1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/netwerk/streamconv/converters/mozTXTToHTMLConv.cpp Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,1389 @@
1.4 +/* -*- Mode: C; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
1.5 +/* This Source Code Form is subject to the terms of the Mozilla Public
1.6 + * License, v. 2.0. If a copy of the MPL was not distributed with this
1.7 + * file, You can obtain one at http://mozilla.org/MPL/2.0/. */
1.8 +
1.9 +#include "mozTXTToHTMLConv.h"
1.10 +#include "nsNetUtil.h"
1.11 +#include "nsUnicharUtils.h"
1.12 +#include "nsCRT.h"
1.13 +#include "nsIExternalProtocolHandler.h"
1.14 +#include "nsIIOService.h"
1.15 +
1.16 +#include <algorithm>
1.17 +
1.18 +#ifdef DEBUG_BenB_Perf
1.19 +#include "prtime.h"
1.20 +#include "prinrval.h"
1.21 +#endif
1.22 +
1.23 +const double growthRate = 1.2;
1.24 +
1.25 +// Bug 183111, editor now replaces multiple spaces with leading
1.26 +// 0xA0's and a single ending space, so need to treat 0xA0's as spaces.
1.27 +// 0xA0 is the Latin1/Unicode character for "non-breaking space (nbsp)"
1.28 +// Also recognize the Japanese ideographic space 0x3000 as a space.
1.29 +static inline bool IsSpace(const char16_t aChar)
1.30 +{
1.31 + return (nsCRT::IsAsciiSpace(aChar) || aChar == 0xA0 || aChar == 0x3000);
1.32 +}
1.33 +
1.34 +// Escape Char will take ch, escape it and append the result to
1.35 +// aStringToAppendTo
1.36 +void
1.37 +mozTXTToHTMLConv::EscapeChar(const char16_t ch, nsString& aStringToAppendTo,
1.38 + bool inAttribute)
1.39 +{
1.40 + switch (ch)
1.41 + {
1.42 + case '<':
1.43 + aStringToAppendTo.AppendLiteral("<");
1.44 + break;
1.45 + case '>':
1.46 + aStringToAppendTo.AppendLiteral(">");
1.47 + break;
1.48 + case '&':
1.49 + aStringToAppendTo.AppendLiteral("&");
1.50 + break;
1.51 + case '"':
1.52 + if (inAttribute)
1.53 + {
1.54 + aStringToAppendTo.AppendLiteral(""");
1.55 + break;
1.56 + }
1.57 + // else fall through
1.58 + default:
1.59 + aStringToAppendTo += ch;
1.60 + }
1.61 +
1.62 + return;
1.63 +}
1.64 +
1.65 +// EscapeStr takes the passed in string and
1.66 +// escapes it IN PLACE.
1.67 +void
1.68 +mozTXTToHTMLConv::EscapeStr(nsString& aInString, bool inAttribute)
1.69 +{
1.70 + // the replace substring routines
1.71 + // don't seem to work if you have a character
1.72 + // in the in string that is also in the replacement
1.73 + // string! =(
1.74 + //aInString.ReplaceSubstring("&", "&");
1.75 + //aInString.ReplaceSubstring("<", "<");
1.76 + //aInString.ReplaceSubstring(">", ">");
1.77 + for (uint32_t i = 0; i < aInString.Length();)
1.78 + {
1.79 + switch (aInString[i])
1.80 + {
1.81 + case '<':
1.82 + aInString.Cut(i, 1);
1.83 + aInString.Insert(NS_LITERAL_STRING("<"), i);
1.84 + i += 4; // skip past the integers we just added
1.85 + break;
1.86 + case '>':
1.87 + aInString.Cut(i, 1);
1.88 + aInString.Insert(NS_LITERAL_STRING(">"), i);
1.89 + i += 4; // skip past the integers we just added
1.90 + break;
1.91 + case '&':
1.92 + aInString.Cut(i, 1);
1.93 + aInString.Insert(NS_LITERAL_STRING("&"), i);
1.94 + i += 5; // skip past the integers we just added
1.95 + break;
1.96 + case '"':
1.97 + if (inAttribute)
1.98 + {
1.99 + aInString.Cut(i, 1);
1.100 + aInString.Insert(NS_LITERAL_STRING("""), i);
1.101 + i += 6;
1.102 + break;
1.103 + }
1.104 + // else fall through
1.105 + default:
1.106 + i++;
1.107 + }
1.108 + }
1.109 +}
1.110 +
1.111 +void
1.112 +mozTXTToHTMLConv::UnescapeStr(const char16_t * aInString, int32_t aStartPos, int32_t aLength, nsString& aOutString)
1.113 +{
1.114 + const char16_t * subString = nullptr;
1.115 + for (uint32_t i = aStartPos; int32_t(i) - aStartPos < aLength;)
1.116 + {
1.117 + int32_t remainingChars = i - aStartPos;
1.118 + if (aInString[i] == '&')
1.119 + {
1.120 + subString = &aInString[i];
1.121 + if (!nsCRT::strncmp(subString, MOZ_UTF16("<"), std::min(4, aLength - remainingChars)))
1.122 + {
1.123 + aOutString.Append(char16_t('<'));
1.124 + i += 4;
1.125 + }
1.126 + else if (!nsCRT::strncmp(subString, MOZ_UTF16(">"), std::min(4, aLength - remainingChars)))
1.127 + {
1.128 + aOutString.Append(char16_t('>'));
1.129 + i += 4;
1.130 + }
1.131 + else if (!nsCRT::strncmp(subString, MOZ_UTF16("&"), std::min(5, aLength - remainingChars)))
1.132 + {
1.133 + aOutString.Append(char16_t('&'));
1.134 + i += 5;
1.135 + }
1.136 + else if (!nsCRT::strncmp(subString, MOZ_UTF16("""), std::min(6, aLength - remainingChars)))
1.137 + {
1.138 + aOutString.Append(char16_t('"'));
1.139 + i += 6;
1.140 + }
1.141 + else
1.142 + {
1.143 + aOutString += aInString[i];
1.144 + i++;
1.145 + }
1.146 + }
1.147 + else
1.148 + {
1.149 + aOutString += aInString[i];
1.150 + i++;
1.151 + }
1.152 + }
1.153 +}
1.154 +
1.155 +void
1.156 +mozTXTToHTMLConv::CompleteAbbreviatedURL(const char16_t * aInString, int32_t aInLength,
1.157 + const uint32_t pos, nsString& aOutString)
1.158 +{
1.159 + NS_ASSERTION(int32_t(pos) < aInLength, "bad args to CompleteAbbreviatedURL, see bug #190851");
1.160 + if (int32_t(pos) >= aInLength)
1.161 + return;
1.162 +
1.163 + if (aInString[pos] == '@')
1.164 + {
1.165 + // only pre-pend a mailto url if the string contains a .domain in it..
1.166 + //i.e. we want to linkify johndoe@foo.com but not "let's meet @8pm"
1.167 + nsDependentString inString(aInString, aInLength);
1.168 + if (inString.FindChar('.', pos) != kNotFound) // if we have a '.' after the @ sign....
1.169 + {
1.170 + aOutString.AssignLiteral("mailto:");
1.171 + aOutString += aInString;
1.172 + }
1.173 + }
1.174 + else if (aInString[pos] == '.')
