fix the problem with parsing HTML comments (closes bug 1116708; based on patch 1168583)
[wxWidgets.git] / src / html / htmltag.cpp
1 /////////////////////////////////////////////////////////////////////////////
2 // Name: src/html/htmltag.cpp
3 // Purpose: wxHtmlTag class (represents single tag)
4 // Author: Vaclav Slavik
5 // RCS-ID: $Id$
6 // Copyright: (c) 1999 Vaclav Slavik
7 // Licence: wxWindows licence
8 /////////////////////////////////////////////////////////////////////////////
9
10 #include "wx/wxprec.h"
11
12 #ifdef __BORLANDC__
13 #pragma hdrstop
14 #endif
15
16 #if wxUSE_HTML
17
18 #include "wx/html/htmltag.h"
19
20 #ifndef WXPRECOMP
21 #include "wx/colour.h"
22 #endif
23
24 #include "wx/html/htmlpars.h"
25 #include <stdio.h> // for vsscanf
26 #include <stdarg.h>
27
28
29 //-----------------------------------------------------------------------------
30 // wxHtmlTagsCache
31 //-----------------------------------------------------------------------------
32
33 struct wxHtmlCacheItem
34 {
35 // this is "pos" value passed to wxHtmlTag's constructor.
36 // it is position of '<' character of the tag
37 int Key;
38
39 // end positions for the tag:
40 // end1 is '<' of ending tag,
41 // end2 is '>' or both are
42 // -1 if there is no ending tag for this one...
43 // or -2 if this is ending tag </...>
44 int End1, End2;
45
46 // name of this tag
47 wxChar *Name;
48 };
49
50
51 IMPLEMENT_CLASS(wxHtmlTagsCache,wxObject)
52
53 #define CACHE_INCREMENT 64
54
55 bool wxIsCDATAElement(const wxChar *tag)
56 {
57 return (wxStrcmp(tag, _T("SCRIPT")) == 0) ||
58 (wxStrcmp(tag, _T("STYLE")) == 0);
59 }
60
61 wxHtmlTagsCache::wxHtmlTagsCache(const wxString& source)
62 {
63 const wxChar *src = source.c_str();
64 int lng = source.length();
65 wxChar tagBuffer[256];
66
67 m_Cache = NULL;
68 m_CacheSize = 0;
69 m_CachePos = 0;
70
71 for ( int pos = 0; pos < lng; pos++ )
72 {
73 if (src[pos] == wxT('<')) // tag found:
74 {
75 // don't cache comment tags
76 wxString::const_iterator iter = source.begin() + pos;
77 if ( wxHtmlParser::SkipCommentTag(iter, source.end()) )
78 {
79 pos = iter - source.begin();
80 continue;
81 }
82
83 if (m_CacheSize % CACHE_INCREMENT == 0)
84 m_Cache = (wxHtmlCacheItem*) realloc(m_Cache, (m_CacheSize + CACHE_INCREMENT) * sizeof(wxHtmlCacheItem));
85 int tg = m_CacheSize++;
86 int stpos = pos++;
87 m_Cache[tg].Key = stpos;
88
89 int i;
90 for ( i = 0;
91 pos < lng && i < (int)WXSIZEOF(tagBuffer) - 1 &&
92 src[pos] != wxT('>') && !wxIsspace(src[pos]);
93 i++, pos++ )
94 {
95 tagBuffer[i] = (wxChar)wxToupper(src[pos]);
96 }
97 tagBuffer[i] = _T('\0');
98
99 m_Cache[tg].Name = new wxChar[i+1];
100 memcpy(m_Cache[tg].Name, tagBuffer, (i+1)*sizeof(wxChar));
101
102 while (pos < lng && src[pos] != wxT('>')) pos++;
103
104 if (src[stpos+1] == wxT('/')) // ending tag:
105 {
106 m_Cache[tg].End1 = m_Cache[tg].End2 = -2;
107 // find matching begin tag:
108 for (i = tg; i >= 0; i--)
109 if ((m_Cache[i].End1 == -1) && (wxStrcmp(m_Cache[i].Name, tagBuffer+1) == 0))
110 {
111 m_Cache[i].End1 = stpos;
112 m_Cache[i].End2 = pos + 1;
113 break;
114 }
115 }
116 else
117 {
118 m_Cache[tg].End1 = m_Cache[tg].End2 = -1;
119
120 if (wxIsCDATAElement(tagBuffer))
