summaryrefslogtreecommitdiffstats
path: root/net
diff options
context:
space:
mode:
Diffstat (limited to 'net')
-rw-r--r--net/base/net_util.cc38
-rw-r--r--net/base/net_util.h31
-rw-r--r--net/base/net_util_unittest.cc188
-rw-r--r--net/url_request/url_request_context.h11
4 files changed, 174 insertions, 94 deletions
diff --git a/net/base/net_util.cc b/net/base/net_util.cc
index 9610e40..e5c53eb 100644
--- a/net/base/net_util.cc
+++ b/net/base/net_util.cc
@@ -247,17 +247,22 @@ bool DecodeBQEncoding(const std::string& part, RFC2047EncodingType enc_type,
}
bool DecodeWord(const std::string& encoded_word,
+ const std::string& referrer_charset,
bool *is_rfc2047,
std::string* output) {
- // TODO(jungshik) : Revisit this later. Do we want to pass through non-ASCII
- // strings which can be mozibake? WinHTTP converts a raw 8bit string
- // UTF-16 assuming it's in the OS default encoding.
if (!IsStringASCII(encoded_word)) {
- // Try falling back to the NativeMB encoding if the raw input is not UTF-8.
+ // Try UTF-8, referrer_charset and the native OS default charset in turn.
if (IsStringUTF8(encoded_word)) {
*output = encoded_word;
} else {
- *output = WideToUTF8(base::SysNativeMBToWide(encoded_word));
+ std::wstring wide_output;
+ if (!referrer_charset.empty() &&
+ CodepageToWide(encoded_word, referrer_charset.c_str(),
+ OnStringUtilConversionError::FAIL, &wide_output)) {
+ *output = WideToUTF8(wide_output);
+ } else {
+ *output = WideToUTF8(base::SysNativeMBToWide(encoded_word));
+ }
}
*is_rfc2047 = false;
return true;
@@ -357,7 +362,9 @@ bool DecodeWord(const std::string& encoded_word,
return false;
}
-bool DecodeParamValue(const std::string& input, std::string* output) {
+bool DecodeParamValue(const std::string& input,
+ const std::string& referrer_charset,
+ std::string* output) {
std::string tmp;
// Tokenize with whitespace characters.
StringTokenizer t(input, " \t\n\r");
@@ -378,7 +385,8 @@ bool DecodeParamValue(const std::string& input, std::string* output) {
// in a single encoded-word. Firefox/Thunderbird do not support
// it, either.
std::string decoded;
- if (!DecodeWord(t.token(), &is_previous_token_rfc2047, &decoded))
+ if (!DecodeWord(t.token(), referrer_charset, &is_previous_token_rfc2047,
+ &decoded))
return false;
tmp.append(decoded);
}
@@ -683,7 +691,8 @@ std::string GetSpecificHeader(const std::string& headers,
return GetSpecificHeaderT(headers, name);
}
-std::wstring GetFileNameFromCD(const std::string& header) {
+std::wstring GetFileNameFromCD(const std::string& header,
+ const std::string& referrer_charset) {
std::string param_value = GetHeaderParamValue(header, "filename");
if (param_value.empty()) {
// Some servers use 'name' parameter.
@@ -692,7 +701,7 @@ std::wstring GetFileNameFromCD(const std::string& header) {
if (param_value.empty())
return std::wstring();
std::string decoded;
- if (DecodeParamValue(param_value, &decoded))
+ if (DecodeParamValue(param_value, referrer_charset, &decoded))
return UTF8ToWide(decoded);
return std::wstring();
}
@@ -863,8 +872,10 @@ std::wstring StripWWW(const std::wstring& text) {
std::wstring GetSuggestedFilename(const GURL& url,
const std::string& content_disposition,
+ const std::string& referrer_charset,
const std::wstring& default_name) {
- std::wstring filename = GetFileNameFromCD(content_disposition);
+ std::wstring filename = GetFileNameFromCD(content_disposition,
+ referrer_charset);
if (!filename.empty()) {
// Remove any path information the server may have sent, take the name
// only.
