---|
| 35 |
private import std.stream; |
|---|
| 36 |
private import std.stdio; |
|---|
| 37 |
private import std.string; |
|---|
| 38 |
private import xpath.lex_util; |
|---|
| 39 |
|
|---|
| 40 |
private |
|---|
| 41 |
{ |
|---|
| 42 |
class token_redef : public token_syntax_decoder |
|---|
| 43 |
{ |
|---|
| 44 |
protected : |
|---|
| 45 |
/// pointer to an xpath_stream instance |
|---|
| 46 |
xpath_stream xsp_stream; |
|---|
| 47 |
public : |
|---|
| 48 |
/// constructor |
|---|
| 49 |
this (xpath_stream xsp_in) ///< Pointer to an xpath_stream instance |
|---|
| 50 |
{ |
|---|
| 51 |
xsp_stream = xsp_in; |
|---|
| 52 |
} |
|---|
| 53 |
/// dispatcher of the xpath_stream::v_action |
|---|
| 54 |
final void v_action (xpath_construct xc_rule, uint u_sub, uint u_variable = 0, char[] cp_literal = "") |
|---|
| 55 |
{ |
|---|
| 56 |
xsp_stream.v_action (xc_rule, u_sub, u_variable, cp_literal); |
|---|
| 57 |
debug { |
|---|
| 58 |
auto File f = new File("Streamer", FileMode.Append); |
|---|
| 59 |
f.writefln("'%s' ('%d', '%d', '%s')", cp_disp_construct(xc_rule), u_sub, u_variable, replace(cp_literal,"\n","\\n")); |
|---|
| 60 |
} |
|---|
| 61 |
} |
|---|
| 62 |
/// dispatcher of the xpath_stream::i_get_action_counter |
|---|
| 63 |
final int i_get_action_counter () {return xsp_stream.i_get_action_counter;} |
|---|
| 64 |
} |
|---|
| 65 |
|
|---|
| 66 |
} |
|---|
| 67 |
|
|---|
| 68 |
public |
|---|
| 69 |
{ |
|---|
| 70 |
|
|---|
| 71 |
/** |
|---|
| 72 |
A specialized version of byte_stream for XPath |
|---|
| 73 |
*/ |
|---|
| 74 |
class xpath_stream : public byte_stream |
|---|
| 75 |
{ |
|---|
| 76 |
protected : |
|---|
| 77 |
/// List of tokens |
|---|
| 78 |
token_syntax_decoder tlp_list; |
|---|
| 79 |
|
|---|
| 80 |
public : |
|---|
| 81 |
/// constructor |
|---|
| 82 |
this (char[] cp_in) |
|---|
| 83 |
{ |
|---|
| 84 |
super (cp_in); |
|---|
| 85 |
tlp_list = new token_redef(this); |
|---|
| 86 |
} |
|---|
| 87 |
/// destructor |
|---|
| 88 |
~ this() |
|---|
| 89 |
{ |
|---|
| 90 |
delete tlp_list; |
|---|
| 91 |
} |
|---|
| 92 |
/// Decode the byte stream, and construct the lexical list |
|---|
| 93 |
void v_lexico_decode () |
|---|
| 94 |
{ |
|---|
| 95 |
enum _state {s_init, s_ncname, s_number, s_literal_1, s_literal_2, s_end}; |
|---|
| 96 |
_state state; |
|---|
| 97 |
lexico lex_new, lex_next; |
|---|
| 98 |
uint u_size; |
|---|
| 99 |
bool o_dot_in_number; |
|---|
| 100 |
|
|---|
| 101 |
u_size = 0; |
|---|
| 102 |
o_dot_in_number = false; |
|---|
| 103 |
state = _state.s_init; |
|---|
| 104 |
while (state != _state.s_end) |
|---|
| 105 |
{ |
|---|
| 106 |
lex_next = lex_get_class (b_top ()); |
|---|
| 107 |
debug writefln("Parsing: %s [%d] (%s)", cast(char)b_top(), cast(int)b_top(), cp_disp_class_lex(lex_next)); |
|---|
| 108 |
switch (state) |
|---|
| 109 |
{ |
|---|
| 110 |
case _state.s_init : |
|---|
| 111 |
switch (lex_next) |
|---|
| 112 |