121 {
122 // store the orig pos in case we are missing the closing
123 // tag (see below)
124 wxInt32 old_pos = pos;
125 bool foundCloseTag = false;
126
127 // find next matching tag
128 int tag_len = wxStrlen(tagBuffer);
129 while (pos < lng)
130 {
131 // find the ending tag
132 while (pos + 1 < lng &&
133 (src[pos] != '<' || src[pos+1] != '/'))
134 ++pos;
135 if (src[pos] == '<')
136 ++pos;
137
138 // see if it matches
139 int match_pos = 0;
140 while (pos < lng && match_pos < tag_len && src[pos] != '>' && src[pos] != '<') {
141 // cast to wxChar needed to suppress warning in
142 // Unicode build
143 if ((wxChar)wxToupper(src[pos]) == tagBuffer[match_pos]) {
144 ++match_pos;
145 }
146 else if (src[pos] == wxT(' ') || src[pos] == wxT('\n') ||
147 src[pos] == wxT('\r') || src[pos] == wxT('\t')) {
148 // need to skip over these
149 }
150 else {
151 match_pos = 0;
152 }
153 ++pos;
154 }
155
156 // found a match
157 if (match_pos == tag_len)
158 {
159 pos = pos - tag_len - 3;
160 foundCloseTag = true;
161 break;
162 }
163 else // keep looking for the closing tag
164 {
165 ++pos;
166 }
167 }
168 if (!foundCloseTag)
169 {
170 // we didn't find closing tag; this means the markup
171 // is incorrect and the best thing we can do is to
172 // ignore the unclosed tag and continue parsing as if
173 // it didn't exist:
174 pos = old_pos;
175 }
176 }
177 }
178 }
179 }
180
181 // ok, we're done, now we'll free .Name members of cache - we don't need it anymore:
182 for (int i = 0; i < m_CacheSize; i++)
183 {
184 delete[] m_Cache[i].Name;
185 m_Cache[i].Name = NULL;
186 }
187 }
188
189 void wxHtmlTagsCache::QueryTag(int at, int* end1, int* end2)
190 {
191 if (m_Cache == NULL) return;
192 if (m_Cache[m_CachePos].Key != at)
193 {
194 int delta = (at < m_Cache[m_CachePos].Key) ? -1 : 1;
195 do
196 {
197 if ( m_CachePos < 0 || m_CachePos == m_CacheSize )
198 {
199 // something is very wrong with HTML, give up by returning an
200 // impossibly large value which is going to be ignored by the
201 // caller
202 *end1 =
203 *end2 = INT_MAX;
204 return;
205 }
206
207 m_CachePos += delta;
208 }
209 while (m_Cache[m_CachePos].Key != at);
210 }
211 *end1 = m_Cache[m_CachePos].End1;
212 *end2 = m_Cache[m_CachePos].End2;
213 }
214
215
216
217
218 //-----------------------------------------------------------------------------
219 // wxHtmlTag
220 //-----------------------------------------------------------------------------
221
222 IMPLEMENT_CLASS(wxHtmlTag,wxObject)
223
224 wxHtmlTag::wxHtmlTag(wxHtmlTag *parent,
225 const wxString& source, int pos, int end_pos,
226 wxHtmlTagsCache *cache,
227 wxHtmlEntitiesParser *entParser) : wxObject()
228 {
229 /* Setup DOM relations */
230
231 m_Next = NULL;
232 m_FirstChild = m_LastChild = NULL;
233 m_Parent = parent;
234 if (parent)
235 {
236 m_Prev = m_Parent->m_LastChild;
237 if (m_Prev == NULL)
238 m_Parent->m_FirstChild = this;
239 else
240 m_Prev->m_Next = this;
241 m_Parent->m_LastChild = this;
242 }
243 else
244 m_Prev = NULL;
245
246 /* Find parameters and their values: */
247
248 int i;
249 wxChar c;
250
251 // fill-in name, params and begin pos:
252 i = pos+1;
253
254 // find tag's name and convert it to uppercase:
255 while ((i < end_pos) &&
256 ((c = source[i++]) != wxT(' ') && c != wxT('\r') &&
257 c != wxT('\n') && c != wxT('\t') &&
258 c != wxT('>')))
259 {
260 if ((c >= wxT('a')) && (c <= wxT('z')))
261 c -= (wxT('a') - wxT('A'));
262 m_Name << c;
263 }
264
265 // if the tag has parameters, read them and "normalize" them,
266 // i.e. convert to uppercase, replace whitespaces by spaces and
267 // remove whitespaces around '=':
268 if (source[i-1] != wxT('>'))
269 {
270 #define IS_WHITE(c) (c == wxT(' ') || c == wxT('\r') || \
271 c == wxT('\n') || c == wxT('\t'))
272 wxString pname, pvalue;
273 wxChar quote;
274 enum
275 {
276 ST_BEFORE_NAME = 1,
277 ST_NAME,
278 ST_BEFORE_EQ,
279 ST_BEFORE_VALUE,
280 ST_VALUE
281 } state;
282
283 quote = 0;
284 state = ST_BEFORE_NAME;
285 while (i < end_pos)
286 {
287 c = source[i++];
288
289 if (c == wxT('>') && !(state == ST_VALUE && quote != 0))
290 {
291 if (state == ST_BEFORE_EQ || state == ST_NAME)
292 {
293 m_ParamNames.Add(pname);
294 m_ParamValues.Add(wxEmptyString);
295 }
296 else if (state == ST_VALUE && quote == 0)
297 {
298 m_ParamNames.Add(pname);
299 if (entParser)
300 m_ParamValues.Add(entParser->Parse(pvalue));
301 else
302 m_ParamValues.Add(pvalue);
303 }
304 break;
305 }
306 switch (state)
307 {
308 case ST_BEFORE_NAME:
309 if (!IS_WHITE(c))
310 {
311 pname = c;
312 state = ST_NAME;
313 }
314 break;
315 case ST_NAME:
316 if (IS_WHITE(c))
317 state = ST_BEFORE_EQ;
318 else if (c == wxT('='))
319 state = ST_BEFORE_VALUE;
320 else
321 pname << c;
322 break;
323 case ST_BEFORE_EQ:
324 if (c == wxT('='))
325 state = ST_BEFORE_VALUE;
326 else if (!IS_WHITE(c))
327 {
328 m_ParamNames.Add(pname);
329 m_ParamValues.Add(wxEmptyString);
330 pname = c;
331 state = ST_NAME;
332 }
333 break;
334 case ST_BEFORE_VALUE:
335 if (!IS_WHITE(c))
336 {
337 if (c == wxT('"') || c == wxT('\''))
338 quote = c, pvalue = wxEmptyString;
339 else
340 quote = 0, pvalue = c;
341 state = ST_VALUE;
342 }
343 break;
344 case ST_VALUE:
345 if ((quote != 0 && c == quote) ||
346 (quote == 0 && IS_WHITE(c)))
347 {
348 m_ParamNames.Add(pname);
349 if (quote == 0)
350 {
351 // VS: backward compatibility, no real reason,
352 // but wxHTML code relies on this... :(
353 pvalue.MakeUpper();
354 }
355 if (entParser)
356 m_ParamValues.Add(entParser->Parse(pvalue));
357 else
358 m_ParamValues.Add(pvalue);
359 state = ST_BEFORE_NAME;
360 }
361 else
362 pvalue << c;
363 break;
364 }
365 }
366
367 #undef IS_WHITE
368 }
369 m_Begin = i;
370
371 cache->QueryTag(pos, &m_End1, &m_End2);
372 if (m_End1 > end_pos) m_End1 = end_pos;
373 if (m_End2 > end_pos) m_End2 = end_pos;
374 }
375
376 wxHtmlTag::~wxHtmlTag()
377 {
378 wxHtmlTag *t1, *t2;
379 t1 = m_FirstChild;
380 while (t1)
381 {
382 t2 = t1->GetNextSibling();
383 delete t1;
384 t1 = t2;
385 }
386 }
387
388 bool wxHtmlTag::HasParam(const wxString& par) const
389 {
390 return (m_ParamNames.Index(par, false) != wxNOT_FOUND);
391 }
392
393 wxString wxHtmlTag::GetParam(const wxString& par, bool with_commas) const
394 {
395 int index = m_ParamNames.Index(par, false);
396 if (index == wxNOT_FOUND)
397 return wxEmptyString;
398 if (with_commas)
399 {
400 // VS: backward compatibility, seems to be never used by wxHTML...