@@ -901,13 +912,6 @@ std::wstring GetSuggestedFilename(const GURL& url,
return filename;
}
-std::wstring GetSuggestedFilename(const GURL& url,
- const std::wstring& content_disposition,
- const std::wstring& default_name) {
- return GetSuggestedFilename(
- url, WideToUTF8(content_disposition), default_name);
-}
-
bool IsPortAllowedByDefault(int port) {
int array_size = arraysize(kRestrictedPorts);
for (int i = 0; i < array_size; i++) {
diff --git a/net/base/net_util.h b/net/base/net_util.h
index 2ad81bf..9aba789 100644
--- a/net/base/net_util.h
+++ b/net/base/net_util.h
@@ -71,12 +71,20 @@ std::wstring GetHeaderParamValue(const std::wstring& field,
std::string GetHeaderParamValue(const std::string& field,
const std::string& param_name);
-// Return the filename extracted from Content-Disposition header. Only two
-// formats are supported: a. %-escaped UTF-8 b. RFC 2047.
+// Return the filename extracted from Content-Disposition header. The following
+// formats are tried in order listed below:
//
-// A non-ASCII param value is just returned as it is (assuming a NativeMB
-// encoding). When a param value is ASCII, but is not in one of two forms
-// supported, it is returned as it is unless it's pretty close to two supported
+// 1. RFC 2047
+// 2. Raw-8bit-characters :
+// a. UTF-8, b. referrer_charset, c. default os codepage.
+// 3. %-escaped UTF-8.
+//
+// In step 2, if referrer_charset is empty(i.e. unknown), 2b is skipped.
+// In step 3, the fallback charsets tried in step 2 are not tried. We
+// can consider doing that later.
+//
+// When a param value is ASCII, but is not in format #1 or format #3 above,
+// it is returned as it is unless it's pretty close to two supported
// formats but not well-formed. In that case, an empty string is returned.
//
// In any case, a caller must check for the empty return value and resort to
@@ -90,7 +98,8 @@ std::string GetHeaderParamValue(const std::string& field,
// other caller is a unit test. Need to figure out expose this function only to
// net_util_unittest.
//
-std::wstring GetFileNameFromCD(const std::string& header);
+std::wstring GetFileNameFromCD(const std::string& header,
+ const std::string& referrer_charset);
// Converts the given host name to unicode characters, APPENDING them to the
// the given output string. This can be called for any host name, if the
@@ -133,14 +142,12 @@ std::wstring StripWWW(const std::wstring& text);
// Gets the filename from the raw Content-Disposition header (as read from the
// network). Otherwise uses the last path component name or hostname from
// |url|. Note: it's possible for the suggested filename to be empty (e.g.,
-// file:/// or view-cache:).
+// file:/// or view-cache:). referrer_charset is used as one of charsets
+// to interpret a raw 8bit string in C-D header (after interpreting
+// as UTF-8 fails). See the comment for GetFilenameFromCD for more details.
std::wstring GetSuggestedFilename(const GURL& url,
const std::string& content_disposition,
- const std::wstring& default_name);
-
-// DEPRECATED: Please use the above version of this method.
-std::wstring GetSuggestedFilename(const GURL& url,
- const std::wstring& content_disposition,
+ const std::string& referrer_charset,
const std::wstring& default_name);
// Checks the given port against a list of ports which are restricted by
diff --git a/net/base/net_util_unittest.cc b/net/base/net_util_unittest.cc
index f9bc7f7..1a3bcdc 100644
--- a/net/base/net_util_unittest.cc
+++ b/net/base/net_util_unittest.cc
@@ -39,6 +39,7 @@ struct HeaderParamCase {
struct FileNameCDCase {
const char* header_field;
+ const char* referrer_charset;
const wchar_t* expected;
};
@@ -58,7 +59,8 @@ struct IDNTestCase {
struct SuggestedFilenameCase {
const char* url;
- const wchar_t* content_disp_header;
+ const char* content_disp_header;
+ const char* referrer_charset;
const wchar_t* default_filename;
const wchar_t* expected_filename;
};
@@ -299,75 +301,96 @@ TEST(NetUtilTest, GetHeaderParamValue) {
TEST(NetUtilTest, GetFileNameFromCD) {
const FileNameCDCase tests[] = {
// Test various forms of C-D header fields emitted by web servers.
