netwerk/streamconv/converters/mozTXTToHTMLConv.cpp@deefc01c0e14
netwerk/streamconv/converters/mozTXTToHTMLConv.cpp
Thu, 15 Jan 2015 21:03:48 +0100
- author
- Michael Schloh von Bennewitz <michael@schloh.com>
- date
- Thu, 15 Jan 2015 21:03:48 +0100
- branch
- TOR_BUG_9701
- changeset 11
- deefc01c0e14
- permissions
- -rw-r--r--
Integrate friendly tips from Tor colleagues to make (or not) 4.5 alpha 3;
This includes removal of overloaded (but unused) methods, and addition of
a overlooked call to DataStruct::SetData(nsISupports, uint32_t, bool.)
1 /* -*- Mode: C; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
2 /* This Source Code Form is subject to the terms of the Mozilla Public
3 * License, v. 2.0. If a copy of the MPL was not distributed with this
4 * file, You can obtain one at http://mozilla.org/MPL/2.0/. */
6 #include "mozTXTToHTMLConv.h"
7 #include "nsNetUtil.h"
8 #include "nsUnicharUtils.h"
9 #include "nsCRT.h"
10 #include "nsIExternalProtocolHandler.h"
11 #include "nsIIOService.h"
13 #include <algorithm>
15 #ifdef DEBUG_BenB_Perf
16 #include "prtime.h"
17 #include "prinrval.h"
18 #endif
20 const double growthRate = 1.2;
22 // Bug 183111, editor now replaces multiple spaces with leading
23 // 0xA0's and a single ending space, so need to treat 0xA0's as spaces.
24 // 0xA0 is the Latin1/Unicode character for "non-breaking space (nbsp)"
25 // Also recognize the Japanese ideographic space 0x3000 as a space.
26 static inline bool IsSpace(const char16_t aChar)
27 {
28 return (nsCRT::IsAsciiSpace(aChar) || aChar == 0xA0 || aChar == 0x3000);
29 }
31 // Escape Char will take ch, escape it and append the result to
32 // aStringToAppendTo
33 void
34 mozTXTToHTMLConv::EscapeChar(const char16_t ch, nsString& aStringToAppendTo,
35 bool inAttribute)
36 {
37 switch (ch)
38 {
39 case '<':
40 aStringToAppendTo.AppendLiteral("<");
41 break;
42 case '>':
43 aStringToAppendTo.AppendLiteral(">");
44 break;
45 case '&':
46 aStringToAppendTo.AppendLiteral("&");
47 break;
48 case '"':
49 if (inAttribute)
50 {
51 aStringToAppendTo.AppendLiteral(""");
52 break;
53 }
54 // else fall through
55 default:
56 aStringToAppendTo += ch;
57 }
59 return;
60 }
62 // EscapeStr takes the passed in string and
63 // escapes it IN PLACE.
64 void
65 mozTXTToHTMLConv::EscapeStr(nsString& aInString, bool inAttribute)
66 {
67 // the replace substring routines
68 // don't seem to work if you have a character
69 // in the in string that is also in the replacement
70 // string! =(
71 //aInString.ReplaceSubstring("&", "&");
72 //aInString.ReplaceSubstring("<", "<");
73 //aInString.ReplaceSubstring(">", ">");
74 for (uint32_t i = 0; i < aInString.Length();)
75 {
76 switch (aInString[i])
77 {
78 case '<':
79 aInString.Cut(i, 1);
80 aInString.Insert(NS_LITERAL_STRING("<"), i);
81 i += 4; // skip past the integers we just added
82 break;
83 case '>':
84 aInString.Cut(i, 1);
85 aInString.Insert(NS_LITERAL_STRING(">"), i);
86 i += 4; // skip past the integers we just added
87 break;
88 case '&':
89 aInString.Cut(i, 1);
90 aInString.Insert(NS_LITERAL_STRING("&"), i);
91 i += 5; // skip past the integers we just added
92 break;
93 case '"':
94 if (inAttribute)
95 {
96 aInString.Cut(i, 1);
97 aInString.Insert(NS_LITERAL_STRING("""), i);
98 i += 6;
99 break;
100 }
101 // else fall through
102 default:
103 i++;
104 }
105 }
106 }
108 void
109 mozTXTToHTMLConv::UnescapeStr(const char16_t * aInString, int32_t aStartPos, int32_t aLength, nsString& aOutString)
110 {
111 const char16_t * subString = nullptr;
112 for (uint32_t i = aStartPos; int32_t(i) - aStartPos < aLength;)
113 {
114 int32_t remainingChars = i - aStartPos;
115 if (aInString[i] == '&')
116 {
117 subString = &aInString[i];
118 if (!nsCRT::strncmp(subString, MOZ_UTF16("<"), std::min(4, aLength - remainingChars)))
119 {
120 aOutString.Append(char16_t('<'));
121 i += 4;
122 }
123 else if (!nsCRT::strncmp(subString, MOZ_UTF16(">"), std::min(4, aLength - remainingChars)))
124 {
125 aOutString.Append(char16_t('>'));
126 i += 4;
127 }
128 else if (!nsCRT::strncmp(subString, MOZ_UTF16("&"), std::min(5, aLength - remainingChars)))
129 {
130 aOutString.Append(char16_t('&'));
131 i += 5;
132 }
133 else if (!nsCRT::strncmp(subString, MOZ_UTF16("""), std::min(6, aLength - remainingChars)))
134 {
135 aOutString.Append(char16_t('"'));
136 i += 6;
137 }
138 else
139 {
140 aOutString += aInString[i];
141 i++;
142 }
143 }
144 else
145 {
146 aOutString += aInString[i];
147 i++;
148 }
149 }
150 }
152 void
153 mozTXTToHTMLConv::CompleteAbbreviatedURL(const char16_t * aInString, int32_t aInLength,
154 const uint32_t pos, nsString& aOutString)
155 {
156 NS_ASSERTION(int32_t(pos) < aInLength, "bad args to CompleteAbbreviatedURL, see bug #190851");
157 if (int32_t(pos) >= aInLength)
158 return;
160 if (aInString[pos] == '@')
161 {
162 // only pre-pend a mailto url if the string contains a .domain in it..
163 //i.e. we want to linkify johndoe@foo.com but not "let's meet @8pm"
164 nsDependentString inString(aInString, aInLength);
165 if (inString.FindChar('.', pos) != kNotFound) // if we have a '.' after the @ sign....
