1 | /**************************************************************************
|
---|
2 | *
|
---|
3 | * TextEl.cpp -- Data structures for parsed documents
|
---|
4 | * Copyright (C) 1999 Rodger McNab
|
---|
5 | *
|
---|
6 | * This program is free software; you can redistribute it and/or modify
|
---|
7 | * it under the terms of the GNU General Public License as published by
|
---|
8 | * the Free Software Foundation; either version 2 of the License, or
|
---|
9 | * (at your option) any later version.
|
---|
10 | *
|
---|
11 | * This program is distributed in the hope that it will be useful,
|
---|
12 | * but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
13 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
14 | * GNU General Public License for more details.
|
---|
15 | *
|
---|
16 | * You should have received a copy of the GNU General Public License
|
---|
17 | * along with this program; if not, write to the Free Software
|
---|
18 | * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
19 | *
|
---|
20 | **************************************************************************/
|
---|
21 |
|
---|
22 | #include "TextEl.h"
|
---|
23 | #include "ctype.h"
|
---|
24 |
|
---|
25 | #define REC_TERM '\002'
|
---|
26 | #define PARA_TERM '\003'
|
---|
27 |
|
---|
28 |
|
---|
29 | void TextEl::Clear () {
|
---|
30 | elType = TextE;
|
---|
31 | tagName.erase (tagName.begin(), tagName.end());
|
---|
32 | text.erase (text.begin(), text.end());
|
---|
33 | }
|
---|
34 |
|
---|
35 |
|
---|
36 | static bool ReadTextTag (FileBuf &buf, TextEl &el) {
|
---|
37 | // we must have some sort of tag
|
---|
38 | unsigned char c;
|
---|
39 | if (!buf.Peek(c)) return false;
|
---|
40 |
|
---|
41 | // get the '<'
|
---|
42 | if (c != '<') return false;
|
---|
43 | el.text.push_back (c);
|
---|
44 | if (!buf.Next() || !buf.Peek (c)) return false;
|
---|
45 |
|
---|
46 | // see if we have a '/'
|
---|
47 | if (c == '/') {
|
---|
48 | // we have a closing tag
|
---|
49 | el.text.push_back (c);
|
---|
50 | if (!buf.Next() || !buf.Peek (c)) return false;
|
---|
51 | el.elType = CloseTagE;
|
---|
52 | } else {
|
---|
53 | // we have an opening tag
|
---|
54 | el.elType = OpenTagE;
|
---|
55 | }
|
---|
56 |
|
---|
57 | // get the tag name
|
---|
58 | while (isalnum(c)) {
|
---|
59 | el.text.push_back (c);
|
---|
60 | el.tagName.push_back (c);
|
---|
61 | if (!buf.Next() || !buf.Peek (c)) return false;
|
---|
62 | }
|
---|
63 |
|
---|
64 | // get everything up to the closing '>'
|
---|
65 | while (c != '>') {
|
---|
66 | el.text.push_back (c);
|
---|
67 | if (!buf.Next() || !buf.Peek (c)) return false;
|
---|
68 | }
|
---|
69 |
|
---|
70 | // get the closing '>'
|
---|
71 | el.text.push_back (c);
|
---|
72 | buf.Next();
|
---|
73 |
|
---|
74 | // if there is no tag name then make this a text element
|
---|
75 | if (el.tagName.empty()) el.elType = TextE;
|
---|
76 |
|
---|
77 | return true;
|
---|
78 | }
|
---|
79 |
|
---|
80 | static bool ReadTextTag (u_char ** buffer, TextEl &el) {
|
---|
81 | // we must have some sort of tag
|
---|
82 | u_char * buffer_pos = *buffer;
|
---|
83 | if (*buffer_pos == '\0' || *buffer_pos != '<') {
|
---|
84 | return false;
|
---|
85 | }
|
---|
86 | el.text.push_back('<');
|
---|
87 | buffer_pos ++;
|
---|