1.175 + {
1.176 + if (ItMatchesDelimited(aInString, aInLength,
1.177 + MOZ_UTF16("www."), 4, LT_IGNORE, LT_IGNORE))
1.178 + {
1.179 + aOutString.AssignLiteral("http://");
1.180 + aOutString += aInString;
1.181 + }
1.182 + else if (ItMatchesDelimited(aInString,aInLength, MOZ_UTF16("ftp."), 4, LT_IGNORE, LT_IGNORE))
1.183 + {
1.184 + aOutString.AssignLiteral("ftp://");
1.185 + aOutString += aInString;
1.186 + }
1.187 + }
1.188 +}
1.189 +
1.190 +bool
1.191 +mozTXTToHTMLConv::FindURLStart(const char16_t * aInString, int32_t aInLength,
1.192 + const uint32_t pos, const modetype check,
1.193 + uint32_t& start)
1.194 +{
1.195 + switch(check)
1.196 + { // no breaks, because end of blocks is never reached
1.197 + case RFC1738:
1.198 + {
1.199 + if (!nsCRT::strncmp(&aInString[std::max(int32_t(pos - 4), 0)], MOZ_UTF16("<URL:"), 5))
1.200 + {
1.201 + start = pos + 1;
1.202 + return true;
1.203 + }
1.204 + else
1.205 + return false;
1.206 + }
1.207 + case RFC2396E:
1.208 + {
1.209 + nsString temp(aInString, aInLength);
1.210 + int32_t i = pos <= 0 ? kNotFound : temp.RFindCharInSet(MOZ_UTF16("<>\""), pos - 1);
1.211 + if (i != kNotFound && (temp[uint32_t(i)] == '<' ||
1.212 + temp[uint32_t(i)] == '"'))
1.213 + {
1.214 + start = uint32_t(++i);
1.215 + return start < pos;
1.216 + }
1.217 + else
1.218 + return false;
1.219 + }
1.220 + case freetext:
1.221 + {
1.222 + int32_t i = pos - 1;
1.223 + for (; i >= 0 && (
1.224 + nsCRT::IsAsciiAlpha(aInString[uint32_t(i)]) ||
1.225 + nsCRT::IsAsciiDigit(aInString[uint32_t(i)]) ||
1.226 + aInString[uint32_t(i)] == '+' ||
1.227 + aInString[uint32_t(i)] == '-' ||
1.228 + aInString[uint32_t(i)] == '.'
1.229 + ); i--)
1.230 + ;
1.231 + if (++i >= 0 && uint32_t(i) < pos && nsCRT::IsAsciiAlpha(aInString[uint32_t(i)]))
1.232 + {
1.233 + start = uint32_t(i);
1.234 + return true;
1.235 + }
1.236 + else
1.237 + return false;
1.238 + }
1.239 + case abbreviated:
1.240 + {
1.241 + int32_t i = pos - 1;
1.242 + // This disallows non-ascii-characters for email.
1.243 + // Currently correct, but revisit later after standards changed.
1.244 + bool isEmail = aInString[pos] == (char16_t)'@';
1.245 + // These chars mark the start of the URL
1.246 + for (; i >= 0
1.247 + && aInString[uint32_t(i)] != '>' && aInString[uint32_t(i)] != '<'
1.248 + && aInString[uint32_t(i)] != '"' && aInString[uint32_t(i)] != '\''
1.249 + && aInString[uint32_t(i)] != '`' && aInString[uint32_t(i)] != ','
1.250 + && aInString[uint32_t(i)] != '{' && aInString[uint32_t(i)] != '['
1.251 + && aInString[uint32_t(i)] != '(' && aInString[uint32_t(i)] != '|'
1.252 + && aInString[uint32_t(i)] != '\\'
1.253 + && !IsSpace(aInString[uint32_t(i)])
1.254 + && (!isEmail || nsCRT::IsAscii(aInString[uint32_t(i)]))
1.255 + ; i--)
1.256 + ;
1.257 + if
1.258 + (
1.259 + ++i >= 0 && uint32_t(i) < pos
1.260 + &&
1.261 + (
1.262 + nsCRT::IsAsciiAlpha(aInString[uint32_t(i)]) ||
1.263 + nsCRT::IsAsciiDigit(aInString[uint32_t(i)])
1.264 + )
1.265 + )
1.266 + {
1.267 + start = uint32_t(i);
1.268 + return true;
1.269 + }
1.270 + else
1.271 + return false;
1.272 + }
1.273 + default:
1.274 + return false;
1.275 + } //switch
1.276 +}
1.277 +
1.278 +bool
1.279 +mozTXTToHTMLConv::FindURLEnd(const char16_t * aInString, int32_t aInStringLength, const uint32_t pos,
1.280 + const modetype check, const uint32_t start, uint32_t& end)
1.281 +{
1.282 + switch(check)
1.283 + { // no breaks, because end of blocks is never reached
1.284 + case RFC1738:
1.285 + case RFC2396E:
1.286 + {
1.287 + nsString temp(aInString, aInStringLength);
1.288 +
1.289 + int32_t i = temp.FindCharInSet(MOZ_UTF16("<>\""), pos + 1);
1.290 + if (i != kNotFound && temp[uint32_t(i--)] ==
1.291 + (check == RFC1738 || temp[start - 1] == '<' ? '>' : '"'))
1.292 + {
1.293 + end = uint32_t(i);
1.294 + return end > pos;
1.295 + }
1.296 + return false;
1.297 + }
1.298 + case freetext:
1.299 + case abbreviated:
1.300 + {
1.301 + uint32_t i = pos + 1;
1.302 + bool isEmail = aInString[pos] == (char16_t)'@';
1.303 + bool seenOpeningParenthesis = false; // there is a '(' earlier in the URL
1.304 + bool seenOpeningSquareBracket = false; // there is a '[' earlier in the URL
1.305 + for (; int32_t(i) < aInStringLength; i++)
1.306 + {
1.307 + // These chars mark the end of the URL
1.308 + if (aInString[i] == '>' || aInString[i] == '<' ||
1.309 + aInString[i] == '"' || aInString[i] == '`' ||
1.310 + aInString[i] == '}' || aInString[i] == '{' ||
1.311 + aInString[i] == '|' ||
1.312 + (aInString[i] == ')' && !seenOpeningParenthesis) ||
1.313 + (aInString[i] == ']' && !seenOpeningSquareBracket) ||
1.314 + // Allow IPv6 adresses like http://[1080::8:800:200C:417A]/foo.
1.315 + (aInString[i] == '[' && i > 2 &&
1.316 + (aInString[i - 1] != '/' || aInString[i - 2] != '/')) ||
1.317 + IsSpace(aInString[i]))
1.318 + break;
1.319 + // Disallow non-ascii-characters for email.
1.320 + // Currently correct, but revisit later after standards changed.
1.321 + if (isEmail && (
1.322 + aInString[i] == '(' || aInString[i] == '\'' ||
1.323 + !nsCRT::IsAscii(aInString[i])))
1.324 + break;
1.325 + if (aInString[i] == '(')
1.326 + seenOpeningParenthesis = true;
1.327 + if (aInString[i] == '[')
1.328 + seenOpeningSquareBracket = true;
1.329 + }
1.330 + // These chars are allowed in the middle of the URL, but not at end.
1.331 + // Technically they are, but are used in normal text after the URL.