166 {
167 aOutString.AssignLiteral("mailto:");
168 aOutString += aInString;
169 }
170 }
171 else if (aInString[pos] == '.')
172 {
173 if (ItMatchesDelimited(aInString, aInLength,
174 MOZ_UTF16("www."), 4, LT_IGNORE, LT_IGNORE))
175 {
176 aOutString.AssignLiteral("http://");
177 aOutString += aInString;
178 }
179 else if (ItMatchesDelimited(aInString,aInLength, MOZ_UTF16("ftp."), 4, LT_IGNORE, LT_IGNORE))
180 {
181 aOutString.AssignLiteral("ftp://");
182 aOutString += aInString;
183 }
184 }
185 }
187 bool
188 mozTXTToHTMLConv::FindURLStart(const char16_t * aInString, int32_t aInLength,
189 const uint32_t pos, const modetype check,
190 uint32_t& start)
191 {
192 switch(check)
193 { // no breaks, because end of blocks is never reached
194 case RFC1738:
195 {
196 if (!nsCRT::strncmp(&aInString[std::max(int32_t(pos - 4), 0)], MOZ_UTF16("<URL:"), 5))
197 {
198 start = pos + 1;
199 return true;
200 }
201 else
202 return false;
203 }
204 case RFC2396E:
205 {
206 nsString temp(aInString, aInLength);
207 int32_t i = pos <= 0 ? kNotFound : temp.RFindCharInSet(MOZ_UTF16("<>\""), pos - 1);
208 if (i != kNotFound && (temp[uint32_t(i)] == '<' ||
209 temp[uint32_t(i)] == '"'))
210 {
211 start = uint32_t(++i);
212 return start < pos;
213 }
214 else
215 return false;
216 }
217 case freetext:
218 {
219 int32_t i = pos - 1;
220 for (; i >= 0 && (
221 nsCRT::IsAsciiAlpha(aInString[uint32_t(i)]) ||
222 nsCRT::IsAsciiDigit(aInString[uint32_t(i)]) ||
223 aInString[uint32_t(i)] == '+' ||
224 aInString[uint32_t(i)] == '-' ||
225 aInString[uint32_t(i)] == '.'
226 ); i--)
227 ;
228 if (++i >= 0 && uint32_t(i) < pos && nsCRT::IsAsciiAlpha(aInString[uint32_t(i)]))
229 {
230 start = uint32_t(i);
231 return true;
232 }
233 else
234 return false;
235 }
236 case abbreviated:
237 {
238 int32_t i = pos - 1;
239 // This disallows non-ascii-characters for email.
240 // Currently correct, but revisit later after standards changed.
241 bool isEmail = aInString[pos] == (char16_t)'@';
242 // These chars mark the start of the URL
243 for (; i >= 0
244 && aInString[uint32_t(i)] != '>' && aInString[uint32_t(i)] != '<'
245 && aInString[uint32_t(i)] != '"' && aInString[uint32_t(i)] != '\''
246 && aInString[uint32_t(i)] != '`' && aInString[uint32_t(i)] != ','
247 && aInString[uint32_t(i)] != '{' && aInString[uint32_t(i)] != '['
248 && aInString[uint32_t(i)] != '(' && aInString[uint32_t(i)] != '|'
249 && aInString[uint32_t(i)] != '\\'
250 && !IsSpace(aInString[uint32_t(i)])
251 && (!isEmail || nsCRT::IsAscii(aInString[uint32_t(i)]))
252 ; i--)
253 ;
254 if
255 (
256 ++i >= 0 && uint32_t(i) < pos
257 &&
258 (
259 nsCRT::IsAsciiAlpha(aInString[uint32_t(i)]) ||
260 nsCRT::IsAsciiDigit(aInString[uint32_t(i)])
261 )
262 )
263 {
264 start = uint32_t(i);
265 return true;
266 }
267 else
268 return false;
269 }
270 default:
271 return false;
272 } //switch
273 }
275 bool
276 mozTXTToHTMLConv::FindURLEnd(const char16_t * aInString, int32_t aInStringLength, const uint32_t pos,
277 const modetype check, const uint32_t start, uint32_t& end)
278 {
279 switch(check)
280 { // no breaks, because end of blocks is never reached
281 case RFC1738:
282 case RFC2396E:
283 {
284 nsString temp(aInString, aInStringLength);
286 int32_t i = temp.FindCharInSet(MOZ_UTF16("<>\""), pos + 1);
287 if (i != kNotFound && temp[uint32_t(i--)] ==
288 (check == RFC1738 || temp[start - 1] == '<' ? '>' : '"'))
289 {
290 end = uint32_t(i);
291 return end > pos;
292 }
293 return false;
294 }
295 case freetext:
296 case abbreviated:
297 {
298 uint32_t i = pos + 1;
299 bool isEmail = aInString[pos] == (char16_t)'@';
300 bool seenOpeningParenthesis = false; // there is a '(' earlier in the URL
301 bool seenOpeningSquareBracket = false; // there is a '[' earlier in the URL
302 for (; int32_t(i) < aInStringLength; i++)
303 {
304 // These chars mark the end of the URL
305 if (aInString[i] == '>' || aInString[i] == '<' ||
306 aInString[i] == '"' || aInString[i] == '`' ||
307 aInString[i] == '}' || aInString[i] == '{' ||
308 aInString[i] == '|' ||
309 (aInString[i] == ')' && !seenOpeningParenthesis) ||
310 (aInString[i] == ']' && !seenOpeningSquareBracket) ||
311 // Allow IPv6 adresses like http://[1080::8:800:200C:417A]/foo.
312 (aInString[i] == '[' && i > 2 &&
313 (aInString[i - 1] != '/' || aInString[i - 2] != '/')) ||
314 IsSpace(aInString[i]))
315 break;
316 // Disallow non-ascii-characters for email.
317 // Currently correct, but revisit later after standards changed.
318 if (isEmail && (
319 aInString[i] == '(' || aInString[i] == '\'' ||
320 !nsCRT::IsAscii(aInString[i])))
321 break;
322 if (aInString[i] == '(')
323 seenOpeningParenthesis = true;
324 if (aInString[i] == '[')
325 seenOpeningSquareBracket = true;
326 }
327 // These chars are allowed in the middle of the URL, but not at end.