88 | if (*buffer_pos == '\0') {
|
---|
89 | *buffer = buffer_pos;
|
---|
90 | return false;
|
---|
91 | }
|
---|
92 | // see if we have a '/'
|
---|
93 | if (*buffer_pos == '/') {
|
---|
94 | el.text.push_back('/');
|
---|
95 | el.elType = CloseTagE;
|
---|
96 | buffer_pos++;
|
---|
97 | if (*buffer_pos == '\0') {
|
---|
98 | *buffer = buffer_pos;
|
---|
99 | return false;
|
---|
100 | }
|
---|
101 | } else {
|
---|
102 | el.elType = OpenTagE;
|
---|
103 | }
|
---|
104 | // get the tag name
|
---|
105 | while (isalnum(*buffer_pos)) {
|
---|
106 | el.text.push_back (*buffer_pos);
|
---|
107 | el.tagName.push_back (*buffer_pos);
|
---|
108 | buffer_pos++;
|
---|
109 |
|
---|
110 | if (*buffer_pos == '\0') {
|
---|
111 | *buffer = buffer_pos;
|
---|
112 | return false;
|
---|
113 | }
|
---|
114 |
|
---|
115 | }
|
---|
116 | cout << "tag name = "<<el.tagName<<endl;
|
---|
117 | // get everything up to closing '>'
|
---|
118 | while (*buffer_pos != '>') {
|
---|
119 | el.text.push_back(*buffer_pos);
|
---|
120 | buffer_pos++;
|
---|
121 | if (*buffer_pos == '\0') {
|
---|
122 | *buffer = buffer_pos;
|
---|
123 | return false;
|
---|
124 | }
|
---|
125 |
|
---|
126 | }
|
---|
127 | // get teh closing '>'
|
---|
128 | el.text.push_back('>');
|
---|
129 | buffer_pos++;
|
---|
130 | *buffer = buffer_pos;
|
---|
131 | cout << "at end of readtexttag, buffer is "<<*buffer<<endl;
|
---|
132 | // if there is no tag name then make this a text element
|
---|
133 | if (el.tagName.empty()) el.elType = TextE;
|
---|
134 |
|
---|
135 | return true;
|
---|
136 | }
|
---|
137 |
|
---|
138 | static void ToggleParaTag (TextEl &el, bool &compatInPara) {
|
---|
139 | SetCStr (el.tagName, "Paragraph", 9);
|
---|
140 | el.text.erase (el.text.begin(), el.text.end());
|
---|
141 | if (compatInPara) {
|
---|
142 | el.elType = CloseTagE;
|
---|
143 | el.text.push_back (PARA_TERM);
|
---|
144 | } else el.elType = OpenTagE;
|
---|
145 | compatInPara = !compatInPara;
|
---|
146 | }
|
---|
147 |
|
---|
148 | static void SetRecTag (TextEl &el, TextElType elType) {
|
---|
149 | el.elType = elType;
|
---|
150 | SetCStr (el.tagName, "Document", 8);
|
---|
151 | el.text.erase (el.text.begin(), el.text.end());
|
---|
152 | if (elType == CloseTagE)
|
---|
153 | el.text.push_back (REC_TERM);
|
---|
154 | }
|
---|
155 |
|
---|
156 |
|
---|
157 | bool ReadTextEl (FileBuf &buf, TextEl &el,
|
---|
158 | bool compatMode, bool &compatInPara) {
|
---|
159 | el.Clear();
|
---|
160 |
|
---|
161 | unsigned char c;
|
---|
162 | if (!buf.Peek (c)) return false;
|
---|
163 |
|
---|
164 | if (compatMode) {
|
---|
165 | if (c == PARA_TERM) {
|
---|
166 | ToggleParaTag (el, compatInPara);
|
---|
167 | if (!compatInPara) buf.Next();
|
---|
168 | return true;
|
---|
169 | }
|
---|
170 | if (c == REC_TERM) {
|
---|
171 | if (compatInPara) {
|
---|
172 | // need to close this paragraph
|
---|
173 | ToggleParaTag (el, compatInPara);
|
---|
174 | return true;
|
---|
175 | }
|
---|
176 | // can close this document
|
---|
177 | buf.Next();
|
---|
178 | SetRecTag (el, CloseTagE);
|
---|
179 | return true;
|
---|
180 | }
|
---|
181 | if (!compatInPara) {
|
---|
182 | // need to open a paragraph
|
---|
183 | ToggleParaTag (el, compatInPara);
|
---|
184 | return true;
|
---|
185 | }
|
---|
186 | }
|