1.332 + while (--i > pos && (
1.333 + aInString[i] == '.' || aInString[i] == ',' || aInString[i] == ';' ||
1.334 + aInString[i] == '!' || aInString[i] == '?' || aInString[i] == '-' ||
1.335 + aInString[i] == ':' || aInString[i] == '\''
1.336 + ))
1.337 + ;
1.338 + if (i > pos)
1.339 + {
1.340 + end = i;
1.341 + return true;
1.342 + }
1.343 + return false;
1.344 + }
1.345 + default:
1.346 + return false;
1.347 + } //switch
1.348 +}
1.349 +
1.350 +void
1.351 +mozTXTToHTMLConv::CalculateURLBoundaries(const char16_t * aInString, int32_t aInStringLength,
1.352 + const uint32_t pos, const uint32_t whathasbeendone,
1.353 + const modetype check, const uint32_t start, const uint32_t end,
1.354 + nsString& txtURL, nsString& desc,
1.355 + int32_t& replaceBefore, int32_t& replaceAfter)
1.356 +{
1.357 + uint32_t descstart = start;
1.358 + switch(check)
1.359 + {
1.360 + case RFC1738:
1.361 + {
1.362 + descstart = start - 5;
1.363 + desc.Append(&aInString[descstart], end - descstart + 2); // include "<URL:" and ">"
1.364 + replaceAfter = end - pos + 1;
1.365 + } break;
1.366 + case RFC2396E:
1.367 + {
1.368 + descstart = start - 1;
1.369 + desc.Append(&aInString[descstart], end - descstart + 2); // include brackets
1.370 + replaceAfter = end - pos + 1;
1.371 + } break;
1.372 + case freetext:
1.373 + case abbreviated:
1.374 + {
1.375 + descstart = start;
1.376 + desc.Append(&aInString[descstart], end - start + 1); // don't include brackets
1.377 + replaceAfter = end - pos;
1.378 + } break;
1.379 + default: break;
1.380 + } //switch
1.381 +
1.382 + EscapeStr(desc, false);
1.383 +
1.384 + txtURL.Append(&aInString[start], end - start + 1);
1.385 + txtURL.StripWhitespace();
1.386 +
1.387 + // FIX ME
1.388 + nsAutoString temp2;
1.389 + ScanTXT(&aInString[descstart], pos - descstart, ~kURLs /*prevents loop*/ & whathasbeendone, temp2);
1.390 + replaceBefore = temp2.Length();
1.391 + return;
1.392 +}
1.393 +
1.394 +bool mozTXTToHTMLConv::ShouldLinkify(const nsCString& aURL)
1.395 +{
1.396 + if (!mIOService)
1.397 + return false;
1.398 +
1.399 + nsAutoCString scheme;
1.400 + nsresult rv = mIOService->ExtractScheme(aURL, scheme);
1.401 + if(NS_FAILED(rv))
1.402 + return false;
1.403 +
1.404 + // Get the handler for this scheme.
1.405 + nsCOMPtr<nsIProtocolHandler> handler;
1.406 + rv = mIOService->GetProtocolHandler(scheme.get(), getter_AddRefs(handler));
1.407 + if(NS_FAILED(rv))
1.408 + return false;
1.409 +
1.410 + // Is it an external protocol handler? If not, linkify it.
1.411 + nsCOMPtr<nsIExternalProtocolHandler> externalHandler = do_QueryInterface(handler);
1.412 + if (!externalHandler)
1.413 + return true; // handler is built-in, linkify it!
1.414 +
1.415 + // If external app exists for the scheme then linkify it.
1.416 + bool exists;
1.417 + rv = externalHandler->ExternalAppExistsForScheme(scheme, &exists);
1.418 + return(NS_SUCCEEDED(rv) && exists);
1.419 +}
1.420 +
1.421 +bool
1.422 +mozTXTToHTMLConv::CheckURLAndCreateHTML(
1.423 + const nsString& txtURL, const nsString& desc, const modetype mode,
1.424 + nsString& outputHTML)
1.425 +{
1.426 + // Create *uri from txtURL
1.427 + nsCOMPtr<nsIURI> uri;
1.428 + nsresult rv;
1.429 + // Lazily initialize mIOService
1.430 + if (!mIOService)
1.431 + {
1.432 + mIOService = do_GetIOService();
1.433 +
1.434 + if (!mIOService)
1.435 + return false;
1.436 + }
1.437 +
1.438 + // See if the url should be linkified.
1.439 + NS_ConvertUTF16toUTF8 utf8URL(txtURL);
1.440 + if (!ShouldLinkify(utf8URL))
1.441 + return false;
1.442 +
1.443 + // it would be faster if we could just check to see if there is a protocol
1.444 + // handler for the url and return instead of actually trying to create a url...
1.445 + rv = mIOService->NewURI(utf8URL, nullptr, nullptr, getter_AddRefs(uri));
1.446 +
1.447 + // Real work
1.448 + if (NS_SUCCEEDED(rv) && uri)
1.449 + {
1.450 + outputHTML.AssignLiteral("<a class=\"moz-txt-link-");
1.451 + switch(mode)
1.452 + {
1.453 + case RFC1738:
1.454 + outputHTML.AppendLiteral("rfc1738");
1.455 + break;
1.456 + case RFC2396E:
1.457 + outputHTML.AppendLiteral("rfc2396E");
1.458 + break;
1.459 + case freetext:
1.460 + outputHTML.AppendLiteral("freetext");
1.461 + break;
1.462 + case abbreviated:
1.463 + outputHTML.AppendLiteral("abbreviated");
1.464 + break;
1.465 + default: break;
1.466 + }
1.467 + nsAutoString escapedURL(txtURL);
1.468 + EscapeStr(escapedURL, true);
1.469 +
1.470 + outputHTML.AppendLiteral("\" href=\"");
1.471 + outputHTML += escapedURL;
1.472 + outputHTML.AppendLiteral("\">");
1.473 + outputHTML += desc;
1.474 + outputHTML.AppendLiteral("</a>");
1.475 + return true;
1.476 + }
1.477 + else
1.478 + return false;
1.479 +}
1.480 +
1.481 +NS_IMETHODIMP mozTXTToHTMLConv::FindURLInPlaintext(const char16_t * aInString, int32_t aInLength, int32_t aPos, int32_t * aStartPos, int32_t * aEndPos)
1.482 +{
1.483 + // call FindURL on the passed in string
1.484 + nsAutoString outputHTML; // we'll ignore the generated output HTML
1.485 +
1.486 + *aStartPos = -1;
1.487 + *aEndPos = -1;
1.488 +
1.489 + FindURL(aInString, aInLength, aPos, kURLs, outputHTML, *aStartPos, *aEndPos);
1.490 +
1.491 + return NS_OK;
1.492 +}
1.493 +
1.494 +bool
1.495 +mozTXTToHTMLConv::FindURL(const char16_t * aInString, int32_t aInLength, const uint32_t pos,
1.496 + const uint32_t whathasbeendone,
1.497 + nsString& outputHTML, int32_t& replaceBefore, int32_t& replaceAfter)
1.498 +{
1.499 + enum statetype {unchecked, invalid, startok, endok, success};
1.500 + static const modetype ranking[] = {RFC1738, RFC2396E, freetext, abbreviated};
1.501 +
1.502 + statetype state[mozTXTToHTMLConv_lastMode + 1]; // 0(=unknown)..lastMode
1.503 + /* I don't like this abuse of enums as index for the array,
1.504 + but I don't know a better method */
1.505 +
1.506 + // Define, which modes to check
1.507 + /* all modes but abbreviated are checked for text[pos] == ':',
1.508 + only abbreviated for '.', RFC2396E and abbreviated for '@' */
1.509 + for (modetype iState = unknown; iState <= mozTXTToHTMLConv_lastMode;
1.510 + iState = modetype(iState + 1))
1.511 + state[iState] = aInString[pos] == ':' ? unchecked : invalid;
1.512 + switch (aInString[pos])
1.513 + {
1.514 + case '@':
1.515 + state[RFC2396E] = unchecked;
1.516 + // no break here
1.517 + case '.':
1.518 + state[abbreviated] = unchecked;
1.519 + break;
1.520 + case ':':
1.521 + state[abbreviated] = invalid;
1.522 + break;
1.523 + default:
1.524 + break;
1.525 + }
1.526 +
1.527 + // Test, first successful mode wins, sequence defined by |ranking|
1.528 + int32_t iCheck = 0; // the currently tested modetype
1.529 + modetype check = ranking[iCheck];
1.530 + for (; iCheck < mozTXTToHTMLConv_numberOfModes && state[check] != success;
1.531 + iCheck++)
1.532 + /* check state from last run.