- {"content-disposition: inline; filename=\"abcde.pdf\"", L"abcde.pdf"},
- {"content-disposition: inline; name=\"abcde.pdf\"", L"abcde.pdf"},
- {"content-disposition: attachment; filename=abcde.pdf", L"abcde.pdf"},
- {"content-disposition: attachment; name=abcde.pdf", L"abcde.pdf"},
- {"content-disposition: attachment; filename=abc,de.pdf", L"abc,de.pdf"},
- {"content-disposition: filename=abcde.pdf", L"abcde.pdf"},
- {"content-disposition: filename= abcde.pdf", L"abcde.pdf"},
- {"content-disposition: filename =abcde.pdf", L"abcde.pdf"},
- {"content-disposition: filename = abcde.pdf", L"abcde.pdf"},
- {"content-disposition: filename\t=abcde.pdf", L"abcde.pdf"},
- {"content-disposition: filename \t\t =abcde.pdf", L"abcde.pdf"},
- {"content-disposition: name=abcde.pdf", L"abcde.pdf"},
- {"content-disposition: inline; filename=\"abc%20de.pdf\"", L"abc de.pdf"},
+ {"content-disposition: inline; filename=\"abcde.pdf\"", "", L"abcde.pdf"},
+ {"content-disposition: inline; name=\"abcde.pdf\"", "", L"abcde.pdf"},
+ {"content-disposition: attachment; filename=abcde.pdf", "", L"abcde.pdf"},
+ {"content-disposition: attachment; name=abcde.pdf", "", L"abcde.pdf"},
+ {"content-disposition: attachment; filename=abc,de.pdf", "", L"abc,de.pdf"},
+ {"content-disposition: filename=abcde.pdf", "", L"abcde.pdf"},
+ {"content-disposition: filename= abcde.pdf", "", L"abcde.pdf"},
+ {"content-disposition: filename =abcde.pdf", "", L"abcde.pdf"},
+ {"content-disposition: filename = abcde.pdf", "", L"abcde.pdf"},
+ {"content-disposition: filename\t=abcde.pdf", "", L"abcde.pdf"},
+ {"content-disposition: filename \t\t =abcde.pdf", "", L"abcde.pdf"},
+ {"content-disposition: name=abcde.pdf", "", L"abcde.pdf"},
+ {"content-disposition: inline; filename=\"abc%20de.pdf\"", "",
+ L"abc de.pdf"},
// Whitespaces are converted to a space.
- {"content-disposition: inline; filename=\"abc \t\nde.pdf\"",
+ {"content-disposition: inline; filename=\"abc \t\nde.pdf\"", "",
L"abc de.pdf"},
// %-escaped UTF-8
{"Content-Disposition: attachment; filename=\"%EC%98%88%EC%88%A0%20"
- "%EC%98%88%EC%88%A0.jpg\"", L"\xc608\xc220 \xc608\xc220.jpg"},
+ "%EC%98%88%EC%88%A0.jpg\"", "", L"\xc608\xc220 \xc608\xc220.jpg"},
{"Content-Disposition: attachment; filename=\"%F0%90%8C%B0%F0%90%8C%B1"
- "abc.jpg\"", L"\U00010330\U00010331abc.jpg"},
+ "abc.jpg\"", "", L"\U00010330\U00010331abc.jpg"},
{"Content-Disposition: attachment; filename=\"%EC%98%88%EC%88%A0 \n"
- "%EC%98%88%EC%88%A0.jpg\"", L"\xc608\xc220 \xc608\xc220.jpg"},
+ "%EC%98%88%EC%88%A0.jpg\"", "", L"\xc608\xc220 \xc608\xc220.jpg"},
// RFC 2047 with various charsets and Q/B encodings
{"Content-Disposition: attachment; filename=\"=?EUC-JP?Q?=B7=DD=BD="
- "D13=2Epng?=\"", L"\x82b8\x8853" L"3.png"},
+ "D13=2Epng?=\"", "", L"\x82b8\x8853" L"3.png"},
{"Content-Disposition: attachment; filename==?eUc-Kr?b?v7m8+iAzLnBuZw==?=",
- L"\xc608\xc220 3.png"},
+ "", L"\xc608\xc220 3.png"},
{"Content-Disposition: attachment; filename==?utf-8?Q?=E8=8A=B8=E8"
- "=A1=93_3=2Epng?=", L"\x82b8\x8853 3.png"},
+ "=A1=93_3=2Epng?=", "", L"\x82b8\x8853 3.png"},
{"Content-Disposition: attachment; filename==?utf-8?Q?=F0=90=8C=B0"
- "_3=2Epng?=", L"\U00010330 3.png"},
- {"Content-Disposition: inline; filename=\"=?iso88591?Q?caf=e3_=2epng?=\"",
- L"caf\x00e3 .png"},
+ "_3=2Epng?=", "", L"\U00010330 3.png"},
+ {"Content-Disposition: inline; filename=\"=?iso88591?Q?caf=e9_=2epng?=\"",
+ "", L"caf\x00e9 .png"},
// Space after an encode word should be removed.
