1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/netwerk/streamconv/converters/nsDirIndexParser.cpp Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,434 @@
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 +/* This parsing code originally lived in xpfe/components/directory/ - bbaetz */
1.10 +
1.11 +#include "mozilla/ArrayUtils.h"
1.12 +
1.13 +#include "prprf.h"
1.14 +
1.15 +#include "nsDirIndexParser.h"
1.16 +#include "nsEscape.h"
1.17 +#include "nsIInputStream.h"
1.18 +#include "nsCRT.h"
1.19 +#include "mozilla/dom/FallbackEncoding.h"
1.20 +#include "nsITextToSubURI.h"
1.21 +#include "nsIDirIndex.h"
1.22 +#include "nsServiceManagerUtils.h"
1.23 +
1.24 +using namespace mozilla;
1.25 +
1.26 +NS_IMPL_ISUPPORTS(nsDirIndexParser,
1.27 + nsIRequestObserver,
1.28 + nsIStreamListener,
1.29 + nsIDirIndexParser)
1.30 +
1.31 +nsDirIndexParser::nsDirIndexParser() {
1.32 +}
1.33 +
1.34 +nsresult
1.35 +nsDirIndexParser::Init() {
1.36 + mLineStart = 0;
1.37 + mHasDescription = false;
1.38 + mFormat = nullptr;
1.39 + mozilla::dom::FallbackEncoding::FromLocale(mEncoding);
1.40 +
1.41 + nsresult rv;
1.42 + // XXX not threadsafe
1.43 + if (gRefCntParser++ == 0)
1.44 + rv = CallGetService(NS_ITEXTTOSUBURI_CONTRACTID, &gTextToSubURI);
1.45 + else
1.46 + rv = NS_OK;
1.47 +
1.48 + return rv;
1.49 +}
1.50 +
1.51 +nsDirIndexParser::~nsDirIndexParser() {
1.52 + delete[] mFormat;
1.53 + // XXX not threadsafe
1.54 + if (--gRefCntParser == 0) {
1.55 + NS_IF_RELEASE(gTextToSubURI);
1.56 + }
1.57 +}
1.58 +
1.59 +NS_IMETHODIMP
1.60 +nsDirIndexParser::SetListener(nsIDirIndexListener* aListener) {
1.61 + mListener = aListener;
1.62 + return NS_OK;
1.63 +}
1.64 +
1.65 +NS_IMETHODIMP
1.66 +nsDirIndexParser::GetListener(nsIDirIndexListener** aListener) {
1.67 + NS_IF_ADDREF(*aListener = mListener.get());
1.68 + return NS_OK;
1.69 +}
1.70 +
1.71 +NS_IMETHODIMP
1.72 +nsDirIndexParser::GetComment(char** aComment) {
1.73 + *aComment = ToNewCString(mComment);
1.74 +
1.75 + if (!*aComment)
1.76 + return NS_ERROR_OUT_OF_MEMORY;
1.77 +
1.78 + return NS_OK;
1.79 +}
1.80 +
1.81 +NS_IMETHODIMP
1.82 +nsDirIndexParser::SetEncoding(const char* aEncoding) {
1.83 + mEncoding.Assign(aEncoding);
1.84 + return NS_OK;
1.85 +}
1.86 +
1.87 +NS_IMETHODIMP
1.88 +nsDirIndexParser::GetEncoding(char** aEncoding) {
1.89 + *aEncoding = ToNewCString(mEncoding);
1.90 +
1.91 + if (!*aEncoding)
1.92 + return NS_ERROR_OUT_OF_MEMORY;
1.93 +
1.94 + return NS_OK;
1.95 +}
1.96 +
1.97 +NS_IMETHODIMP
1.98 +nsDirIndexParser::OnStartRequest(nsIRequest* aRequest, nsISupports* aCtxt) {
1.99 + return NS_OK;
1.100 +}
1.101 +
1.102 +NS_IMETHODIMP
1.103 +nsDirIndexParser::OnStopRequest(nsIRequest *aRequest, nsISupports *aCtxt,
1.104 + nsresult aStatusCode) {
1.105 + // Finish up
1.106 + if (mBuf.Length() > (uint32_t) mLineStart) {
1.107 + ProcessData(aRequest, aCtxt);
1.108 + }
1.109 +
1.110 + return NS_OK;
1.111 +}
1.112 +
1.113 +nsDirIndexParser::Field
1.114 +nsDirIndexParser::gFieldTable[] = {
1.115 + { "Filename", FIELD_FILENAME },
1.116 + { "Description", FIELD_DESCRIPTION },
1.117 + { "Content-Length", FIELD_CONTENTLENGTH },
1.118 + { "Last-Modified", FIELD_LASTMODIFIED },
1.119 + { "Content-Type", FIELD_CONTENTTYPE },
1.120 + { "File-Type", FIELD_FILETYPE },
1.121 + { nullptr, FIELD_UNKNOWN }
1.122 +};
1.123 +
1.124 +nsrefcnt nsDirIndexParser::gRefCntParser = 0;
1.125 +nsITextToSubURI *nsDirIndexParser::gTextToSubURI;
1.126 +
1.127 +nsresult
1.128 +nsDirIndexParser::ParseFormat(const char* aFormatStr) {
1.129 + // Parse a "200" format line, and remember the fields and their
1.130 + // ordering in mFormat. Multiple 200 lines stomp on each other.