1.533 + If this is the first, check this one, which isn't = success yet */
1.534 + {
1.535 + check = ranking[iCheck];
1.536 +
1.537 + uint32_t start, end;
1.538 +
1.539 + if (state[check] == unchecked)
1.540 + if (FindURLStart(aInString, aInLength, pos, check, start))
1.541 + state[check] = startok;
1.542 +
1.543 + if (state[check] == startok)
1.544 + if (FindURLEnd(aInString, aInLength, pos, check, start, end))
1.545 + state[check] = endok;
1.546 +
1.547 + if (state[check] == endok)
1.548 + {
1.549 + nsAutoString txtURL, desc;
1.550 + int32_t resultReplaceBefore, resultReplaceAfter;
1.551 +
1.552 + CalculateURLBoundaries(aInString, aInLength, pos, whathasbeendone, check, start, end,
1.553 + txtURL, desc,
1.554 + resultReplaceBefore, resultReplaceAfter);
1.555 +
1.556 + if (aInString[pos] != ':')
1.557 + {
1.558 + nsAutoString temp = txtURL;
1.559 + txtURL.SetLength(0);
1.560 + CompleteAbbreviatedURL(temp.get(),temp.Length(), pos - start, txtURL);
1.561 + }
1.562 +
1.563 + if (!txtURL.IsEmpty() && CheckURLAndCreateHTML(txtURL, desc, check,
1.564 + outputHTML))
1.565 + {
1.566 + replaceBefore = resultReplaceBefore;
1.567 + replaceAfter = resultReplaceAfter;
1.568 + state[check] = success;
1.569 + }
1.570 + } // if
1.571 + } // for
1.572 + return state[check] == success;
1.573 +}
1.574 +
1.575 +bool
1.576 +mozTXTToHTMLConv::ItMatchesDelimited(const char16_t * aInString,
1.577 + int32_t aInLength, const char16_t* rep, int32_t aRepLen,
1.578 + LIMTYPE before, LIMTYPE after)
1.579 +{
1.580 +
1.581 + // this little method gets called a LOT. I found we were spending a
1.582 + // lot of time just calculating the length of the variable "rep"
1.583 + // over and over again every time we called it. So we're now passing
1.584 + // an integer in here.
1.585 + int32_t textLen = aInLength;
1.586 +
1.587 + if
1.588 + (
1.589 + ((before == LT_IGNORE && (after == LT_IGNORE || after == LT_DELIMITER))
1.590 + && textLen < aRepLen) ||
1.591 + ((before != LT_IGNORE || (after != LT_IGNORE && after != LT_DELIMITER))
1.592 + && textLen < aRepLen + 1) ||
1.593 + (before != LT_IGNORE && after != LT_IGNORE && after != LT_DELIMITER
1.594 + && textLen < aRepLen + 2)
1.595 + )
1.596 + return false;
1.597 +
1.598 + char16_t text0 = aInString[0];
1.599 + char16_t textAfterPos = aInString[aRepLen + (before == LT_IGNORE ? 0 : 1)];
1.600 +
1.601 + if
1.602 + (
1.603 + (before == LT_ALPHA
1.604 + && !nsCRT::IsAsciiAlpha(text0)) ||
1.605 + (before == LT_DIGIT
1.606 + && !nsCRT::IsAsciiDigit(text0)) ||
1.607 + (before == LT_DELIMITER
1.608 + &&
1.609 + (
1.610 + nsCRT::IsAsciiAlpha(text0) ||
1.611 + nsCRT::IsAsciiDigit(text0) ||
1.612 + text0 == *rep
1.613 + )) ||
1.614 + (after == LT_ALPHA
1.615 + && !nsCRT::IsAsciiAlpha(textAfterPos)) ||
1.616 + (after == LT_DIGIT
1.617 + && !nsCRT::IsAsciiDigit(textAfterPos)) ||
1.618 + (after == LT_DELIMITER
1.619 + &&
1.620 + (
1.621 + nsCRT::IsAsciiAlpha(textAfterPos) ||
1.622 + nsCRT::IsAsciiDigit(textAfterPos) ||
1.623 + textAfterPos == *rep
1.624 + )) ||
1.625 + !Substring(Substring(aInString, aInString+aInLength),
1.626 + (before == LT_IGNORE ? 0 : 1),
1.627 + aRepLen).Equals(Substring(rep, rep+aRepLen),
1.628 + nsCaseInsensitiveStringComparator())
1.629 + )
1.630 + return false;
1.631 +
1.632 + return true;
1.633 +}
1.634 +
1.635 +uint32_t
1.636 +mozTXTToHTMLConv::NumberOfMatches(const char16_t * aInString, int32_t aInStringLength,
1.637 + const char16_t* rep, int32_t aRepLen, LIMTYPE before, LIMTYPE after)
1.638 +{
1.639 + uint32_t result = 0;
1.640 +
1.641 + for (int32_t i = 0; i < aInStringLength; i++)
1.642 + {
1.643 + const char16_t * indexIntoString = &aInString[i];
1.644 + if (ItMatchesDelimited(indexIntoString, aInStringLength - i, rep, aRepLen, before, after))
1.645 + result++;
1.646 + }
1.647 + return result;
1.648 +}
1.649 +
1.650 +
1.651 +// NOTE: the converted html for the phrase is appended to aOutString
1.652 +// tagHTML and attributeHTML are plain ASCII (literal strings, in fact)
1.653 +bool
1.654 +mozTXTToHTMLConv::StructPhraseHit(const char16_t * aInString, int32_t aInStringLength, bool col0,
1.655 + const char16_t* tagTXT, int32_t aTagTXTLen,
1.656 + const char* tagHTML, const char* attributeHTML,
1.657 + nsString& aOutString, uint32_t& openTags)
1.658 +{
1.659 + /* We're searching for the following pattern:
1.660 + LT_DELIMITER - "*" - ALPHA -
1.661 + [ some text (maybe more "*"-pairs) - ALPHA ] "*" - LT_DELIMITER.
1.662 + <strong> is only inserted, if existence of a pair could be verified
1.663 + We use the first opening/closing tag, if we can choose */
1.664 +
1.665 + const char16_t * newOffset = aInString;
1.666 + int32_t newLength = aInStringLength;
1.667 + if (!col0) // skip the first element?