328 // Technically they are, but are used in normal text after the URL.
329 while (--i > pos && (
330 aInString[i] == '.' || aInString[i] == ',' || aInString[i] == ';' ||
331 aInString[i] == '!' || aInString[i] == '?' || aInString[i] == '-' ||
332 aInString[i] == ':' || aInString[i] == '\''
333 ))
334 ;
335 if (i > pos)
336 {
337 end = i;
338 return true;
339 }
340 return false;
341 }
342 default:
343 return false;
344 } //switch
345 }
347 void
348 mozTXTToHTMLConv::CalculateURLBoundaries(const char16_t * aInString, int32_t aInStringLength,
349 const uint32_t pos, const uint32_t whathasbeendone,
350 const modetype check, const uint32_t start, const uint32_t end,
351 nsString& txtURL, nsString& desc,
352 int32_t& replaceBefore, int32_t& replaceAfter)
353 {
354 uint32_t descstart = start;
355 switch(check)
356 {
357 case RFC1738:
358 {
359 descstart = start - 5;
360 desc.Append(&aInString[descstart], end - descstart + 2); // include "<URL:" and ">"
361 replaceAfter = end - pos + 1;
362 } break;
363 case RFC2396E:
364 {
365 descstart = start - 1;
366 desc.Append(&aInString[descstart], end - descstart + 2); // include brackets
367 replaceAfter = end - pos + 1;
368 } break;
369 case freetext:
370 case abbreviated:
371 {
372 descstart = start;
373 desc.Append(&aInString[descstart], end - start + 1); // don't include brackets
374 replaceAfter = end - pos;
375 } break;
376 default: break;
377 } //switch
379 EscapeStr(desc, false);
381 txtURL.Append(&aInString[start], end - start + 1);
382 txtURL.StripWhitespace();
384 // FIX ME
385 nsAutoString temp2;
386 ScanTXT(&aInString[descstart], pos - descstart, ~kURLs /*prevents loop*/ & whathasbeendone, temp2);
387 replaceBefore = temp2.Length();
388 return;
389 }
391 bool mozTXTToHTMLConv::ShouldLinkify(const nsCString& aURL)
392 {
393 if (!mIOService)
394 return false;
396 nsAutoCString scheme;
397 nsresult rv = mIOService->ExtractScheme(aURL, scheme);
398 if(NS_FAILED(rv))
399 return false;
401 // Get the handler for this scheme.
402 nsCOMPtr<nsIProtocolHandler> handler;
403 rv = mIOService->GetProtocolHandler(scheme.get(), getter_AddRefs(handler));
404 if(NS_FAILED(rv))
405 return false;
407 // Is it an external protocol handler? If not, linkify it.
408 nsCOMPtr<nsIExternalProtocolHandler> externalHandler = do_QueryInterface(handler);
409 if (!externalHandler)
410 return true; // handler is built-in, linkify it!
412 // If external app exists for the scheme then linkify it.
413 bool exists;
414 rv = externalHandler->ExternalAppExistsForScheme(scheme, &exists);
415 return(NS_SUCCEEDED(rv) && exists);
416 }
418 bool
419 mozTXTToHTMLConv::CheckURLAndCreateHTML(
420 const nsString& txtURL, const nsString& desc, const modetype mode,
421 nsString& outputHTML)
422 {
423 // Create *uri from txtURL
424 nsCOMPtr<nsIURI> uri;
425 nsresult rv;
426 // Lazily initialize mIOService
427 if (!mIOService)
428 {
429 mIOService = do_GetIOService();
431 if (!mIOService)
432 return false;
433 }
435 // See if the url should be linkified.
436 NS_ConvertUTF16toUTF8 utf8URL(txtURL);
437 if (!ShouldLinkify(utf8URL))
438 return false;
440 // it would be faster if we could just check to see if there is a protocol
441 // handler for the url and return instead of actually trying to create a url...
442 rv = mIOService->NewURI(utf8URL, nullptr, nullptr, getter_AddRefs(uri));
444 // Real work
445 if (NS_SUCCEEDED(rv) && uri)
446 {
447 outputHTML.AssignLiteral("<a class=\"moz-txt-link-");
448 switch(mode)
449 {
450 case RFC1738:
451 outputHTML.AppendLiteral("rfc1738");
452 break;
453 case RFC2396E:
454 outputHTML.AppendLiteral("rfc2396E");
455 break;
456 case freetext:
457 outputHTML.AppendLiteral("freetext");
458 break;
459 case abbreviated:
460 outputHTML.AppendLiteral("abbreviated");
461 break;
462 default: break;
463 }
464 nsAutoString escapedURL(txtURL);
465 EscapeStr(escapedURL, true);
467 outputHTML.AppendLiteral("\" href=\"");
468 outputHTML += escapedURL;
469 outputHTML.AppendLiteral("\">");
470 outputHTML += desc;
471 outputHTML.AppendLiteral("</a>");
472 return true;
473 }
474 else
475 return false;
476 }
478 NS_IMETHODIMP mozTXTToHTMLConv::FindURLInPlaintext(const char16_t * aInString, int32_t aInLength, int32_t aPos, int32_t * aStartPos, int32_t * aEndPos)
479 {
480 // call FindURL on the passed in string
481 nsAutoString outputHTML; // we'll ignore the generated output HTML
483 *aStartPos = -1;
484 *aEndPos = -1;
486 FindURL(aInString, aInLength, aPos, kURLs, outputHTML, *aStartPos, *aEndPos);
488 return NS_OK;
489 }
491 bool
492 mozTXTToHTMLConv::FindURL(const char16_t * aInString, int32_t aInLength, const uint32_t pos,
493 const uint32_t whathasbeendone,
494 nsString& outputHTML, int32_t& replaceBefore, int32_t& replaceAfter)
495 {
496 enum statetype {unchecked, invalid, startok, endok, success};
497 static const modetype ranking[] = {RFC1738, RFC2396E, freetext, abbreviated};
499 statetype state[mozTXTToHTMLConv_lastMode + 1]; // 0(=unknown)..lastMode
500 /* I don't like this abuse of enums as index for the array,
501 but I don't know a better method */
503 // Define, which modes to check
504 /* all modes but abbreviated are checked for text[pos] == ':',
505 only abbreviated for '.', RFC2396E and abbreviated for '@' */
506 for (modetype iState = unknown; iState <= mozTXTToHTMLConv_lastMode;
507 iState = modetype(iState + 1))
508 state[iState] = aInString[pos] == ':' ? unchecked : invalid;
509 switch (aInString[pos])
510 {
511 case '@':
512 state[RFC2396E] = unchecked;
513 // no break here
514 case '.':
515 state[abbreviated] = unchecked;
516 break;
517 case ':':
518 state[abbreviated] = invalid;
519 break;
520 default:
521 break;
522 }
524 // Test, first successful mode wins, sequence defined by |ranking|
525 int32_t iCheck = 0; // the currently tested modetype
526 modetype check = ranking[iCheck];
527 for (; iCheck < mozTXTToHTMLConv_numberOfModes && state[check] != success;
528 iCheck++)
529 /* check state from last run.