---|
187 |
|
---|
188 | // see if we have some sort of tag
|
---|
189 | if (c == '<') return ReadTextTag (buf, el);
|
---|
190 |
|
---|
191 | // read in a text element
|
---|
192 | el.elType = TextE;
|
---|
193 | while (c != '<' && !(compatMode && (c == PARA_TERM || c == REC_TERM))) {
|
---|
194 | el.text.push_back (c);
|
---|
195 | if (!buf.Next() || !buf.Peek (c)) break;
|
---|
196 | }
|
---|
197 |
|
---|
198 | return true;
|
---|
199 | }
|
---|
200 | bool ReadTextEl (u_char **buffer, TextEl &el,
|
---|
201 | bool compatMode, bool &compatInPara) {
|
---|
202 | el.Clear();
|
---|
203 | u_char * buffer_pos = *buffer;
|
---|
204 | if (*buffer_pos == '\0') return false;
|
---|
205 | if (compatMode) {
|
---|
206 | if (*buffer_pos == PARA_TERM) {
|
---|
207 | ToggleParaTag (el, compatInPara);
|
---|
208 | if (!compatInPara) buffer_pos++;
|
---|
209 | *buffer = buffer_pos;
|
---|
210 | return true;
|
---|
211 | }
|
---|
212 | if (*buffer_pos == REC_TERM) {
|
---|
213 | if (compatInPara) {
|
---|
214 | // need to close this paragraph
|
---|
215 | ToggleParaTag (el, compatInPara);
|
---|
216 | *buffer = buffer_pos;
|
---|
217 | return true;
|
---|
218 | }
|
---|
219 | // can close this document
|
---|
220 | buffer_pos++;
|
---|
221 | SetRecTag (el, CloseTagE);
|
---|
222 | *buffer = buffer_pos;
|
---|
223 | return true;
|
---|
224 | }
|
---|
225 | if (!compatInPara) {
|
---|
226 | // need to open a paragraph
|
---|
227 | ToggleParaTag (el, compatInPara);
|
---|
228 | *buffer = buffer_pos;
|
---|
229 | return true;
|
---|
230 | }
|
---|
231 | }
|
---|
232 |
|
---|
233 | // see if we have some sort of tag
|
---|
234 | if (*buffer_pos == '<') return ReadTextTag (buffer, el);
|
---|
235 |
|
---|
236 | // read in a text element
|
---|
237 | el.elType = TextE;
|
---|
238 | while (*buffer_pos != '<' && !(compatMode && (*buffer_pos == PARA_TERM || *buffer_pos == REC_TERM))) {
|
---|
239 | el.text.push_back (*buffer_pos);
|
---|
240 | buffer_pos++;
|
---|
241 | if (*buffer_pos == '\0') break;
|
---|
242 | }
|
---|
243 | *buffer = buffer_pos;
|
---|
244 | cout << "text element: "<<el.text<<endl;
|
---|
245 | return true;
|
---|
246 | }
|
---|
247 |
|
---|
248 | static void AddTextEl (TextElArray &doc, unsigned long &docLen,
|
---|
249 | const TextEl &el) {
|
---|
250 | doc.push_back (el);
|
---|
251 | docLen += el.text.size();
|
---|
252 |
|
---|
253 | // cout << "elType: " << el.elType << "\n";
|
---|
254 | // cout << "tagName: " << el.tagName << "\n";
|
---|
255 | // cout << "text: \"" << el.text << "\"\n\n";
|
---|
256 | }
|
---|
257 |
|
---|
258 |
|
---|
259 | bool ReadDoc (FileBuf &buf, const UCArray &docTag,
|
---|
260 | TextElArray &doc, unsigned long &docLen,
|
---|
261 | bool compatMode) {
|
---|
262 | bool compatInPara = false;
|
---|
263 | bool foundDocEl = false;
|
---|
264 | TextEl el;
|
---|
265 |
|
---|
266 | doc.erase (doc.begin(), doc.end());
|
---|
267 | docLen = 0;
|
---|
268 |
|
---|
269 | if (compatMode) {
|
---|
270 | // add <Document><Paragraph>
|
---|
271 | SetRecTag (el, OpenTagE);
|
---|
272 | AddTextEl (doc, docLen, el);
|
---|
273 | ToggleParaTag (el, compatInPara);
|
---|
274 | AddTextEl (doc, docLen, el);
|
---|
275 |
|
---|
276 | } else {
|
---|
277 | // look for an opening docTag
|
---|
278 | do {
|
---|