{ |
|---|
| 113 |
case lexico.lex_bchar : |
|---|
| 114 |
case lexico.lex_under : |
|---|
| 115 |
// [XML:4] NCName ::= (Letter | '_') (NCNameChar)* |
|---|
| 116 |
|
|---|
| 117 |
u_size = 1; |
|---|
| 118 |
state = _state.s_ncname; |
|---|
| 119 |
b_pop (); |
|---|
| 120 |
break; |
|---|
| 121 |
case lexico.lex_null : |
|---|
| 122 |
state = _state.s_end; |
|---|
| 123 |
break; |
|---|
| 124 |
case lexico.lex_digit : |
|---|
| 125 |
u_size = 1; |
|---|
| 126 |
state = _state.s_number; |
|---|
| 127 |
o_dot_in_number = false; |
|---|
| 128 |
b_pop (); |
|---|
| 129 |
break; |
|---|
| 130 |
case lexico.lex_dot : |
|---|
| 131 |
if (lex_get_class (b_forward (1)) == lexico.lex_digit) |
|---|
| 132 |
{ |
|---|
| 133 |
// [30] Number ::= Digits ('.' Digits?)? | '.' Digits |
|---|
| 134 |
// [31] Digits ::= [0-9]+ |
|---|
| 135 |
u_size = 1; |
|---|
| 136 |
state = _state.s_number; |
|---|
| 137 |
o_dot_in_number = true; |
|---|
| 138 |
b_pop (); |
|---|
| 139 |
} |
|---|
| 140 |
else |
|---|
| 141 |
{ |
|---|
| 142 |
tlp_list.v_add_token (lex_next, bp_get_backward (1), 1); |
|---|
| 143 |
b_pop (); |
|---|
| 144 |
} |
|---|
| 145 |
break; |
|---|
| 146 |
|
|---|
| 147 |
case lexico.lex_1_quote : |
|---|
| 148 |
// [29] Literal ::= '"' [^"]* '"' | "'" [^']* "'" |
|---|
| 149 |
u_size = 0; |
|---|
| 150 |
b_pop (); |
|---|
| 151 |
state = _state.s_literal_1; |
|---|
| 152 |
break; |
|---|
| 153 |
|
|---|
| 154 |
case lexico.lex_2_quote : |
|---|
| 155 |
// [29] Literal ::= '"' [^"]* '"' | "'" [^']* "'" |
|---|
| 156 |
u_size = 0; |
|---|
| 157 |
b_pop (); |
|---|
| 158 |
state = _state.s_literal_2; |
|---|
| 159 |
break; |
|---|
| 160 |
|
|---|
| 161 |
default : |
|---|
| 162 |
tlp_list.v_add_token (lex_next, bp_get_backward (1), 1); |
|---|
| 163 |
b_pop (); |
|---|
| 164 |
break; |
|---|
| 165 |
} |
|---|
| 166 |
break; |
|---|
| 167 |
case _state.s_literal_1 : |
|---|
| 168 |
// [29] Literal ::= '"' [^"]* '"' | "'" [^']* "'" |
|---|
| 169 |
switch (lex_next) |
|---|
| 170 |
{ |
|---|
| 171 |
case lexico.lex_1_quote : |
|---|
| 172 |
tlp_list.v_add_token (lexico.lex_literal, bp_get_backward (u_size + 1), u_size); |
|---|
| 173 |
b_pop (); |
|---|
| 174 |
state = _state.s_init; |
|---|
| 175 |
break; |
|---|
| 176 |
default : |
|---|
| 177 |
u_size++; |
|---|
| 178 |
b_pop (); |
|---|
| 179 |
break; |
|---|
| 180 |
} |
|---|
| 181 |
break; |
|---|
| 182 |
case _state.s_literal_2 : |
|---|
| 183 |
// [29] Literal ::= '"' [^"]* '"' | "'" [^']* "'" |
|---|
| 184 |
switch (lex_next) |
|---|
| 185 |
{ |
|---|
| 186 |
case lexico.lex_2_quote : |
|---|
| 187 |
tlp_list.v_add_token (lexico.lex_literal, bp_get_backward (u_size + 1), u_size); |
|---|
| 188 |
b_pop (); |
|---|
| 189 |
state = _state.s_init; |
|---|
| 190 |
break; |
|---|
| 191 |
default : |
|---|
| 192 |
u_size++; |
|---|
| 193 |
b_pop (); |
|---|
| 194 |
break; |
|---|
| 195 |
} |
|---|
| 196 |
break; |