401 wxString s;
402 s << wxT('"') << m_ParamValues[index] << wxT('"');
403 return s;
404 }
405 else
406 return m_ParamValues[index];
407 }
408
409 int wxHtmlTag::ScanParam(const wxString& par,
410 const wxChar *format,
411 void *param) const
412 {
413 wxString parval = GetParam(par);
414 return wxSscanf(parval, format, param);
415 }
416
417 bool wxHtmlTag::GetParamAsColour(const wxString& par, wxColour *clr) const
418 {
419 wxCHECK_MSG( clr, false, _T("invalid colour argument") );
420
421 wxString str = GetParam(par);
422
423 // handle colours defined in HTML 4.0 first:
424 if (str.length() > 1 && str[0] != _T('#'))
425 {
426 #define HTML_COLOUR(name, r, g, b) \
427 if (str.IsSameAs(wxT(name), false)) \
428 { clr->Set(r, g, b); return true; }
429 HTML_COLOUR("black", 0x00,0x00,0x00)
430 HTML_COLOUR("silver", 0xC0,0xC0,0xC0)
431 HTML_COLOUR("gray", 0x80,0x80,0x80)
432 HTML_COLOUR("white", 0xFF,0xFF,0xFF)
433 HTML_COLOUR("maroon", 0x80,0x00,0x00)
434 HTML_COLOUR("red", 0xFF,0x00,0x00)
435 HTML_COLOUR("purple", 0x80,0x00,0x80)
436 HTML_COLOUR("fuchsia", 0xFF,0x00,0xFF)
437 HTML_COLOUR("green", 0x00,0x80,0x00)
438 HTML_COLOUR("lime", 0x00,0xFF,0x00)
439 HTML_COLOUR("olive", 0x80,0x80,0x00)
440 HTML_COLOUR("yellow", 0xFF,0xFF,0x00)
441 HTML_COLOUR("navy", 0x00,0x00,0x80)
442 HTML_COLOUR("blue", 0x00,0x00,0xFF)
443 HTML_COLOUR("teal", 0x00,0x80,0x80)
444 HTML_COLOUR("aqua", 0x00,0xFF,0xFF)
445 #undef HTML_COLOUR
446 }
447
448 // then try to parse #rrggbb representations or set from other well
449 // known names (note that this doesn't strictly conform to HTML spec,
450 // but it doesn't do real harm -- but it *must* be done after the standard
451 // colors are handled above):
452 if (clr->Set(str))
453 return true;
454
455 return false;
456 }
457
458 bool wxHtmlTag::GetParamAsInt(const wxString& par, int *clr) const
459 {
460 if (!HasParam(par)) return false;
461 long i;
462 bool succ = GetParam(par).ToLong(&i);
463 *clr = (int)i;
464 return succ;
465 }
466
467 wxString wxHtmlTag::GetAllParams() const
468 {
469 // VS: this function is for backward compatibility only,
470 // never used by wxHTML
471 wxString s;
472 size_t cnt = m_ParamNames.GetCount();
473 for (size_t i = 0; i < cnt; i++)
474 {
475 s << m_ParamNames[i];
476 s << wxT('=');
477 if (m_ParamValues[i].Find(wxT('"')) != wxNOT_FOUND)
478 s << wxT('\'') << m_ParamValues[i] << wxT('\'');
479 else
480 s << wxT('"') << m_ParamValues[i] << wxT('"');
481 }
482 return s;
483 }
484
485 wxHtmlTag *wxHtmlTag::GetFirstSibling() const
486 {
487 if (m_Parent)
488 return m_Parent->m_FirstChild;
489 else
490 {
491 wxHtmlTag *cur = (wxHtmlTag*)this;
492 while (cur->m_Prev)
493 cur = cur->m_Prev;
494 return cur;
495 }
496 }
497
498 wxHtmlTag *wxHtmlTag::GetLastSibling() const
499 {
500 if (m_Parent)
501 return m_Parent->m_LastChild;
502 else
503 {
504 wxHtmlTag *cur = (wxHtmlTag*)this;
505 while (cur->m_Next)
506 cur = cur->m_Next;
507 return cur;
508 }
509 }
510
511 wxHtmlTag *wxHtmlTag::GetNextTag() const
512 {
513 if (m_FirstChild) return m_FirstChild;
514 if (m_Next) return m_Next;
515 wxHtmlTag *cur = m_Parent;
516 if (!cur) return NULL;
517 while (cur->m_Parent && !cur->m_Next)
518 cur = cur->m_Parent;
519 return cur->m_Next;
520 }
521
522 #endif