530 If this is the first, check this one, which isn't = success yet */
531 {
532 check = ranking[iCheck];
534 uint32_t start, end;
536 if (state[check] == unchecked)
537 if (FindURLStart(aInString, aInLength, pos, check, start))
538 state[check] = startok;
540 if (state[check] == startok)
541 if (FindURLEnd(aInString, aInLength, pos, check, start, end))
542 state[check] = endok;
544 if (state[check] == endok)
545 {
546 nsAutoString txtURL, desc;
547 int32_t resultReplaceBefore, resultReplaceAfter;
549 CalculateURLBoundaries(aInString, aInLength, pos, whathasbeendone, check, start, end,
550 txtURL, desc,
551 resultReplaceBefore, resultReplaceAfter);
553 if (aInString[pos] != ':')
554 {
555 nsAutoString temp = txtURL;
556 txtURL.SetLength(0);
557 CompleteAbbreviatedURL(temp.get(),temp.Length(), pos - start, txtURL);
558 }
560 if (!txtURL.IsEmpty() && CheckURLAndCreateHTML(txtURL, desc, check,
561 outputHTML))
562 {
563 replaceBefore = resultReplaceBefore;
564 replaceAfter = resultReplaceAfter;
565 state[check] = success;
566 }
567 } // if
568 } // for
569 return state[check] == success;
570 }
572 bool
573 mozTXTToHTMLConv::ItMatchesDelimited(const char16_t * aInString,
574 int32_t aInLength, const char16_t* rep, int32_t aRepLen,
575 LIMTYPE before, LIMTYPE after)
576 {
578 // this little method gets called a LOT. I found we were spending a
579 // lot of time just calculating the length of the variable "rep"
580 // over and over again every time we called it. So we're now passing
581 // an integer in here.
582 int32_t textLen = aInLength;
584 if
585 (
586 ((before == LT_IGNORE && (after == LT_IGNORE || after == LT_DELIMITER))
587 && textLen < aRepLen) ||
588 ((before != LT_IGNORE || (after != LT_IGNORE && after != LT_DELIMITER))
589 && textLen < aRepLen + 1) ||
590 (before != LT_IGNORE && after != LT_IGNORE && after != LT_DELIMITER
591 && textLen < aRepLen + 2)
592 )
593 return false;
595 char16_t text0 = aInString[0];
596 char16_t textAfterPos = aInString[aRepLen + (before == LT_IGNORE ? 0 : 1)];
598 if
599 (
600 (before == LT_ALPHA
601 && !nsCRT::IsAsciiAlpha(text0)) ||
602 (before == LT_DIGIT
603 && !nsCRT::IsAsciiDigit(text0)) ||
604 (before == LT_DELIMITER
605 &&
606 (
607 nsCRT::IsAsciiAlpha(text0) ||
608 nsCRT::IsAsciiDigit(text0) ||
609 text0 == *rep
610 )) ||
611 (after == LT_ALPHA
612 && !nsCRT::IsAsciiAlpha(textAfterPos)) ||
613 (after == LT_DIGIT
614 && !nsCRT::IsAsciiDigit(textAfterPos)) ||
615 (after == LT_DELIMITER
616 &&
617 (
618 nsCRT::IsAsciiAlpha(textAfterPos) ||
619 nsCRT::IsAsciiDigit(textAfterPos) ||
620 textAfterPos == *rep
621 )) ||
622 !Substring(Substring(aInString, aInString+aInLength),
623 (before == LT_IGNORE ? 0 : 1),
624 aRepLen).Equals(Substring(rep, rep+aRepLen),
625 nsCaseInsensitiveStringComparator())
626 )
627 return false;
629 return true;
630 }
632 uint32_t
633 mozTXTToHTMLConv::NumberOfMatches(const char16_t * aInString, int32_t aInStringLength,
634 const char16_t* rep, int32_t aRepLen, LIMTYPE before, LIMTYPE after)
635 {
636 uint32_t result = 0;
638 for (int32_t i = 0; i < aInStringLength; i++)
639 {
640 const char16_t * indexIntoString = &aInString[i];
641 if (ItMatchesDelimited(indexIntoString, aInStringLength - i, rep, aRepLen, before, after))
642 result++;
643 }
644 return result;
645 }
648 // NOTE: the converted html for the phrase is appended to aOutString
649 // tagHTML and attributeHTML are plain ASCII (literal strings, in fact)
650 bool
651 mozTXTToHTMLConv::StructPhraseHit(const char16_t * aInString, int32_t aInStringLength, bool col0,
652 const char16_t* tagTXT, int32_t aTagTXTLen,
653 const char* tagHTML, const char* attributeHTML,
654 nsString& aOutString, uint32_t& openTags)
655 {
656 /* We're searching for the following pattern:
657 LT_DELIMITER - "*" - ALPHA -
658 [ some text (maybe more "*"-pairs) - ALPHA ] "*" - LT_DELIMITER.
659 <strong> is only inserted, if existence of a pair could be verified
660 We use the first opening/closing tag, if we can choose */
662 const char16_t * newOffset = aInString;
663 int32_t newLength = aInStringLength;
664 if (!col0) // skip the first element?