- {"Content-Disposition: inline; filename=\"=?iso88591?Q?caf=E3_?= .png\"",
- L"caf\x00e3 .png"},
+ {"Content-Disposition: inline; filename=\"=?iso88591?Q?caf=E9_?= .png\"",
+ "", L"caf\x00e9 .png"},
// Two encoded words with different charsets (not very likely to be emitted
// by web servers in the wild). Spaces between them are removed.
{"Content-Disposition: inline; filename=\"=?euc-kr?b?v7m8+iAz?="
- " =?ksc5601?q?=BF=B9=BC=FA=2Epng?=\"", L"\xc608\xc220 3\xc608\xc220.png"},
- {"Content-Disposition: attachment; filename=\"=?windows-1252?Q?caf=E3?="
- " =?iso-8859-7?b?4eI=?= .png\"", L"caf\x00e3\x03b1\x03b2.png"},
- // Non-ASCII string is passed through (and treated as UTF-8).
- {"Content-Disposition: attachment; filename=caf\xc3\xa3.png",
- L"caf\x00e3.png"},
+ " =?ksc5601?q?=BF=B9=BC=FA=2Epng?=\"", "",
+ L"\xc608\xc220 3\xc608\xc220.png"},
+ {"Content-Disposition: attachment; filename=\"=?windows-1252?Q?caf=E9?="
+ " =?iso-8859-7?b?4eI=?= .png\"", "", L"caf\x00e9\x03b1\x03b2.png"},
+ // Non-ASCII string is passed through and treated as UTF-8 as long as
+ // it's valid as UTF-8 and regardless of |referrer_charset|.
+ {"Content-Disposition: attachment; filename=caf\xc3\xa9.png",
+ "iso-8859-1", L"caf\x00e9.png"},
+ {"Content-Disposition: attachment; filename=caf\xc3\xa9.png",
+ "", L"caf\x00e9.png"},
+ // Non-ASCII/Non-UTF-8 string. Fall back to the referrer charset.
+ {"Content-Disposition: attachment; filename=caf\xe5.png",
+ "windows-1253", L"caf\x03b5.png"},
+#if 0
+ // Non-ASCII/Non-UTF-8 string. Fall back to the native codepage.
+ // TODO(jungshik): We need to set the OS default codepage
+ // to a specific value before testing. On Windows, we can use
+ // SetThreadLocale().