1.668 + {
1.669 + newOffset = &aInString[1];
1.670 + newLength = aInStringLength - 1;
1.671 + }
1.672 +
1.673 + // opening tag
1.674 + if
1.675 + (
1.676 + ItMatchesDelimited(aInString, aInStringLength, tagTXT, aTagTXTLen,
1.677 + (col0 ? LT_IGNORE : LT_DELIMITER), LT_ALPHA) // is opening tag
1.678 + && NumberOfMatches(newOffset, newLength, tagTXT, aTagTXTLen,
1.679 + LT_ALPHA, LT_DELIMITER) // remaining closing tags
1.680 + > openTags
1.681 + )
1.682 + {
1.683 + openTags++;
1.684 + aOutString.AppendLiteral("<");
1.685 + aOutString.AppendASCII(tagHTML);
1.686 + aOutString.Append(char16_t(' '));
1.687 + aOutString.AppendASCII(attributeHTML);
1.688 + aOutString.AppendLiteral("><span class=\"moz-txt-tag\">");
1.689 + aOutString.Append(tagTXT);
1.690 + aOutString.AppendLiteral("</span>");
1.691 + return true;
1.692 + }
1.693 +
1.694 + // closing tag
1.695 + else if (openTags > 0
1.696 + && ItMatchesDelimited(aInString, aInStringLength, tagTXT, aTagTXTLen, LT_ALPHA, LT_DELIMITER))
1.697 + {
1.698 + openTags--;
1.699 + aOutString.AppendLiteral("<span class=\"moz-txt-tag\">");
1.700 + aOutString.Append(tagTXT);
1.701 + aOutString.AppendLiteral("</span></");
1.702 + aOutString.AppendASCII(tagHTML);
1.703 + aOutString.Append(char16_t('>'));
1.704 + return true;
1.705 + }
1.706 +
1.707 + return false;
1.708 +}
1.709 +
1.710 +
1.711 +bool
1.712 +mozTXTToHTMLConv::SmilyHit(const char16_t * aInString, int32_t aLength, bool col0,
1.713 + const char* tagTXT, const char* imageName,
1.714 + nsString& outputHTML, int32_t& glyphTextLen)
1.715 +{
1.716 + if ( !aInString || !tagTXT || !imageName )
1.717 + return false;
1.718 +
1.719 + int32_t tagLen = strlen(tagTXT);
1.720 +
1.721 + uint32_t delim = (col0 ? 0 : 1) + tagLen;
1.722 +
1.723 + if
1.724 + (
1.725 + (col0 || IsSpace(aInString[0]))
1.726 + &&
1.727 + (
1.728 + aLength <= int32_t(delim) ||
1.729 + IsSpace(aInString[delim]) ||
1.730 + (aLength > int32_t(delim + 1)
1.731 + &&
1.732 + (
1.733 + aInString[delim] == '.' ||
1.734 + aInString[delim] == ',' ||
1.735 + aInString[delim] == ';' ||
1.736 + aInString[delim] == '8' ||
1.737 + aInString[delim] == '>' ||
1.738 + aInString[delim] == '!' ||
1.739 + aInString[delim] == '?'
1.740 + )
1.741 + && IsSpace(aInString[delim + 1]))
1.742 + )
1.743 + && ItMatchesDelimited(aInString, aLength, NS_ConvertASCIItoUTF16(tagTXT).get(), tagLen,
1.744 + col0 ? LT_IGNORE : LT_DELIMITER, LT_IGNORE)
1.745 + // Note: tests at different pos for LT_IGNORE and LT_DELIMITER
1.746 + )
1.747 + {
1.748 + if (!col0)
1.749 + {
1.750 + outputHTML.Truncate();
1.751 + outputHTML.Append(char16_t(' '));
1.752 + }
1.753 +
1.754 + outputHTML.AppendLiteral("<span class=\""); // <span class="
1.755 + AppendASCIItoUTF16(imageName, outputHTML); // e.g. smiley-frown
1.756 + outputHTML.AppendLiteral("\" title=\""); // " title="
1.757 + AppendASCIItoUTF16(tagTXT, outputHTML); // smiley tooltip
1.758 + outputHTML.AppendLiteral("\"><span>"); // "><span>
1.759 + AppendASCIItoUTF16(tagTXT, outputHTML); // original text
1.760 + outputHTML.AppendLiteral("</span></span>"); // </span></span>
1.761 + glyphTextLen = (col0 ? 0 : 1) + tagLen;
1.762 + return true;
1.763 + }
1.764 +
1.765 + return false;
1.766 +}
1.767 +
1.768 +// the glyph is appended to aOutputString instead of the original string...
1.769 +bool
1.770 +mozTXTToHTMLConv::GlyphHit(const char16_t * aInString, int32_t aInLength, bool col0,
1.771 + nsString& aOutputString, int32_t& glyphTextLen)
1.772 +{
1.773 + char16_t text0 = aInString[0];
1.774 + char16_t text1 = aInString[1];
1.775 + char16_t firstChar = (col0 ? text0 : text1);
1.776 +
1.777 + // temporary variable used to store the glyph html text
1.778 + nsAutoString outputHTML;
1.779 + bool bTestSmilie;
1.780 + bool bArg = false;
1.781 + int i;
1.782 +
1.783 + // refactor some of this mess to avoid code duplication and speed execution a bit
1.784 + // there are two cases that need to be tried one after another. To avoid a lot of
1.785 + // duplicate code, rolling into a loop
1.786 +
1.787 + i = 0;
1.788 + while ( i < 2 )
1.789 + {
1.790 + bTestSmilie = false;
1.791 + if ( !i && (firstChar == ':' || firstChar == ';' || firstChar == '=' || firstChar == '>' || firstChar == '8' || firstChar == 'O'))
1.792 + {
1.793 + // first test passed
1.794 +
1.795 + bTestSmilie = true;
1.796 + bArg = col0;
1.797 + }
1.798 + if ( i && col0 && ( text1 == ':' || text1 == ';' || text1 == '=' || text1 == '>' || text1 == '8' || text1 == 'O' ) )
1.799 + {
1.800 + // second test passed
1.801 +
1.802 + bTestSmilie = true;
1.803 + bArg = false;
1.804 + }
1.805 + if ( bTestSmilie && (
1.806 + SmilyHit(aInString, aInLength, bArg,
1.807 + ":-)",
1.808 + "moz-smiley-s1", // smile
1.809 + outputHTML, glyphTextLen) ||
1.810 +
1.811 + SmilyHit(aInString, aInLength, bArg,
1.812 + ":)",
1.813 + "moz-smiley-s1", // smile
1.814 + outputHTML, glyphTextLen) ||
1.815 +
1.816 + SmilyHit(aInString, aInLength, bArg,
1.817 + ":-D",
1.818 + "moz-smiley-s5", // laughing
1.819 + outputHTML, glyphTextLen) ||
1.820 +
1.821 + SmilyHit(aInString, aInLength, bArg,
1.822 + ":-(",