1.131 +
1.132 + // Lets find out how many elements we have.
1.133 + // easier to do this then realloc
1.134 + const char* pos = aFormatStr;
1.135 + unsigned int num = 0;
1.136 + do {
1.137 + while (*pos && nsCRT::IsAsciiSpace(char16_t(*pos)))
1.138 + ++pos;
1.139 +
1.140 + ++num;
1.141 + // There are a maximum of six allowed header fields (doubled plus
1.142 + // terminator, just in case) -- Bug 443299
1.143 + if (num > (2 * ArrayLength(gFieldTable)))
1.144 + return NS_ERROR_UNEXPECTED;
1.145 +
1.146 + if (! *pos)
1.147 + break;
1.148 +
1.149 + while (*pos && !nsCRT::IsAsciiSpace(char16_t(*pos)))
1.150 + ++pos;
1.151 +
1.152 + } while (*pos);
1.153 +
1.154 + delete[] mFormat;
1.155 + mFormat = new int[num+1];
1.156 + // Prevent nullptr Deref - Bug 443299
1.157 + if (mFormat == nullptr)
1.158 + return NS_ERROR_OUT_OF_MEMORY;
1.159 + mFormat[num] = -1;
1.160 +
1.161 + int formatNum=0;
1.162 + do {
1.163 + while (*aFormatStr && nsCRT::IsAsciiSpace(char16_t(*aFormatStr)))
1.164 + ++aFormatStr;
1.165 +
1.166 + if (! *aFormatStr)
1.167 + break;
1.168 +
1.169 + nsAutoCString name;
1.170 + int32_t len = 0;
1.171 + while (aFormatStr[len] && !nsCRT::IsAsciiSpace(char16_t(aFormatStr[len])))
1.172 + ++len;
1.173 + name.SetCapacity(len + 1);
1.174 + name.Append(aFormatStr, len);
1.175 + aFormatStr += len;
1.176 +
1.177 + // Okay, we're gonna monkey with the nsStr. Bold!
1.178 + name.SetLength(nsUnescapeCount(name.BeginWriting()));
1.179 +
1.180 + // All tokens are case-insensitive - http://www.mozilla.org/projects/netlib/dirindexformat.html
1.181 + if (name.LowerCaseEqualsLiteral("description"))
1.182 + mHasDescription = true;
1.183 +
1.184 + for (Field* i = gFieldTable; i->mName; ++i) {
1.185 + if (name.EqualsIgnoreCase(i->mName)) {
1.186 + mFormat[formatNum] = i->mType;
1.187 + ++formatNum;
1.188 + break;
1.189 + }
1.190 + }
1.191 +
1.192 + } while (*aFormatStr);
1.193 +
1.194 + return NS_OK;
1.195 +}
1.196 +
1.197 +nsresult
1.198 +nsDirIndexParser::ParseData(nsIDirIndex *aIdx, char* aDataStr) {
1.199 + // Parse a "201" data line, using the field ordering specified in
1.200 + // mFormat.
1.201 +
1.202 + if (!mFormat) {
1.203 + // Ignore if we haven't seen a format yet.
1.204 + return NS_OK;
1.205 + }
1.206 +
1.207 + nsresult rv = NS_OK;
1.208 +
1.209 + nsAutoCString filename;
1.210 +
1.211 + for (int32_t i = 0; mFormat[i] != -1; ++i) {
1.212 + // If we've exhausted the data before we run out of fields, just
1.213 + // bail.