+ {"Content-Disposition: attachment; filename=\xb0\xa1\xb0\xa2.png",
+ "", L"\xac00\xac01.png"},
+#endif
// Failure cases
// Invalid hex-digit "G"
- {"Content-Disposition: attachment; filename==?iiso88591?Q?caf=EG?=", L""},
+ {"Content-Disposition: attachment; filename==?iiso88591?Q?caf=EG?=", "",
+ L""},
// Incomplete RFC 2047 encoded-word (missing '='' at the end)
- {"Content-Disposition: attachment; filename==?iso88591?Q?caf=E3?", L""},
+ {"Content-Disposition: attachment; filename==?iso88591?Q?caf=E3?", "", L""},
// Extra character at the end of an encoded word
- {"Content-Disposition: attachment; filename==?iso88591?Q?caf=E3?==", L""},
+ {"Content-Disposition: attachment; filename==?iso88591?Q?caf=E3?==",
+ "", L""},
// Extra token at the end of an encoded word
- {"Content-Disposition: attachment; filename==?iso88591?Q?caf=E3?=?", L""},
- {"Content-Disposition: attachment; filename==?iso88591?Q?caf=E3?=?=", L""},
+ {"Content-Disposition: attachment; filename==?iso88591?Q?caf=E3?=?",
+ "", L""},
+ {"Content-Disposition: attachment; filename==?iso88591?Q?caf=E3?=?=",
+ "", L""},
// Incomplete hex-escaped chars
{"Content-Disposition: attachment; filename==?windows-1252?Q?=63=61=E?=",
- L""},
- {"Content-Disposition: attachment; filename=%EC%98%88%EC%88%A", L""},
+ "", L""},
+ {"Content-Disposition: attachment; filename=%EC%98%88%EC%88%A", "", L""},
// %-escaped non-UTF-8 encoding is an "error"
- {"Content-Disposition: attachment; filename=%B7%DD%BD%D1.png", L""},
+ {"Content-Disposition: attachment; filename=%B7%DD%BD%D1.png", "", L""},
// Two RFC 2047 encoded words in a row without a space is an error.
{"Content-Disposition: attachment; filename==?windows-1252?Q?caf=E3?="
- "=?iso-8859-7?b?4eIucG5nCg==?=", L""},
+ "=?iso-8859-7?b?4eIucG5nCg==?=", "", L""},
};
for (size_t i = 0; i < ARRAYSIZE_UNSAFE(tests); ++i) {
EXPECT_EQ(tests[i].expected,
- net::GetFileNameFromCD(tests[i].header_field));
+ net::GetFileNameFromCD(tests[i].header_field,
+ tests[i].referrer_charset));
}
}
@@ -669,97 +692,132 @@ TEST(NetUtilTest, StripWWW) {
TEST(NetUtilTest, GetSuggestedFilename) {
const SuggestedFilenameCase test_cases[] = {
{"http://www.google.com/",
- L"Content-disposition: attachment; filename=test.html",
+ "Content-disposition: attachment; filename=test.html",
+ "",
L"",
L"test.html"},
{"http://www.google.com/",
- L"Content-disposition: attachment; filename=\"test.html\"",
+ "Content-disposition: attachment; filename=\"test.html\"",
+ "",
L"",
L"test.html"},
{"http://www.google.com/path/test.html",
- L"Content-disposition: attachment",
+ "Content-disposition: attachment",
+ "",
L"",
L"test.html"},
{"http://www.google.com/path/test.html",
- L"Content-disposition: attachment;",
+ "Content-disposition: attachment;",
+ "",
L"",
L"test.html"},
{"http://www.google.com/",
- L"",
+ "",
+ "",
L"",
L"www.google.com"},
{"http://www.google.com/test.html",
- L"",
+ "",
+ "",
L"",
L"test.html"},
// Now that we use googleurl's ExtractFileName, this case falls back
// to the hostname. If this behavior is not desirable, we'd better
// change ExtractFileName (in url_parse).
{"http://www.google.com/path/",
- L"",
+ "",
+ "",
L"",
L"www.google.com"},
{"http://www.google.com/path",
- L"",
+ "",
+ "",
L"",
L"path"},
{"file:///",
- L"",
+ "",
+ "",
L"",
L"download"},
{"view-cache:",
- L"",
+ "",
+ "",
L"",
L"download"},
{"http://www.google.com/",
- L"Content-disposition: attachment; filename =\"test.html\"",
+ "Content-disposition: attachment; filename =\"test.html\"",
+ "",
L"download",
L"test.html"},
{"http://www.google.com/",
- L"",
+ "",
+ "",
L"download",
L"download"},
{"http://www.google.com/",
- L"Content-disposition: attachment; filename=\"../test.html\"",
+ "Content-disposition: attachment; filename=\"../test.html\"",
+ "",
L"",
L"test.html"},
{"http://www.google.com/",
- L"Content-disposition: attachment; filename=\"..\"",
+ "Content-disposition: attachment; filename=\"..\"",
+ "",
L"download",
L"download"},
{"http://www.google.com/test.html",
- L"Content-disposition: attachment; filename=\"..\"",
+ "Content-disposition: attachment; filename=\"..\"",
+ "",
L"download",
L"test.html"},
// Below is a small subset of cases taken from GetFileNameFromCD test above.