665 {
666 newOffset = &aInString[1];
667 newLength = aInStringLength - 1;
668 }
670 // opening tag
671 if
672 (
673 ItMatchesDelimited(aInString, aInStringLength, tagTXT, aTagTXTLen,
674 (col0 ? LT_IGNORE : LT_DELIMITER), LT_ALPHA) // is opening tag
675 && NumberOfMatches(newOffset, newLength, tagTXT, aTagTXTLen,
676 LT_ALPHA, LT_DELIMITER) // remaining closing tags
677 > openTags
678 )
679 {
680 openTags++;
681 aOutString.AppendLiteral("<");
682 aOutString.AppendASCII(tagHTML);
683 aOutString.Append(char16_t(' '));
684 aOutString.AppendASCII(attributeHTML);
685 aOutString.AppendLiteral("><span class=\"moz-txt-tag\">");
686 aOutString.Append(tagTXT);
687 aOutString.AppendLiteral("</span>");
688 return true;
689 }
691 // closing tag
692 else if (openTags > 0
693 && ItMatchesDelimited(aInString, aInStringLength, tagTXT, aTagTXTLen, LT_ALPHA, LT_DELIMITER))
694 {
695 openTags--;
696 aOutString.AppendLiteral("<span class=\"moz-txt-tag\">");
697 aOutString.Append(tagTXT);
698 aOutString.AppendLiteral("</span></");
699 aOutString.AppendASCII(tagHTML);
700 aOutString.Append(char16_t('>'));
701 return true;
702 }
704 return false;
705 }
708 bool
709 mozTXTToHTMLConv::SmilyHit(const char16_t * aInString, int32_t aLength, bool col0,
710 const char* tagTXT, const char* imageName,
711 nsString& outputHTML, int32_t& glyphTextLen)
712 {
713 if ( !aInString || !tagTXT || !imageName )
714 return false;
716 int32_t tagLen = strlen(tagTXT);
718 uint32_t delim = (col0 ? 0 : 1) + tagLen;
720 if
721 (
722 (col0 || IsSpace(aInString[0]))
723 &&
724 (
725 aLength <= int32_t(delim) ||
726 IsSpace(aInString[delim]) ||
727 (aLength > int32_t(delim + 1)
728 &&
729 (
730 aInString[delim] == '.' ||
731 aInString[delim] == ',' ||
732 aInString[delim] == ';' ||
733 aInString[delim] == '8' ||
734 aInString[delim] == '>' ||
735 aInString[delim] == '!' ||
736 aInString[delim] == '?'
737 )
738 && IsSpace(aInString[delim + 1]))
739 )
740 && ItMatchesDelimited(aInString, aLength, NS_ConvertASCIItoUTF16(tagTXT).get(), tagLen,
741 col0 ? LT_IGNORE : LT_DELIMITER, LT_IGNORE)
742 // Note: tests at different pos for LT_IGNORE and LT_DELIMITER
743 )
744 {
745 if (!col0)
746 {
747 outputHTML.Truncate();
748 outputHTML.Append(char16_t(' '));
749 }
751 outputHTML.AppendLiteral("<span class=\""); // <span class="
752 AppendASCIItoUTF16(imageName, outputHTML); // e.g. smiley-frown
753 outputHTML.AppendLiteral("\" title=\""); // " title="
754 AppendASCIItoUTF16(tagTXT, outputHTML); // smiley tooltip
755 outputHTML.AppendLiteral("\"><span>"); // "><span>
756 AppendASCIItoUTF16(tagTXT, outputHTML); // original text
757 outputHTML.AppendLiteral("</span></span>"); // </span></span>
758 glyphTextLen = (col0 ? 0 : 1) + tagLen;
759 return true;
760 }
762 return false;
763 }
765 // the glyph is appended to aOutputString instead of the original string...
766 bool
767 mozTXTToHTMLConv::GlyphHit(const char16_t * aInString, int32_t aInLength, bool col0,
768 nsString& aOutputString, int32_t& glyphTextLen)
769 {
770 char16_t text0 = aInString[0];
771 char16_t text1 = aInString[1];
772 char16_t firstChar = (col0 ? text0 : text1);
774 // temporary variable used to store the glyph html text
775 nsAutoString outputHTML;
776 bool bTestSmilie;
777 bool bArg = false;
778 int i;
780 // refactor some of this mess to avoid code duplication and speed execution a bit
781 // there are two cases that need to be tried one after another. To avoid a lot of
782 // duplicate code, rolling into a loop
784 i = 0;
785 while ( i < 2 )
786 {
787 bTestSmilie = false;
788 if ( !i && (firstChar == ':' || firstChar == ';' || firstChar == '=' || firstChar == '>' || firstChar == '8' || firstChar == 'O'))
789 {
790 // first test passed
792 bTestSmilie = true;
793 bArg = col0;
794 }
795 if ( i && col0 && ( text1 == ':' || text1 == ';' || text1 == '=' || text1 == '>' || text1 == '8' || text1 == 'O' ) )
796 {
797 // second test passed
799 bTestSmilie = true;
800 bArg = false;
801 }
802 if ( bTestSmilie && (
803 SmilyHit(aInString, aInLength, bArg,
804 ":-)",
805 "moz-smiley-s1", // smile
806 outputHTML, glyphTextLen) ||
808 SmilyHit(aInString, aInLength, bArg,
809 ":)",
810 "moz-smiley-s1", // smile
811 outputHTML, glyphTextLen) ||
813 SmilyHit(aInString, aInLength, bArg,
814 ":-D",
815 "moz-smiley-s5", // laughing
816 outputHTML, glyphTextLen) ||
