+/*
+ * Copyright (C) 1996-2023 The Squid Software Foundation and contributors
+ *
+ * Squid software is distributed under GPLv2+ license and includes
+ * contributions from numerous individuals and organizations.
+ * Please see the COPYING and CONTRIBUTORS files for details.
+ */
+
#include "squid.h"
-#include "Debug.h"
+#include "base/Raw.h"
+#include "debug/Stream.h"
#include "http/one/ResponseParser.h"
#include "http/ProtocolVersion.h"
#include "parser/Tokenizer.h"
-#include "profiler/Profiler.h"
+#include "sbuf/Stream.h"
#include "SquidConfig.h"
const SBuf Http::One::ResponseParser::IcyMagic("ICY ");
// NP: we found the protocol version and consumed it already.
// just need the status code and reason phrase
-const int
-Http::One::ResponseParser::parseResponseStatusAndReason(::Parser::Tokenizer &tok, const CharacterSet &WspDelim)
+int
+Http::One::ResponseParser::parseResponseStatusAndReason(Tokenizer &tok)
{
- if (!completedStatus_) {
- debugs(74, 9, "seek status-code in: " << tok.remaining().substr(0,10) << "...");
- /* RFC 7230 section 3.1.2 - status code is 3 DIGIT octets.
- * There is no limit on what those octets may be.
- * 000 through 999 are all valid.
- */
- int64_t statusValue;
- if (tok.int64(statusValue, 10, false, 3) && tok.skipOne(WspDelim)) {
-
- debugs(74, 6, "found int64 status-code=" << statusValue);
- statusCode_ = static_cast<Http::StatusCode>(statusValue);
-
+ try {
+ if (!completedStatus_) {
+ debugs(74, 9, "seek status-code in: " << tok.remaining().substr(0,10) << "...");
+ ParseResponseStatus(tok, statusCode_);
buf_ = tok.remaining(); // resume checkpoint
completedStatus_ = true;
-
- } else if (tok.atEnd()) {
- debugs(74, 6, "Parser needs more data");
- return 0; // need more to be sure we have it all
-
- } else {
- debugs(74, 6, "invalid status-line. invalid code.");
- return -1; // invalid status, a single SP terminator required
}
// NOTE: any whitespace after the single SP is part of the reason phrase.
- }
-
- if (tok.atEnd())
- return 0; // need more to be sure we have it all
- /* RFC 7230 says we SHOULD ignore the reason phrase content
- * but it has a definite valid vs invalid character set.
- * We interpret the SHOULD as ignoring absence and syntax, but
- * producing an error if it contains an invalid octet.
- */
-
- debugs(74, 9, "seek reason-phrase in: " << tok.remaining().substr(0,50) << "...");
+ /* RFC 7230 says we SHOULD ignore the reason phrase content
+ * but it has a definite valid vs invalid character set.
+ * We interpret the SHOULD as ignoring absence and syntax, but
+ * producing an error if it contains an invalid octet.
+ */
- // if we got here we are still looking for reason-phrase bytes
- static const CharacterSet phraseChars = CharacterSet::WSP + CharacterSet::VCHAR + CharacterSet::OBSTEXT;
- (void)tok.prefix(reasonPhrase_, phraseChars); // optional, no error if missing
- if (skipLineTerminator(tok)) {
- debugs(74, DBG_DATA, "parse remaining buf={length=" << tok.remaining().length() << ", data='" << tok.remaining() << "'}");
+ debugs(74, 9, "seek reason-phrase in: " << tok.remaining().substr(0,50) << "...");
+ // if we got here we are still looking for reason-phrase bytes
+ static const CharacterSet phraseChars = CharacterSet::WSP + CharacterSet::VCHAR + CharacterSet::OBSTEXT;
+ (void)tok.prefix(reasonPhrase_, phraseChars); // optional, no error if missing
+ skipLineTerminator(tok);
buf_ = tok.remaining(); // resume checkpoint
+ debugs(74, DBG_DATA, Raw("leftovers", buf_.rawContent(), buf_.length()));
return 1;
- }
- reasonPhrase_.clear();
-
- if (tok.atEnd())
+ } catch (const InsufficientInput &) {
+ reasonPhrase_.clear();
return 0; // need more to be sure we have it all
-
- debugs(74, 6, "invalid status-line. garbage in reason phrase.");
+ } catch (const std::exception &ex) {
+ debugs(74, 6, "invalid status-line: " << ex.what());
+ }
return -1;
}
+void
+Http::One::ResponseParser::ParseResponseStatus(Tokenizer &tok, StatusCode &code)
+{
+ int64_t statusValue;
+ if (tok.int64(statusValue, 10, false, 3) && tok.skipOne(Parser::DelimiterCharacters())) {
+ debugs(74, 6, "raw status-code=" << statusValue);
+ code = static_cast<StatusCode>(statusValue); // may be invalid
+
+ // RFC 7230 Section 3.1.2 says status-code is exactly three DIGITs
+ if (code <= 99)
+ throw TextException(ToSBuf("status-code too short: ", code), Here());
+
+ // Codes with a non-standard first digit (a.k.a. response class) are
+ // considered semantically invalid per the following HTTP WG discussion:
+ // https://lists.w3.org/Archives/Public/ietf-http-wg/2010AprJun/0354.html
+ if (code >= 600)
+ throw TextException(ToSBuf("status-code from an invalid response class: ", code), Here());
+ } else if (tok.atEnd()) {
+ throw InsufficientInput();
+ } else {
+ throw TextException("syntactically invalid status-code area", Here());
+ }
+}
+
/**
* Attempt to parse the method field out of an HTTP message status-line.