{"http://www.google.com/",
- L"Content-Disposition: attachment; filename=\"%EC%98%88%EC%88%A0%20"
- L"%EC%98%88%EC%88%A0.jpg\"",
+ "Content-Disposition: attachment; filename=\"%EC%98%88%EC%88%A0%20"
+ "%EC%98%88%EC%88%A0.jpg\"",
+ "",
L"",
L"\uc608\uc220 \uc608\uc220.jpg"},
{"http://www.google.com/%EC%98%88%EC%88%A0%20%EC%98%88%EC%88%A0.jpg",
- L"",
+ "",
+ "",
L"download",
L"\uc608\uc220 \uc608\uc220.jpg"},
{"http://www.google.com/",
- L"Content-disposition: attachment;",
+ "Content-disposition: attachment;",
+ "",
L"\uB2E4\uC6B4\uB85C\uB4DC",
L"\uB2E4\uC6B4\uB85C\uB4DC"},
{"http://www.google.com/",
- L"Content-Disposition: attachment; filename=\"=?EUC-JP?Q?=B7=DD=BD="
- L"D13=2Epng?=\"",
+ "Content-Disposition: attachment; filename=\"=?EUC-JP?Q?=B7=DD=BD="
+ "D13=2Epng?=\"",
+ "",
L"download",
L"\u82b8\u88533.png"},
+ {"http://www.example.com/images?id=3",
+ "Content-Disposition: attachment; filename=caf\xc3\xa9.png",
+ "iso-8859-1",
+ L"",
+ L"caf\u00e9.png"},
+ {"http://www.example.com/images?id=3",
+ "Content-Disposition: attachment; filename=caf\xe5.png",
+ "windows-1253",
+ L"",
+ L"caf\u03b5.png"},
+ {"http://www.example.com/file?id=3",
+ "Content-Disposition: attachment; name=\xcf\xc2\xd4\xd8.zip",
+ "GBK",
+ L"",
+ L"\u4e0b\u8f7d.zip"},
// Invalid C-D header. Extracts filename from url.
{"http://www.google.com/test.html",
- L"Content-Disposition: attachment; filename==?iiso88591?Q?caf=EG?=",
+ "Content-Disposition: attachment; filename==?iiso88591?Q?caf=EG?=",
+ "",
L"",
L"test.html"},
};
for (size_t i = 0; i < ARRAYSIZE_UNSAFE(test_cases); ++i) {
std::wstring filename = net::GetSuggestedFilename(
GURL(test_cases[i].url), test_cases[i].content_disp_header,
- test_cases[i].default_filename);
+ test_cases[i].referrer_charset, test_cases[i].default_filename);
EXPECT_EQ(std::wstring(test_cases[i].expected_filename), filename);
}
}
diff --git a/net/url_request/url_request_context.h b/net/url_request/url_request_context.h
index 8c09c90..29735ac 100644
--- a/net/url_request/url_request_context.h
+++ b/net/url_request/url_request_context.h
@@ -70,6 +70,13 @@ class URLRequestContext :
return EmptyString();
}
+ // In general, referrer_charset is not known when URLRequestContext is
+ // constructed. So, we need a setter.
+ const std::string& referrer_charset() const { return referrer_charset_; }
+ void set_referrer_charset(const std::string& charset) {
+ referrer_charset_ = charset;
+ }
+
protected:
friend class base::RefCountedThreadSafe<URLRequestContext>;
@@ -85,6 +92,10 @@ class URLRequestContext :
net::FtpAuthCache ftp_auth_cache_;
std::string accept_language_;
std::string accept_charset_;
+ // The charset of the referrer where this request comes from. It's not
+ // used in communication with a server but is used to construct a suggested
+ // filename for file download.
+ std::string referrer_charset_;
private:
DISALLOW_COPY_AND_ASSIGN(URLRequestContext);
generated by cgit v1.1 at 2025-01-20 11:12:07 +0000