818 SmilyHit(aInString, aInLength, bArg,
819 ":-(",
820 "moz-smiley-s2", // frown
821 outputHTML, glyphTextLen) ||
823 SmilyHit(aInString, aInLength, bArg,
824 ":(",
825 "moz-smiley-s2", // frown
826 outputHTML, glyphTextLen) ||
828 SmilyHit(aInString, aInLength, bArg,
829 ":-[",
830 "moz-smiley-s6", // embarassed
831 outputHTML, glyphTextLen) ||
833 SmilyHit(aInString, aInLength, bArg,
834 ";-)",
835 "moz-smiley-s3", // wink
836 outputHTML, glyphTextLen) ||
838 SmilyHit(aInString, aInLength, col0,
839 ";)",
840 "moz-smiley-s3", // wink
841 outputHTML, glyphTextLen) ||
843 SmilyHit(aInString, aInLength, bArg,
844 ":-\\",
845 "moz-smiley-s7", // undecided
846 outputHTML, glyphTextLen) ||
848 SmilyHit(aInString, aInLength, bArg,
849 ":-P",
850 "moz-smiley-s4", // tongue
851 outputHTML, glyphTextLen) ||
853 SmilyHit(aInString, aInLength, bArg,
854 ";-P",
855 "moz-smiley-s4", // tongue
856 outputHTML, glyphTextLen) ||
858 SmilyHit(aInString, aInLength, bArg,
859 "=-O",
860 "moz-smiley-s8", // surprise
861 outputHTML, glyphTextLen) ||
863 SmilyHit(aInString, aInLength, bArg,
864 ":-*",
865 "moz-smiley-s9", // kiss
866 outputHTML, glyphTextLen) ||
868 SmilyHit(aInString, aInLength, bArg,
869 ">:o",
870 "moz-smiley-s10", // yell
871 outputHTML, glyphTextLen) ||
873 SmilyHit(aInString, aInLength, bArg,
874 ">:-o",
875 "moz-smiley-s10", // yell
876 outputHTML, glyphTextLen) ||
878 SmilyHit(aInString, aInLength, bArg,
879 "8-)",
880 "moz-smiley-s11", // cool
881 outputHTML, glyphTextLen) ||
883 SmilyHit(aInString, aInLength, bArg,
884 ":-$",
885 "moz-smiley-s12", // money
886 outputHTML, glyphTextLen) ||
888 SmilyHit(aInString, aInLength, bArg,
889 ":-!",
890 "moz-smiley-s13", // foot
891 outputHTML, glyphTextLen) ||
893 SmilyHit(aInString, aInLength, bArg,
894 "O:-)",
895 "moz-smiley-s14", // innocent
896 outputHTML, glyphTextLen) ||
898 SmilyHit(aInString, aInLength, bArg,
899 ":'(",
900 "moz-smiley-s15", // cry
901 outputHTML, glyphTextLen) ||
903 SmilyHit(aInString, aInLength, bArg,
904 ":-X",
905 "moz-smiley-s16", // sealed
906 outputHTML, glyphTextLen)
907 )
908 )
909 {
910 aOutputString.Append(outputHTML);
911 return true;
912 }
913 i++;
914 }
915 if (text0 == '\f')
916 {
917 aOutputString.AppendLiteral("<span class='moz-txt-formfeed'></span>");
918 glyphTextLen = 1;
919 return true;
920 }
921 if (text0 == '+' || text1 == '+')
922 {
923 if (ItMatchesDelimited(aInString, aInLength,
924 MOZ_UTF16(" +/-"), 4,
925 LT_IGNORE, LT_IGNORE))
926 {
927 aOutputString.AppendLiteral(" ±");
928 glyphTextLen = 4;
929 return true;
930 }
931 if (col0 && ItMatchesDelimited(aInString, aInLength,
932 MOZ_UTF16("+/-"), 3,
933 LT_IGNORE, LT_IGNORE))
934 {
935 aOutputString.AppendLiteral("±");
936 glyphTextLen = 3;
937 return true;
938 }
939 }
941 // x^2 => x<sup>2</sup>, also handle powers x^-2, x^0.5
942 // implement regular expression /[\dA-Za-z\)\]}]\^-?\d+(\.\d+)*[^\dA-Za-z]/
943 if
944 (
945 text1 == '^'
946 &&
947 (
948 nsCRT::IsAsciiDigit(text0) || nsCRT::IsAsciiAlpha(text0) ||
949 text0 == ')' || text0 == ']' || text0 == '}'
950 )
951 &&
952 (
953 (2 < aInLength && nsCRT::IsAsciiDigit(aInString[2])) ||
954 (3 < aInLength && aInString[2] == '-' && nsCRT::IsAsciiDigit(aInString[3]))
955 )
956 )
957 {
958 // Find first non-digit
959 int32_t delimPos = 3; // skip "^" and first digit (or '-')
960 for (; delimPos < aInLength
961 &&
962 (
963 nsCRT::IsAsciiDigit(aInString[delimPos]) ||
964 (aInString[delimPos] == '.' && delimPos + 1 < aInLength &&
965 nsCRT::IsAsciiDigit(aInString[delimPos + 1]))
966 );
967 delimPos++)
968 ;
970 if (delimPos < aInLength && nsCRT::IsAsciiAlpha(aInString[delimPos]))
971 {
972 return false;
973 }
975 outputHTML.Truncate();
976 outputHTML += text0;
977 outputHTML.AppendLiteral(
978 "<sup class=\"moz-txt-sup\">"
979 "<span style=\"display:inline-block;width:0;height:0;overflow:hidden\">"
980 "^</span>");
982 aOutputString.Append(outputHTML);
983 aOutputString.Append(&aInString[2], delimPos - 2);
984 aOutputString.AppendLiteral("</sup>");
986 glyphTextLen = delimPos /* - 1 + 1 */ ;
987 return true;
988 }
989 /*
990 The following strings are not substituted:
991 |TXT |HTML |Reason
992 +------+---------+----------
993 -> ← Bug #454
994 => ⇐ dito
995 <- → dito
996 <= ⇒ dito
997 (tm) ™ dito
998 1/4 ¼ is triggered by 1/4 Part 1, 2/4 Part 2, ...