1.214 + if (! *aDataStr)
1.215 + break;
1.216 +
1.217 + while (*aDataStr && nsCRT::IsAsciiSpace(*aDataStr))
1.218 + ++aDataStr;
1.219 +
1.220 + char *value = aDataStr;
1.221 +
1.222 + if (*aDataStr == '"' || *aDataStr == '\'') {
1.223 + // it's a quoted string. snarf everything up to the next quote character
1.224 + const char quotechar = *(aDataStr++);
1.225 + ++value;
1.226 + while (*aDataStr && *aDataStr != quotechar)
1.227 + ++aDataStr;
1.228 + *aDataStr++ = '\0';
1.229 +
1.230 + if (! aDataStr) {
1.231 + NS_WARNING("quoted value not terminated");
1.232 + }
1.233 + } else {
1.234 + // it's unquoted. snarf until we see whitespace.
1.235 + value = aDataStr;
1.236 + while (*aDataStr && (!nsCRT::IsAsciiSpace(*aDataStr)))
1.237 + ++aDataStr;
1.238 + *aDataStr++ = '\0';
1.239 + }
1.240 +
1.241 + fieldType t = fieldType(mFormat[i]);
1.242 + switch (t) {
1.243 + case FIELD_FILENAME: {
1.244 + // don't unescape at this point, so that UnEscapeAndConvert() can
1.245 + filename = value;
1.246 +
1.247 + bool success = false;
1.248 +
1.249 + nsAutoString entryuri;
1.250 +
1.251 + if (gTextToSubURI) {
1.252 + char16_t *result = nullptr;
1.253 + if (NS_SUCCEEDED(rv = gTextToSubURI->UnEscapeAndConvert(mEncoding.get(), filename.get(),
1.254 + &result)) && (result)) {
1.255 + if (*result) {
1.256 + aIdx->SetLocation(filename.get());
1.257 + if (!mHasDescription)
1.258 + aIdx->SetDescription(result);
1.259 + success = true;
1.260 + }
1.261 + NS_Free(result);
1.262 + } else {
1.263 + NS_WARNING("UnEscapeAndConvert error");
1.264 + }
1.265 + }
1.266 +
1.267 + if (!success) {
1.268 + // if unsuccessfully at charset conversion, then
1.269 + // just fallback to unescape'ing in-place
1.270 + // XXX - this shouldn't be using UTF8, should it?
1.271 + // when can we fail to get the service, anyway? - bbaetz
1.272 + aIdx->SetLocation(filename.get());
1.273 + if (!mHasDescription) {
1.274 + aIdx->SetDescription(NS_ConvertUTF8toUTF16(value).get());
1.275 + }
1.276 + }
1.277 + }
1.278 + break;
1.279 + case FIELD_DESCRIPTION:
1.280 + nsUnescape(value);
1.281 + aIdx->SetDescription(NS_ConvertUTF8toUTF16(value).get());
1.282 + break;
1.283 + case FIELD_CONTENTLENGTH:
1.284 + {
1.285 + int64_t len;
1.286 + int32_t status = PR_sscanf(value, "%lld", &len);
1.287 + if (status == 1)
1.288 + aIdx->SetSize(len);
1.289 + else
1.290 + aIdx->SetSize(UINT64_MAX); // UINT64_MAX means unknown
1.291 + }
1.292 + break;
1.293 + case FIELD_LASTMODIFIED:
1.294 + {
1.295 + PRTime tm;
1.296 + nsUnescape(value);
1.297 + if (PR_ParseTimeString(value, false, &tm) == PR_SUCCESS) {
1.298 + aIdx->SetLastModified(tm);
1.299 + }
1.300 + }
1.301 + break;
1.302 + case FIELD_CONTENTTYPE:
1.303 + aIdx->SetContentType(value);
1.304 + break;
1.305 + case FIELD_FILETYPE:
1.306 + // unescape in-place
1.307 + nsUnescape(value);
1.308 + if (!nsCRT::strcasecmp(value, "directory")) {
1.309 + aIdx->SetType(nsIDirIndex::TYPE_DIRECTORY);
1.310 + } else if (!nsCRT::strcasecmp(value, "file")) {
1.311 + aIdx->SetType(nsIDirIndex::TYPE_FILE);
1.312 + } else if (!nsCRT::strcasecmp(value, "symbolic-link")) {
1.313 + aIdx->SetType(nsIDirIndex::TYPE_SYMLINK);
1.314 + } else {
1.315 + aIdx->SetType(nsIDirIndex::TYPE_UNKNOWN);
1.316 + }
1.317 + break;
1.318 + case FIELD_UNKNOWN:
1.319 + // ignore
1.320 + break;
1.321 + }
1.322 + }
1.323 +
1.324 + return NS_OK;
1.325 +}
1.326 +
1.327 +NS_IMETHODIMP
1.328 +nsDirIndexParser::OnDataAvailable(nsIRequest *aRequest, nsISupports *aCtxt,
1.329 + nsIInputStream *aStream,
1.330 + uint64_t aSourceOffset,
1.331 + uint32_t aCount) {
1.332 + if (aCount < 1)
1.333 + return NS_OK;
1.334 +
1.335 + int32_t len = mBuf.Length();
1.336 +
1.337 + // Ensure that our mBuf has capacity to hold the data we're about to
1.338 + // read.