|---|
| 197 |
case _state.s_ncname : |
|---|
| 198 |
switch (lex_next) |
|---|
| 199 |
{ |
|---|
| 200 |
// [XML:5] NCNameChar ::= Letter | Digit | '.' | '-' | '_' | CombiningChar | Extender |
|---|
| 201 |
case lexico.lex_bchar : |
|---|
| 202 |
case lexico.lex_digit : |
|---|
| 203 |
case lexico.lex_dot : |
|---|
| 204 |
case lexico.lex_minus : |
|---|
| 205 |
case lexico.lex_under : |
|---|
| 206 |
case lexico.lex_extend : |
|---|
| 207 |
u_size++; |
|---|
| 208 |
b_pop (); |
|---|
| 209 |
debug writefln("Extending wit char: %s", cp_disp_class_lex(lex_next)); |
|---|
| 210 |
break; |
|---|
| 211 |
default : |
|---|
| 212 |
lex_new = lex_test_id (bp_get_backward (u_size + 1), u_size, lex_next); |
|---|
| 213 |
tlp_list.v_add_token (lex_new, bp_get_backward (u_size + 1), u_size); |
|---|
| 214 |
state = _state.s_init; |
|---|
| 215 |
break; |
|---|
| 216 |
} |
|---|
| 217 |
break; |
|---|
| 218 |
case _state.s_number : |
|---|
| 219 |
switch (lex_next) |
|---|
| 220 |
{ |
|---|
| 221 |
// [30] Number ::= Digits ('.' Digits?)? | '.' Digits |
|---|
| 222 |
// [31] Digits ::= [0-9]+ |
|---|
| 223 |
case lexico.lex_dot : |
|---|
| 224 |
if (o_dot_in_number) |
|---|
| 225 |
{ |
|---|
| 226 |
tlp_list.v_add_token (lexico.lex_number, bp_get_backward (u_size + 1), u_size); |
|---|
| 227 |
state = _state.s_init; |
|---|
| 228 |
} |
|---|
| 229 |
else |
|---|
| 230 |
{ |
|---|
| 231 |
o_dot_in_number = true; |
|---|
| 232 |
u_size++; |
|---|
| 233 |
b_pop (); |
|---|
| 234 |
} |
|---|
| 235 |
break; |
|---|
| 236 |
case lexico.lex_digit : |
|---|
| 237 |
u_size++; |
|---|
| 238 |
b_pop (); |
|---|
| 239 |
break; |
|---|
| 240 |
default : |
|---|
| 241 |
tlp_list.v_add_token (lexico.lex_number, bp_get_backward (u_size + 1), u_size); |
|---|
| 242 |
state = _state.s_init; |
|---|
| 243 |
break; |
|---|
| 244 |
} |
|---|
| 245 |
break; |
|---|
| 246 |
} |
|---|
| 247 |
if (lex_next == lexico.lex_null) |
|---|
| 248 |
state = _state.s_end; |
|---|
| 249 |
} |
|---|
| 250 |
} |
|---|
| 251 |
|
|---|
| 252 |
/// Evaluate a XPath expression \n |
|---|
| 253 |
/// Decodes the lexical and syntax contents. |
|---|
| 254 |
void v_evaluate () |
|---|
| 255 |
{ |
|---|
| 256 |
v_lexico_decode (); |
|---|
| 257 |
tlp_list.v_syntax_decode (); |
|---|
| 258 |
} |
|---|
| 259 |
|
|---|
| 260 |
|
|---|
| 261 |
|
|---|
| 262 |
/// Callback used by token_syntax_decoder::v_syntax_decode to notify of an action to be made. Pure virtual |
|---|
| 263 |
abstract void v_action (xpath_construct , uint , uint , char[] ) ; |
|---|
| 264 |
|
|---|
| 265 |
/// Callback used by token_syntax_decoder::v_syntax_decode to know the action counter position. Pure virtual |
|---|
| 266 |
/// \n This can be any kind of nomenclature, provided that the redefinition is coherent |
|---|
| 267 |
abstract int i_get_action_counter () ; |
|---|
| 268 |
} // class xpath_stream |
|---|
| 269 |
|
|---|
| 270 |
} |
|---|