279 | if (!ReadTextEl (buf, el, compatMode, compatInPara)) return false;
|
---|
280 | } while (el.elType != OpenTagE || el.tagName != docTag);
|
---|
281 |
|
---|
282 | AddTextEl (doc, docLen, el);
|
---|
283 | }
|
---|
284 |
|
---|
285 |
|
---|
286 | // get all elements until the closing docTag
|
---|
287 | while (ReadTextEl (buf, el, compatMode, compatInPara)) {
|
---|
288 | foundDocEl = true;
|
---|
289 | AddTextEl (doc, docLen, el);
|
---|
290 | if (el.elType == CloseTagE && el.tagName == docTag)
|
---|
291 | return true;
|
---|
292 | }
|
---|
293 |
|
---|
294 |
|
---|
295 | if (compatMode) {
|
---|
296 | if (!foundDocEl) { // end of text
|
---|
297 | doc.erase (doc.begin(), doc.end());
|
---|
298 | return false;
|
---|
299 | }
|
---|
300 |
|
---|
301 | // if we got here then the eof was encountered before
|
---|
302 | // the closing document tag
|
---|
303 | if (compatInPara) {
|
---|
304 | // need to close this paragraph
|
---|
305 | ToggleParaTag (el, compatInPara);
|
---|
306 | AddTextEl (doc, docLen, el);
|
---|
307 | }
|
---|
308 | // close this document
|
---|
309 | SetRecTag (el, CloseTagE);
|
---|
310 | AddTextEl (doc, docLen, el);
|
---|
311 | }
|
---|
312 |
|
---|
313 | return true;
|
---|
314 | }
|
---|
315 |
|
---|
316 | // copy of readdoc to get TextElArray from u_char buffer
|
---|
317 | bool ReadDoc(u_char ** buffer, const UCArray &docTag,
|
---|
318 | TextElArray &doc, unsigned long &docLen,
|
---|
319 | bool compatMode) {
|
---|
320 | bool compatInPara = false;
|
---|
321 | bool foundDocEl = false;
|
---|
322 | TextEl el;
|
---|
323 |
|
---|
324 | doc.erase (doc.begin(), doc.end());
|
---|
325 | docLen = 0;
|
---|
326 |
|
---|
327 | if (compatMode) {
|
---|
328 | // add <Document><Paragraph>
|
---|
329 | SetRecTag (el, OpenTagE);
|
---|
330 | AddTextEl (doc, docLen, el);
|
---|
331 | ToggleParaTag (el, compatInPara);
|
---|
332 | AddTextEl (doc, docLen, el);
|
---|
333 |
|
---|
334 | } else {
|
---|
335 | // look for an opening docTag
|
---|
336 | do {
|
---|
337 | if (!ReadTextEl (buffer, el, compatMode, compatInPara)) return false;
|
---|
338 | } while (el.elType != OpenTagE || el.tagName != docTag);
|
---|
339 |
|
---|
340 | AddTextEl (doc, docLen, el);
|
---|
341 | }
|
---|
342 |
|
---|
343 |
|
---|
344 | // get all elements until the closing docTag
|
---|
345 | while (ReadTextEl (buffer, el, compatMode, compatInPara)) {
|
---|
346 | foundDocEl = true;
|
---|
347 | AddTextEl (doc, docLen, el);
|
---|
348 | if (el.elType == CloseTagE && el.tagName == docTag)
|
---|
349 | return true;
|
---|
350 | }
|
---|
351 |
|
---|
352 |
|
---|
353 | if (compatMode) {
|
---|
354 | if (!foundDocEl) { // end of text
|
---|
355 | doc.erase (doc.begin(), doc.end());
|
---|
356 | return false;
|
---|
357 | }
|
---|
358 |
|
---|
359 | // if we got here then the eof was encountered before
|
---|
360 | // the closing document tag
|
---|
361 | if (compatInPara) {
|
---|
362 | // need to close this paragraph
|
---|
363 | ToggleParaTag (el, compatInPara);
|
---|
364 | AddTextEl (doc, docLen, el);
|
---|
365 | }
|
---|
366 | // close this document
|
---|
367 | SetRecTag (el, CloseTagE);
|
---|
368 | AddTextEl (doc, docLen, el);
|
---|
369 | }
|
---|
370 |
|
---|
371 | return true;
|
---|
372 | }
|
---|
373 |
|
---|
374 |
|
---|