999 3/4 ¾ dito
1000 1/2 ½ similar
1001 */
1002 return false;
1003 }
1005 /***************************************************************************
1006 Library-internal Interface
1007 ****************************************************************************/
1009 mozTXTToHTMLConv::mozTXTToHTMLConv()
1010 {
1011 }
1013 mozTXTToHTMLConv::~mozTXTToHTMLConv()
1014 {
1015 }
1017 NS_IMPL_ISUPPORTS(mozTXTToHTMLConv,
1018 mozITXTToHTMLConv,
1019 nsIStreamConverter,
1020 nsIStreamListener,
1021 nsIRequestObserver)
1023 int32_t
1024 mozTXTToHTMLConv::CiteLevelTXT(const char16_t *line,
1025 uint32_t& logLineStart)
1026 {
1027 int32_t result = 0;
1028 int32_t lineLength = NS_strlen(line);
1030 bool moreCites = true;
1031 while (moreCites)
1032 {
1033 /* E.g. the following lines count as quote:
1035 > text
1036 //#ifdef QUOTE_RECOGNITION_AGGRESSIVE
1037 >text
1038 //#ifdef QUOTE_RECOGNITION_AGGRESSIVE
1039 > text
1040 ] text
1041 USER> text
1042 USER] text
1043 //#endif
1045 logLineStart is the position of "t" in this example
1046 */
1047 uint32_t i = logLineStart;
1049 #ifdef QUOTE_RECOGNITION_AGGRESSIVE
1050 for (; int32_t(i) < lineLength && IsSpace(line[i]); i++)
1051 ;
1052 for (; int32_t(i) < lineLength && nsCRT::IsAsciiAlpha(line[i])
1053 && nsCRT::IsUpper(line[i]) ; i++)
1054 ;
1055 if (int32_t(i) < lineLength && (line[i] == '>' || line[i] == ']'))
1056 #else
1057 if (int32_t(i) < lineLength && line[i] == '>')
1058 #endif
1059 {
1060 i++;
1061 if (int32_t(i) < lineLength && line[i] == ' ')
1062 i++;
1063 // sendmail/mbox
1064 // Placed here for performance increase
1065 const char16_t * indexString = &line[logLineStart];
1066 // here, |logLineStart < lineLength| is always true
1067 uint32_t minlength = std::min(uint32_t(6), NS_strlen(indexString));
1068 if (Substring(indexString,
1069 indexString+minlength).Equals(Substring(NS_LITERAL_STRING(">From "), 0, minlength),
1070 nsCaseInsensitiveStringComparator()))
1071 //XXX RFC2646
1072 moreCites = false;
1073 else
1074 {
1075 result++;
1076 logLineStart = i;
1077 }
1078 }
1079 else
1080 moreCites = false;
1081 }
1083 return result;
1084 }
1086 void
1087 mozTXTToHTMLConv::ScanTXT(const char16_t * aInString, int32_t aInStringLength, uint32_t whattodo, nsString& aOutString)
1088 {
1089 bool doURLs = 0 != (whattodo & kURLs);
1090 bool doGlyphSubstitution = 0 != (whattodo & kGlyphSubstitution);
1091 bool doStructPhrase = 0 != (whattodo & kStructPhrase);
1093 uint32_t structPhrase_strong = 0; // Number of currently open tags
1094 uint32_t structPhrase_underline = 0;
1095 uint32_t structPhrase_italic = 0;
1096 uint32_t structPhrase_code = 0;
1098 nsAutoString outputHTML; // moved here for performance increase
1100 for(uint32_t i = 0; int32_t(i) < aInStringLength;)
1101 {
1102 if (doGlyphSubstitution)
1103 {
1104 int32_t glyphTextLen;
1105 if (GlyphHit(&aInString[i], aInStringLength - i, i == 0, aOutString, glyphTextLen))
1106 {
1107 i += glyphTextLen;
1108 continue;
1109 }
1110 }
1112 if (doStructPhrase)
1113 {
1114 const char16_t * newOffset = aInString;
1115 int32_t newLength = aInStringLength;
1116 if (i > 0 ) // skip the first element?
1117 {
1118 newOffset = &aInString[i-1];
1119 newLength = aInStringLength - i + 1;
1120 }
1122 switch (aInString[i]) // Performance increase
1123 {
1124 case '*':
1125 if (StructPhraseHit(newOffset, newLength, i == 0,
1126 MOZ_UTF16("*"), 1,
1127 "b", "class=\"moz-txt-star\"",
1128 aOutString, structPhrase_strong))
1129 {
1130 i++;
1131 continue;
1132 }
1133 break;
1134 case '/':
1135 if (StructPhraseHit(newOffset, newLength, i == 0,
1136 MOZ_UTF16("/"), 1,
1137 "i", "class=\"moz-txt-slash\"",
1138 aOutString, structPhrase_italic))
1139 {
1140 i++;
1141 continue;
1142 }
1143 break;
1144 case '_':
1145 if (StructPhraseHit(newOffset, newLength, i == 0,
1146 MOZ_UTF16("_"), 1,
1147 "span" /* <u> is deprecated */,
1148 "class=\"moz-txt-underscore\"",
1149 aOutString, structPhrase_underline))
1150 {
1151 i++;
1152 continue;
1153 }
1154 break;
1155 case '|':
1156 if (StructPhraseHit(newOffset, newLength, i == 0,
1157 MOZ_UTF16("|"), 1,
1158 "code", "class=\"moz-txt-verticalline\"",
1159 aOutString, structPhrase_code))
1160 {
1161 i++;
1162 continue;
1163 }
1164 break;
1165 }
1166 }
1168 if (doURLs)
1169 {
1170 switch (aInString[i])
1171 {
1172 case ':':
1173 case '@':
1174 case '.':
1175 if ( (i == 0 || ((i > 0) && aInString[i - 1] != ' ')) && aInString[i +1] != ' ') // Performance increase
1176 {
1177 int32_t replaceBefore;
1178 int32_t replaceAfter;
1179 if (FindURL(aInString, aInStringLength, i, whattodo,
1180 outputHTML, replaceBefore, replaceAfter)
1181 && structPhrase_strong + structPhrase_italic +
1182 structPhrase_underline + structPhrase_code == 0
1183 /* workaround for bug #19445 */ )
1184 {
1185 aOutString.Cut(aOutString.Length() - replaceBefore, replaceBefore);
1186 aOutString += outputHTML;
1187 i += replaceAfter + 1;
1188 continue;
1189 }
1190 }
1191 break;
1192 } //switch
1193 }
1195 switch (aInString[i])
1196 {
1197 // Special symbols
1198 case '<':
1199 case '>':
1200 case '&':
1201 EscapeChar(aInString[i], aOutString, false);
1202 i++;
1203 break;
1204 // Normal characters
1205 default:
1206 aOutString += aInString[i];
1207 i++;
1208 break;
1209 }
1210 }
1211 }
1213 void
1214 mozTXTToHTMLConv::ScanHTML(nsString& aInString, uint32_t whattodo, nsString &aOutString)
1215 {
1216 // some common variables we were recalculating
1217 // every time inside the for loop...
1218 int32_t lengthOfInString = aInString.Length();
1219 const char16_t * uniBuffer = aInString.get();
1221 #ifdef DEBUG_BenB_Perf
1222 PRTime parsing_start = PR_IntervalNow();
1223 #endif
1225 // Look for simple entities not included in a tags and scan them.
1226 /* Skip all tags ("<[...]>") and content in an a tag ("<a[...]</a>")
1227 or in a tag ("<!--[...]-->").