1.823 + "moz-smiley-s2", // frown
1.824 + outputHTML, glyphTextLen) ||
1.825 +
1.826 + SmilyHit(aInString, aInLength, bArg,
1.827 + ":(",
1.828 + "moz-smiley-s2", // frown
1.829 + outputHTML, glyphTextLen) ||
1.830 +
1.831 + SmilyHit(aInString, aInLength, bArg,
1.832 + ":-[",
1.833 + "moz-smiley-s6", // embarassed
1.834 + outputHTML, glyphTextLen) ||
1.835 +
1.836 + SmilyHit(aInString, aInLength, bArg,
1.837 + ";-)",
1.838 + "moz-smiley-s3", // wink
1.839 + outputHTML, glyphTextLen) ||
1.840 +
1.841 + SmilyHit(aInString, aInLength, col0,
1.842 + ";)",
1.843 + "moz-smiley-s3", // wink
1.844 + outputHTML, glyphTextLen) ||
1.845 +
1.846 + SmilyHit(aInString, aInLength, bArg,
1.847 + ":-\\",
1.848 + "moz-smiley-s7", // undecided
1.849 + outputHTML, glyphTextLen) ||
1.850 +
1.851 + SmilyHit(aInString, aInLength, bArg,
1.852 + ":-P",
1.853 + "moz-smiley-s4", // tongue
1.854 + outputHTML, glyphTextLen) ||
1.855 +
1.856 + SmilyHit(aInString, aInLength, bArg,
1.857 + ";-P",
1.858 + "moz-smiley-s4", // tongue
1.859 + outputHTML, glyphTextLen) ||
1.860 +
1.861 + SmilyHit(aInString, aInLength, bArg,
1.862 + "=-O",
1.863 + "moz-smiley-s8", // surprise
1.864 + outputHTML, glyphTextLen) ||
1.865 +
1.866 + SmilyHit(aInString, aInLength, bArg,
1.867 + ":-*",
1.868 + "moz-smiley-s9", // kiss
1.869 + outputHTML, glyphTextLen) ||
1.870 +
1.871 + SmilyHit(aInString, aInLength, bArg,
1.872 + ">:o",
1.873 + "moz-smiley-s10", // yell
1.874 + outputHTML, glyphTextLen) ||
1.875 +
1.876 + SmilyHit(aInString, aInLength, bArg,
1.877 + ">:-o",
1.878 + "moz-smiley-s10", // yell
1.879 + outputHTML, glyphTextLen) ||
1.880 +
1.881 + SmilyHit(aInString, aInLength, bArg,
1.882 + "8-)",
1.883 + "moz-smiley-s11", // cool
1.884 + outputHTML, glyphTextLen) ||
1.885 +
1.886 + SmilyHit(aInString, aInLength, bArg,
1.887 + ":-$",
1.888 + "moz-smiley-s12", // money
1.889 + outputHTML, glyphTextLen) ||
1.890 +
1.891 + SmilyHit(aInString, aInLength, bArg,
1.892 + ":-!",
1.893 + "moz-smiley-s13", // foot
1.894 + outputHTML, glyphTextLen) ||
1.895 +
1.896 + SmilyHit(aInString, aInLength, bArg,
1.897 + "O:-)",
1.898 + "moz-smiley-s14", // innocent
1.899 + outputHTML, glyphTextLen) ||
1.900 +
1.901 + SmilyHit(aInString, aInLength, bArg,
1.902 + ":'(",
1.903 + "moz-smiley-s15", // cry
1.904 + outputHTML, glyphTextLen) ||
1.905 +
1.906 + SmilyHit(aInString, aInLength, bArg,
1.907 + ":-X",
1.908 + "moz-smiley-s16", // sealed
1.909 + outputHTML, glyphTextLen)
1.910 + )
1.911 + )
1.912 + {
1.913 + aOutputString.Append(outputHTML);
1.914 + return true;
1.915 + }
1.916 + i++;
1.917 + }
1.918 + if (text0 == '\f')
1.919 + {
1.920 + aOutputString.AppendLiteral("<span class='moz-txt-formfeed'></span>");
1.921 + glyphTextLen = 1;
1.922 + return true;
1.923 + }
1.924 + if (text0 == '+' || text1 == '+')
1.925 + {
1.926 + if (ItMatchesDelimited(aInString, aInLength,
1.927 + MOZ_UTF16(" +/-"), 4,
1.928 + LT_IGNORE, LT_IGNORE))
1.929 + {
1.930 + aOutputString.AppendLiteral(" ±");
1.931 + glyphTextLen = 4;
1.932 + return true;
1.933 + }
1.934 + if (col0 && ItMatchesDelimited(aInString, aInLength,
1.935 + MOZ_UTF16("+/-"), 3,
1.936 + LT_IGNORE, LT_IGNORE))
1.937 + {
1.938 + aOutputString.AppendLiteral("±");
1.939 + glyphTextLen = 3;
1.940 + return true;
1.941 + }
1.942 + }
1.943 +
1.944 + // x^2 => x<sup>2</sup>, also handle powers x^-2, x^0.5
1.945 + // implement regular expression /[\dA-Za-z\)\]}]\^-?\d+(\.\d+)*[^\dA-Za-z]/
1.946 + if
1.947 + (
1.948 + text1 == '^'
1.949 + &&
1.950 + (
1.951 + nsCRT::IsAsciiDigit(text0) || nsCRT::IsAsciiAlpha(text0) ||
1.952 + text0 == ')' || text0 == ']' || text0 == '}'
1.953 + )
1.954 + &&
1.955 + (
1.956 + (2 < aInLength && nsCRT::IsAsciiDigit(aInString[2])) ||
1.957 + (3 < aInLength && aInString[2] == '-' && nsCRT::IsAsciiDigit(aInString[3]))
1.958 + )
1.959 + )
1.960 + {
1.961 + // Find first non-digit
1.962 + int32_t delimPos = 3; // skip "^" and first digit (or '-')
1.963 + for (; delimPos < aInLength
1.964 + &&
1.965 + (
1.966 + nsCRT::IsAsciiDigit(aInString[delimPos]) ||
1.967 + (aInString[delimPos] == '.' && delimPos + 1 < aInLength &&
1.968 + nsCRT::IsAsciiDigit(aInString[delimPos + 1]))
1.969 + );
1.970 + delimPos++)
1.971 + ;
1.972 +
1.973 + if (delimPos < aInLength && nsCRT::IsAsciiAlpha(aInString[delimPos]))
1.974 + {
1.975 + return false;
1.976 + }
1.977 +
1.978 + outputHTML.Truncate();
1.979 + outputHTML += text0;
1.980 + outputHTML.AppendLiteral(
1.981 + "<sup class=\"moz-txt-sup\">"
1.982 + "<span style=\"display:inline-block;width:0;height:0;overflow:hidden\">"
1.983 + "^</span>");
1.984 +
1.985 + aOutputString.Append(outputHTML);
1.986 + aOutputString.Append(&aInString[2], delimPos - 2);
1.987 + aOutputString.AppendLiteral("</sup>");
1.988 +
1.989 + glyphTextLen = delimPos /* - 1 + 1 */ ;
1.990 + return true;
1.991 + }
1.992 + /*
1.993 + The following strings are not substituted:
1.994 + |TXT |HTML |Reason
1.995 + +------+---------+----------
1.996 + -> ← Bug #454
1.997 + => ⇐ dito
1.998 + <- → dito
1.999 + <= ⇒ dito
1.1000 + (tm) ™ dito
1.1001 + 1/4 ¼ is triggered by 1/4 Part 1, 2/4 Part 2, ...