*
* \retval 1 successful parse. statusCode_ and maybe reasonPhrase_ are filled and buffer consumed including first delimiter.
* \retval 0 more data is needed to complete the parse
*/
-const int
+int
Http::One::ResponseParser::parseResponseFirstLine()
{
- ::Parser::Tokenizer tok(buf_);
-
- CharacterSet WspDelim = CharacterSet::SP; // strict parse only accepts SP
-
- if (Config.onoff.relaxed_header_parser) {
- // RFC 7230 section 3.5
- // tolerant parser MAY accept any of SP, HTAB, VT (%x0B), FF (%x0C), or bare CR
- // as whitespace between status-line fields
- WspDelim += CharacterSet::HTAB
- + CharacterSet("VT,FF","\x0B\x0C")
- + CharacterSet::CR;
- }
+ Tokenizer tok(buf_);
if (msgProtocol_.protocol != AnyP::PROTO_NONE) {
debugs(74, 6, "continue incremental parse for " << msgProtocol_);
debugs(74, DBG_DATA, "parse remaining buf={length=" << tok.remaining().length() << ", data='" << tok.remaining() << "'}");
// we already found the magic, but not the full line. keep going.
- return parseResponseStatusAndReason(tok, WspDelim);
+ return parseResponseStatusAndReason(tok);
} else if (tok.skip(Http1magic)) {
debugs(74, 6, "found prefix magic " << Http1magic);
// magic contains major version, still need to find minor DIGIT
int64_t verMinor;
+ const auto &WspDelim = DelimiterCharacters();
if (tok.int64(verMinor, 10, false, 1) && tok.skipOne(WspDelim)) {
msgProtocol_.protocol = AnyP::PROTO_HTTP;
msgProtocol_.major = 1;
debugs(74, DBG_DATA, "parse remaining buf={length=" << tok.remaining().length() << ", data='" << tok.remaining() << "'}");
buf_ = tok.remaining(); // resume checkpoint
- return parseResponseStatusAndReason(tok, WspDelim);
+ return parseResponseStatusAndReason(tok);
} else if (tok.atEnd())
return 0; // need more to be sure we have it all
// NP: ICY has no /major.minor details
debugs(74, DBG_DATA, "parse remaining buf={length=" << tok.remaining().length() << ", data='" << tok.remaining() << "'}");
buf_ = tok.remaining(); // resume checkpoint
- return parseResponseStatusAndReason(tok, WspDelim);
-
- } else if (buf_.length() > Http1magic.length() && buf_.length() > IcyMagic.length()) {
+ return parseResponseStatusAndReason(tok);
+ } else if (buf_.length() < Http1magic.length() && Http1magic.startsWith(buf_)) {
+ debugs(74, 7, Raw("valid HTTP/1 prefix", buf_.rawContent(), buf_.length()));
+ return 0;
+ } else if (buf_.length() < IcyMagic.length() && IcyMagic.startsWith(buf_)) {
+ debugs(74, 7, Raw("valid ICY prefix", buf_.rawContent(), buf_.length()));
+ return 0;
+ } else {
debugs(74, 2, "unknown/missing prefix magic. Interpreting as HTTP/0.9");
// found something that looks like an HTTP/0.9 response
// Gateway/Transform it into HTTP/1.1
return 1; // no more parsing
}
- return 0; // need more to parse anything.
+ // unreachable
+ assert(false);
+ return -1;
}
bool
// stage 2: parse the status-line
if (parsingStage_ == HTTP_PARSE_FIRST) {
- PROF_start(HttpParserParseReplyLine);
-
const int retcode = parseResponseFirstLine();
// first-line (or a look-alike) found successfully.
- if (retcode > 0)
+ if (retcode > 0 && parsingStage_ == HTTP_PARSE_FIRST)
parsingStage_ = HTTP_PARSE_MIME;
debugs(74, 5, "status-line: retval " << retcode);
debugs(74, 5, "status-line: proto " << msgProtocol_);
debugs(74, 5, "status-line: status-code " << statusCode_);
debugs(74, 5, "status-line: reason-phrase " << reasonPhrase_);
debugs(74, 5, "Parser: bytes processed=" << (aBuf.length()-buf_.length()));
- PROF_stop(HttpParserParseReplyLine);
// syntax errors already
if (retcode < 0) {
parsingStage_ = HTTP_PARSE_DONE;
- statusCode_ = Http::scInvalidHeader;
+ parseStatusCode = Http::scInvalidHeader;
return false;
}
}
return !needsMoreData();
}
+
projects / thirdparty / squid.git / blobdiff