1.339 + if (!mBuf.SetLength(len + aCount, fallible_t()))
1.340 + return NS_ERROR_OUT_OF_MEMORY;
1.341 +
1.342 + // Now read the data into our buffer.
1.343 + nsresult rv;
1.344 + uint32_t count;
1.345 + rv = aStream->Read(mBuf.BeginWriting() + len, aCount, &count);
1.346 + if (NS_FAILED(rv)) return rv;
1.347 +
1.348 + // Set the string's length according to the amount of data we've read.
1.349 + // Note: we know this to work on nsCString. This isn't guaranteed to
1.350 + // work on other strings.
1.351 + mBuf.SetLength(len + count);
1.352 +
1.353 + return ProcessData(aRequest, aCtxt);
1.354 +}
1.355 +
1.356 +nsresult
1.357 +nsDirIndexParser::ProcessData(nsIRequest *aRequest, nsISupports *aCtxt) {
1.358 + if (!mListener)
1.359 + return NS_ERROR_FAILURE;
1.360 +
1.361 + int32_t numItems = 0;
1.362 +
1.363 + while(true) {
1.364 + ++numItems;
1.365 +
1.366 + int32_t eol = mBuf.FindCharInSet("\n\r", mLineStart);
1.367 + if (eol < 0) break;
1.368 + mBuf.SetCharAt(char16_t('\0'), eol);
1.369 +
1.370 + const char *line = mBuf.get() + mLineStart;
1.371 +
1.372 + int32_t lineLen = eol - mLineStart;
1.373 + mLineStart = eol + 1;
1.374 +
1.375 + if (lineLen >= 4) {
1.376 + nsresult rv;
1.377 + const char *buf = line;
1.378 +
1.379 + if (buf[0] == '1') {
1.380 + if (buf[1] == '0') {
1.381 + if (buf[2] == '0' && buf[3] == ':') {
1.382 + // 100. Human-readable comment line. Ignore
1.383 + } else if (buf[2] == '1' && buf[3] == ':') {
1.384 + // 101. Human-readable information line.
1.385 + mComment.Append(buf + 4);
1.386 +
1.387 + char *value = ((char *)buf) + 4;
1.388 + nsUnescape(value);
1.389 + mListener->OnInformationAvailable(aRequest, aCtxt, NS_ConvertUTF8toUTF16(value));
1.390 +
1.391 + } else if (buf[2] == '2' && buf[3] == ':') {
1.392 + // 102. Human-readable information line, HTML.
1.393 + mComment.Append(buf + 4);
1.394 + }
1.395 + }
1.396 + } else if (buf[0] == '2') {
1.397 + if (buf[1] == '0') {
1.398 + if (buf[2] == '0' && buf[3] == ':') {
1.399 + // 200. Define field names
1.400 + rv = ParseFormat(buf + 4);
1.401 + if (NS_FAILED(rv)) {
1.402 + return rv;
1.403 + }
1.404 + } else if (buf[2] == '1' && buf[3] == ':') {
1.405 + // 201. Field data
1.406 + nsCOMPtr<nsIDirIndex> idx = do_CreateInstance("@mozilla.org/dirIndex;1",&rv);
1.407 + if (NS_FAILED(rv))
1.408 + return rv;
1.409 +
1.410 + rv = ParseData(idx, ((char *)buf) + 4);
1.411 + if (NS_FAILED(rv)) {
1.412 + return rv;
1.413 + }
1.414 +
1.415 + mListener->OnIndexAvailable(aRequest, aCtxt, idx);
1.416 + }
1.417 + }
1.418 + } else if (buf[0] == '3') {
1.419 + if (buf[1] == '0') {
1.420 + if (buf[2] == '0' && buf[3] == ':') {
1.421 + // 300. Self-referring URL
1.422 + } else if (buf[2] == '1' && buf[3] == ':') {
1.423 + // 301. OUR EXTENSION - encoding
1.424 + int i = 4;
1.425 + while (buf[i] && nsCRT::IsAsciiSpace(buf[i]))
1.426 + ++i;
1.427 +
1.428 + if (buf[i])
1.429 + SetEncoding(buf+i);
1.430 + }
1.431 + }
1.432 + }
1.433 + }
1.434 + }
1.435 +
1.436 + return NS_OK;
1.437 +}