1.1002 + 3/4 ¾ dito
1.1003 + 1/2 ½ similar
1.1004 + */
1.1005 + return false;
1.1006 +}
1.1007 +
1.1008 +/***************************************************************************
1.1009 + Library-internal Interface
1.1010 +****************************************************************************/
1.1011 +
1.1012 +mozTXTToHTMLConv::mozTXTToHTMLConv()
1.1013 +{
1.1014 +}
1.1015 +
1.1016 +mozTXTToHTMLConv::~mozTXTToHTMLConv()
1.1017 +{
1.1018 +}
1.1019 +
1.1020 +NS_IMPL_ISUPPORTS(mozTXTToHTMLConv,
1.1021 + mozITXTToHTMLConv,
1.1022 + nsIStreamConverter,
1.1023 + nsIStreamListener,
1.1024 + nsIRequestObserver)
1.1025 +
1.1026 +int32_t
1.1027 +mozTXTToHTMLConv::CiteLevelTXT(const char16_t *line,
1.1028 + uint32_t& logLineStart)
1.1029 +{
1.1030 + int32_t result = 0;
1.1031 + int32_t lineLength = NS_strlen(line);
1.1032 +
1.1033 + bool moreCites = true;
1.1034 + while (moreCites)
1.1035 + {
1.1036 + /* E.g. the following lines count as quote:
1.1037 +
1.1038 + > text
1.1039 + //#ifdef QUOTE_RECOGNITION_AGGRESSIVE
1.1040 + >text
1.1041 + //#ifdef QUOTE_RECOGNITION_AGGRESSIVE
1.1042 + > text
1.1043 + ] text
1.1044 + USER> text
1.1045 + USER] text
1.1046 + //#endif
1.1047 +
1.1048 + logLineStart is the position of "t" in this example
1.1049 + */
1.1050 + uint32_t i = logLineStart;
1.1051 +
1.1052 +#ifdef QUOTE_RECOGNITION_AGGRESSIVE
1.1053 + for (; int32_t(i) < lineLength && IsSpace(line[i]); i++)
1.1054 + ;
1.1055 + for (; int32_t(i) < lineLength && nsCRT::IsAsciiAlpha(line[i])
1.1056 + && nsCRT::IsUpper(line[i]) ; i++)
1.1057 + ;
1.1058 + if (int32_t(i) < lineLength && (line[i] == '>' || line[i] == ']'))
1.1059 +#else
1.1060 + if (int32_t(i) < lineLength && line[i] == '>')
1.1061 +#endif
1.1062 + {
1.1063 + i++;
1.1064 + if (int32_t(i) < lineLength && line[i] == ' ')
1.1065 + i++;
1.1066 + // sendmail/mbox
1.1067 + // Placed here for performance increase
1.1068 + const char16_t * indexString = &line[logLineStart];
1.1069 + // here, |logLineStart < lineLength| is always true
1.1070 + uint32_t minlength = std::min(uint32_t(6), NS_strlen(indexString));
1.1071 + if (Substring(indexString,
1.1072 + indexString+minlength).Equals(Substring(NS_LITERAL_STRING(">From "), 0, minlength),
1.1073 + nsCaseInsensitiveStringComparator()))
1.1074 + //XXX RFC2646
1.1075 + moreCites = false;
1.1076 + else
1.1077 + {
1.1078 + result++;
1.1079 + logLineStart = i;
1.1080 + }
1.1081 + }
1.1082 + else
1.1083 + moreCites = false;
1.1084 + }
1.1085 +
1.1086 + return result;
1.1087 +}
1.1088 +
1.1089 +void
1.1090 +mozTXTToHTMLConv::ScanTXT(const char16_t * aInString, int32_t aInStringLength, uint32_t whattodo, nsString& aOutString)
1.1091 +{
1.1092 + bool doURLs = 0 != (whattodo & kURLs);
1.1093 + bool doGlyphSubstitution = 0 != (whattodo & kGlyphSubstitution);
1.1094 + bool doStructPhrase = 0 != (whattodo & kStructPhrase);
1.1095 +
1.1096 + uint32_t structPhrase_strong = 0; // Number of currently open tags
1.1097 + uint32_t structPhrase_underline = 0;
1.1098 + uint32_t structPhrase_italic = 0;
1.1099 + uint32_t structPhrase_code = 0;
1.1100 +
1.1101 + nsAutoString outputHTML; // moved here for performance increase
1.1102 +
1.1103 + for(uint32_t i = 0; int32_t(i) < aInStringLength;)
1.1104 + {
1.1105 + if (doGlyphSubstitution)
1.1106 + {
1.1107 + int32_t glyphTextLen;
1.1108 + if (GlyphHit(&aInString[i], aInStringLength - i, i == 0, aOutString, glyphTextLen))
1.1109 + {
1.1110 + i += glyphTextLen;
1.1111 + continue;
1.1112 + }
1.1113 + }
1.1114 +
1.1115 + if (doStructPhrase)
1.1116 + {
1.1117 + const char16_t * newOffset = aInString;
1.1118 + int32_t newLength = aInStringLength;
1.1119 + if (i > 0 ) // skip the first element?
1.1120 + {
1.1121 + newOffset = &aInString[i-1];
1.1122 + newLength = aInStringLength - i + 1;
1.1123 + }
1.1124 +
1.1125 + switch (aInString[i]) // Performance increase
1.1126 + {
1.1127 + case '*':
1.1128 + if (StructPhraseHit(newOffset, newLength, i == 0,
1.1129 + MOZ_UTF16("*"), 1,
1.1130 + "b", "class=\"moz-txt-star\"",
1.1131 + aOutString, structPhrase_strong))
1.1132 + {
1.1133 + i++;
1.1134 + continue;
1.1135 + }
1.1136 + break;
1.1137 + case '/':
1.1138 + if (StructPhraseHit(newOffset, newLength, i == 0,
1.1139 + MOZ_UTF16("/"), 1,
1.1140 + "i", "class=\"moz-txt-slash\"",
1.1141 + aOutString, structPhrase_italic))
1.1142 + {
1.1143 + i++;
1.1144 + continue;
1.1145 + }
1.1146 + break;
1.1147 + case '_':
1.1148 + if (StructPhraseHit(newOffset, newLength, i == 0,
1.1149 + MOZ_UTF16("_"), 1,
1.1150 + "span" /* <u> is deprecated */,
1.1151 + "class=\"moz-txt-underscore\"",
1.1152 + aOutString, structPhrase_underline))
1.1153 + {
1.1154 + i++;
1.1155 + continue;
1.1156 + }
1.1157 + break;
1.1158 + case '|':
1.1159 + if (StructPhraseHit(newOffset, newLength, i == 0,
1.1160 + MOZ_UTF16("|"), 1,
1.1161 + "code", "class=\"moz-txt-verticalline\"",
1.1162 + aOutString, structPhrase_code))
1.1163 + {
1.1164 + i++;
1.1165 + continue;
1.1166 + }
1.1167 + break;
1.1168 + }
1.1169 + }
1.1170 +
1.1171 + if (doURLs)
1.1172 + {
1.1173 + switch (aInString[i])
1.1174 + {
1.1175 + case ':':
1.1176 + case '@':
1.1177 + case '.':
1.1178 + if ( (i == 0 || ((i > 0) && aInString[i - 1] != ' ')) && aInString[i +1] != ' ') // Performance increase
1.1179 + {
1.1180 + int32_t replaceBefore;
1.1181 + int32_t replaceAfter;
1.1182 + if (FindURL(aInString, aInStringLength, i, whattodo,
1.1183 + outputHTML, replaceBefore, replaceAfter)
1.1184 + && structPhrase_strong + structPhrase_italic +
1.1185 + structPhrase_underline + structPhrase_code == 0
1.1186 + /* workaround for bug #19445 */ )
1.1187 + {
1.1188 + aOutString.Cut(aOutString.Length() - replaceBefore, replaceBefore);
1.1189 + aOutString += outputHTML;
1.1190 + i += replaceAfter + 1;
1.1191 + continue;
1.1192 + }
1.1193 + }
1.1194 + break;
1.1195 + } //switch
1.1196 + }
1.1197 +
1.1198 + switch (aInString[i])
1.1199 + {
1.1200 + // Special symbols
1.1201 + case '<':
1.1202 + case '>':
1.1203 + case '&':
1.1204 + EscapeChar(aInString[i], aOutString, false);
1.1205 + i++;
1.1206 + break;
1.1207 + // Normal characters
1.1208 + default:
1.1209 + aOutString += aInString[i];
1.1210 + i++;
1.1211 + break;
1.1212 + }
1.1213 + }
1.1214 +}
1.1215 +
1.1216 +void
1.1217 +mozTXTToHTMLConv::ScanHTML(nsString& aInString, uint32_t whattodo, nsString &aOutString)
1.1218 +{
1.1219 + // some common variables we were recalculating
1.1220 + // every time inside the for loop...
1.1221 + int32_t lengthOfInString = aInString.Length();
1.1222 + const char16_t * uniBuffer = aInString.get();
1.1223 +
1.1224 +#ifdef DEBUG_BenB_Perf
1.1225 + PRTime parsing_start = PR_IntervalNow();
1.1226 +#endif
1.1227 +
1.1228 + // Look for simple entities not included in a tags and scan them.
1.1229 + /* Skip all tags ("<[...]>") and content in an a tag ("<a[...]</a>")
1.1230 + or in a tag ("<!--[...]-->").