1228 Unescape the rest (text between tags) and pass it to ScanTXT. */
1229 for (int32_t i = 0; i < lengthOfInString;)
1230 {
1231 if (aInString[i] == '<') // html tag
1232 {
1233 uint32_t start = uint32_t(i);
1234 if (nsCRT::ToLower((char)aInString[uint32_t(i) + 1]) == 'a')
1235 // if a tag, skip until </a>
1236 {
1237 i = aInString.Find("</a>", true, i);
1238 if (i == kNotFound)
1239 i = lengthOfInString;
1240 else
1241 i += 4;
1242 }
1243 else if (aInString[uint32_t(i) + 1] == '!' && aInString[uint32_t(i) + 2] == '-' &&
1244 aInString[uint32_t(i) + 3] == '-')
1245 //if out-commended code, skip until -->
1246 {
1247 i = aInString.Find("-->", false, i);
1248 if (i == kNotFound)
1249 i = lengthOfInString;
1250 else
1251 i += 3;
1253 }
1254 else // just skip tag (attributes etc.)
1255 {
1256 i = aInString.FindChar('>', i);
1257 if (i == kNotFound)
1258 i = lengthOfInString;
1259 else
1260 i++;
1261 }
1262 aOutString.Append(&uniBuffer[start], uint32_t(i) - start);
1263 }
1264 else
1265 {
1266 uint32_t start = uint32_t(i);
1267 i = aInString.FindChar('<', i);
1268 if (i == kNotFound)
1269 i = lengthOfInString;
1271 nsString tempString;
1272 tempString.SetCapacity(uint32_t((uint32_t(i) - start) * growthRate));
1273 UnescapeStr(uniBuffer, start, uint32_t(i) - start, tempString);
1274 ScanTXT(tempString.get(), tempString.Length(), whattodo, aOutString);
1275 }
1276 }
1278 #ifdef DEBUG_BenB_Perf
1279 printf("ScanHTML time: %d ms\n", PR_IntervalToMilliseconds(PR_IntervalNow() - parsing_start));
1280 #endif
1281 }
1283 /****************************************************************************
1284 XPCOM Interface
1285 *****************************************************************************/
1287 NS_IMETHODIMP
1288 mozTXTToHTMLConv::Convert(nsIInputStream *aFromStream,
1289 const char *aFromType,
1290 const char *aToType,
1291 nsISupports *aCtxt, nsIInputStream **_retval)
1292 {
1293 return NS_ERROR_NOT_IMPLEMENTED;
1294 }
1296 NS_IMETHODIMP
1297 mozTXTToHTMLConv::AsyncConvertData(const char *aFromType,
1298 const char *aToType,
1299 nsIStreamListener *aListener, nsISupports *aCtxt) {
1300 return NS_ERROR_NOT_IMPLEMENTED;
1301 }
1303 NS_IMETHODIMP
1304 mozTXTToHTMLConv::OnDataAvailable(nsIRequest* request, nsISupports *ctxt,
1305 nsIInputStream *inStr, uint64_t sourceOffset,
1306 uint32_t count)
1307 {
1308 return NS_ERROR_NOT_IMPLEMENTED;
1309 }
1311 NS_IMETHODIMP
1312 mozTXTToHTMLConv::OnStartRequest(nsIRequest* request, nsISupports *ctxt)
1313 {
1314 return NS_ERROR_NOT_IMPLEMENTED;
1315 }
1317 NS_IMETHODIMP
1318 mozTXTToHTMLConv::OnStopRequest(nsIRequest* request, nsISupports *ctxt,
1319 nsresult aStatus)
1320 {
1321 return NS_ERROR_NOT_IMPLEMENTED;
1322 }
1324 NS_IMETHODIMP
1325 mozTXTToHTMLConv::CiteLevelTXT(const char16_t *line, uint32_t *logLineStart,
1326 uint32_t *_retval)
1327 {
1328 if (!logLineStart || !_retval || !line)
1329 return NS_ERROR_NULL_POINTER;
1330 *_retval = CiteLevelTXT(line, *logLineStart);
1331 return NS_OK;
1332 }
1334 NS_IMETHODIMP
1335 mozTXTToHTMLConv::ScanTXT(const char16_t *text, uint32_t whattodo,
1336 char16_t **_retval)
1337 {
1338 NS_ENSURE_ARG(text);
1340 // FIX ME!!!
1341 nsString outString;
1342 int32_t inLength = NS_strlen(text);
1343 // by setting a large capacity up front, we save time
1344 // when appending characters to the output string because we don't
1345 // need to reallocate and re-copy the characters already in the out String.
1346 NS_ASSERTION(inLength, "ScanTXT passed 0 length string");
1347 if (inLength == 0) {
1348 *_retval = NS_strdup(text);
1349 return NS_OK;
1350 }
1352 outString.SetCapacity(uint32_t(inLength * growthRate));
1353 ScanTXT(text, inLength, whattodo, outString);
1355 *_retval = ToNewUnicode(outString);
1356 return *_retval ? NS_OK : NS_ERROR_OUT_OF_MEMORY;
1357 }
1359 NS_IMETHODIMP
1360 mozTXTToHTMLConv::ScanHTML(const char16_t *text, uint32_t whattodo,
1361 char16_t **_retval)
1362 {
1363 NS_ENSURE_ARG(text);
1365 // FIX ME!!!
1366 nsString outString;
1367 nsString inString (text); // look at this nasty extra copy of the entire input buffer!
1368 outString.SetCapacity(uint32_t(inString.Length() * growthRate));
1370 ScanHTML(inString, whattodo, outString);
1371 *_retval = ToNewUnicode(outString);
1372 return *_retval ? NS_OK : NS_ERROR_OUT_OF_MEMORY;
1373 }
1375 nsresult
1376 MOZ_NewTXTToHTMLConv(mozTXTToHTMLConv** aConv)
1377 {
1378 NS_PRECONDITION(aConv != nullptr, "null ptr");
1379 if (!aConv)
1380 return NS_ERROR_NULL_POINTER;
1382 *aConv = new mozTXTToHTMLConv();
1383 if (!*aConv)
1384 return NS_ERROR_OUT_OF_MEMORY;
1386 NS_ADDREF(*aConv);
1387 // return (*aConv)->Init();
1388 return NS_OK;
1389 }