1.1231 + Unescape the rest (text between tags) and pass it to ScanTXT. */
1.1232 + for (int32_t i = 0; i < lengthOfInString;)
1.1233 + {
1.1234 + if (aInString[i] == '<') // html tag
1.1235 + {
1.1236 + uint32_t start = uint32_t(i);
1.1237 + if (nsCRT::ToLower((char)aInString[uint32_t(i) + 1]) == 'a')
1.1238 + // if a tag, skip until </a>
1.1239 + {
1.1240 + i = aInString.Find("</a>", true, i);
1.1241 + if (i == kNotFound)
1.1242 + i = lengthOfInString;
1.1243 + else
1.1244 + i += 4;
1.1245 + }
1.1246 + else if (aInString[uint32_t(i) + 1] == '!' && aInString[uint32_t(i) + 2] == '-' &&
1.1247 + aInString[uint32_t(i) + 3] == '-')
1.1248 + //if out-commended code, skip until -->
1.1249 + {
1.1250 + i = aInString.Find("-->", false, i);
1.1251 + if (i == kNotFound)
1.1252 + i = lengthOfInString;
1.1253 + else
1.1254 + i += 3;
1.1255 +
1.1256 + }
1.1257 + else // just skip tag (attributes etc.)
1.1258 + {
1.1259 + i = aInString.FindChar('>', i);
1.1260 + if (i == kNotFound)
1.1261 + i = lengthOfInString;
1.1262 + else
1.1263 + i++;
1.1264 + }
1.1265 + aOutString.Append(&uniBuffer[start], uint32_t(i) - start);
1.1266 + }
1.1267 + else
1.1268 + {
1.1269 + uint32_t start = uint32_t(i);
1.1270 + i = aInString.FindChar('<', i);
1.1271 + if (i == kNotFound)
1.1272 + i = lengthOfInString;
1.1273 +
1.1274 + nsString tempString;
1.1275 + tempString.SetCapacity(uint32_t((uint32_t(i) - start) * growthRate));
1.1276 + UnescapeStr(uniBuffer, start, uint32_t(i) - start, tempString);
1.1277 + ScanTXT(tempString.get(), tempString.Length(), whattodo, aOutString);
1.1278 + }
1.1279 + }
1.1280 +
1.1281 +#ifdef DEBUG_BenB_Perf
1.1282 + printf("ScanHTML time: %d ms\n", PR_IntervalToMilliseconds(PR_IntervalNow() - parsing_start));
1.1283 +#endif
1.1284 +}
1.1285 +
1.1286 +/****************************************************************************
1.1287 + XPCOM Interface
1.1288 +*****************************************************************************/
1.1289 +
1.1290 +NS_IMETHODIMP
1.1291 +mozTXTToHTMLConv::Convert(nsIInputStream *aFromStream,
1.1292 + const char *aFromType,
1.1293 + const char *aToType,
1.1294 + nsISupports *aCtxt, nsIInputStream **_retval)
1.1295 +{
1.1296 + return NS_ERROR_NOT_IMPLEMENTED;
1.1297 +}
1.1298 +
1.1299 +NS_IMETHODIMP
1.1300 +mozTXTToHTMLConv::AsyncConvertData(const char *aFromType,
1.1301 + const char *aToType,
1.1302 + nsIStreamListener *aListener, nsISupports *aCtxt) {
1.1303 + return NS_ERROR_NOT_IMPLEMENTED;
1.1304 +}
1.1305 +
1.1306 +NS_IMETHODIMP
1.1307 +mozTXTToHTMLConv::OnDataAvailable(nsIRequest* request, nsISupports *ctxt,
1.1308 + nsIInputStream *inStr, uint64_t sourceOffset,
1.1309 + uint32_t count)
1.1310 +{
1.1311 + return NS_ERROR_NOT_IMPLEMENTED;
1.1312 +}
1.1313 +
1.1314 +NS_IMETHODIMP
1.1315 +mozTXTToHTMLConv::OnStartRequest(nsIRequest* request, nsISupports *ctxt)
1.1316 +{
1.1317 + return NS_ERROR_NOT_IMPLEMENTED;
1.1318 +}
1.1319 +
1.1320 +NS_IMETHODIMP
1.1321 +mozTXTToHTMLConv::OnStopRequest(nsIRequest* request, nsISupports *ctxt,
1.1322 + nsresult aStatus)
1.1323 +{
1.1324 + return NS_ERROR_NOT_IMPLEMENTED;
1.1325 +}
1.1326 +
1.1327 +NS_IMETHODIMP
1.1328 +mozTXTToHTMLConv::CiteLevelTXT(const char16_t *line, uint32_t *logLineStart,
1.1329 + uint32_t *_retval)
1.1330 +{
1.1331 + if (!logLineStart || !_retval || !line)
1.1332 + return NS_ERROR_NULL_POINTER;
1.1333 + *_retval = CiteLevelTXT(line, *logLineStart);
1.1334 + return NS_OK;
1.1335 +}
1.1336 +
1.1337 +NS_IMETHODIMP
1.1338 +mozTXTToHTMLConv::ScanTXT(const char16_t *text, uint32_t whattodo,
1.1339 + char16_t **_retval)
1.1340 +{
1.1341 + NS_ENSURE_ARG(text);
1.1342 +
1.1343 + // FIX ME!!!
1.1344 + nsString outString;
1.1345 + int32_t inLength = NS_strlen(text);
1.1346 + // by setting a large capacity up front, we save time
1.1347 + // when appending characters to the output string because we don't
1.1348 + // need to reallocate and re-copy the characters already in the out String.
1.1349 + NS_ASSERTION(inLength, "ScanTXT passed 0 length string");
1.1350 + if (inLength == 0) {
1.1351 + *_retval = NS_strdup(text);
1.1352 + return NS_OK;
1.1353 + }
1.1354 +
1.1355 + outString.SetCapacity(uint32_t(inLength * growthRate));
1.1356 + ScanTXT(text, inLength, whattodo, outString);
1.1357 +
1.1358 + *_retval = ToNewUnicode(outString);
1.1359 + return *_retval ? NS_OK : NS_ERROR_OUT_OF_MEMORY;
1.1360 +}
1.1361 +
1.1362 +NS_IMETHODIMP
1.1363 +mozTXTToHTMLConv::ScanHTML(const char16_t *text, uint32_t whattodo,
1.1364 + char16_t **_retval)
1.1365 +{
1.1366 + NS_ENSURE_ARG(text);
1.1367 +
1.1368 + // FIX ME!!!
1.1369 + nsString outString;
1.1370 + nsString inString (text); // look at this nasty extra copy of the entire input buffer!
1.1371 + outString.SetCapacity(uint32_t(inString.Length() * growthRate));
1.1372 +
1.1373 + ScanHTML(inString, whattodo, outString);
1.1374 + *_retval = ToNewUnicode(outString);
1.1375 + return *_retval ? NS_OK : NS_ERROR_OUT_OF_MEMORY;
1.1376 +}
1.1377 +
1.1378 +nsresult
1.1379 +MOZ_NewTXTToHTMLConv(mozTXTToHTMLConv** aConv)
1.1380 +{
1.1381 + NS_PRECONDITION(aConv != nullptr, "null ptr");
1.1382 + if (!aConv)
1.1383 + return NS_ERROR_NULL_POINTER;
1.1384 +
1.1385 + *aConv = new mozTXTToHTMLConv();
1.1386 + if (!*aConv)
1.1387 + return NS_ERROR_OUT_OF_MEMORY;
1.1388 +
1.1389 + NS_ADDREF(*aConv);
1.1390 + // return (*aConv)->Init();
1.1391 + return NS_OK;
1.1392 +}
