2 * Copyright (C) 1996-2016 The Squid Software Foundation and contributors
4 * Squid software is distributed under GPLv2+ license and includes
5 * contributions from numerous individuals and organizations.
6 * Please see the COPYING and CONTRIBUTORS files for details.
9 #ifndef _SQUID_SRC_HTTP_ONE_PARSER_H
10 #define _SQUID_SRC_HTTP_ONE_PARSER_H
12 #include "anyp/ProtocolVersion.h"
13 #include "http/one/forward.h"
14 #include "http/StatusCode.h"
15 #include "sbuf/SBuf.h"
22 HTTP_PARSE_NONE
, ///< initialized, but nothing usefully parsed yet
23 HTTP_PARSE_FIRST
, ///< HTTP/1 message first-line
24 HTTP_PARSE_CHUNK_SZ
, ///< HTTP/1.1 chunked encoding chunk-size
25 HTTP_PARSE_CHUNK_EXT
, ///< HTTP/1.1 chunked encoding chunk-ext
26 HTTP_PARSE_CHUNK
, ///< HTTP/1.1 chunked encoding chunk-data
27 HTTP_PARSE_MIME
, ///< HTTP/1 mime-header block
28 HTTP_PARSE_DONE
///< parsed a message header, or reached a terminal syntax error
31 /** HTTP/1.x protocol parser
33 * Works on a raw character I/O buffer and tokenizes the content into
34 * the major CRLF delimited segments of an HTTP/1 procotol message:
36 * \item first-line (request-line / simple-request / status-line)
37 * \item mime-header 0*( header-name ':' SP field-value CRLF)
39 class Parser
: public RefCountable
42 typedef SBuf::size_type size_type
;
44 Parser() : parseStatusCode(Http::scNone
), parsingStage_(HTTP_PARSE_NONE
), hackExpectsMime_(false) {}
47 /// Set this parser back to a default state.
48 /// Will DROP any reference to a buffer (does not free).
49 virtual void clear() = 0;
51 /// attempt to parse a message from the buffer
52 /// \retval true if a full message was found and parsed
53 /// \retval false if incomplete, invalid or no message was found
54 virtual bool parse(const SBuf
&aBuf
) = 0;
56 /** Whether the parser is waiting on more data to complete parsing a message.
57 * Use to distinguish between incomplete data and error results
58 * when parse() returns false.
60 bool needsMoreData() const {return parsingStage_
!=HTTP_PARSE_DONE
;}
62 /// size in bytes of the first line including CRLF terminator
63 virtual size_type
firstLineSize() const = 0;
65 /// size in bytes of the message headers including CRLF terminator(s)
66 /// but excluding first-line bytes
67 size_type
headerBlockSize() const {return mimeHeaderBlock_
.length();}
69 /// size in bytes of HTTP message block, includes first-line and mime headers
70 /// excludes any body/entity/payload bytes
71 /// excludes any garbage prefix before the first-line
72 size_type
messageHeaderSize() const {return firstLineSize() + headerBlockSize();}
74 /// buffer containing HTTP mime headers, excluding message first-line.
75 SBuf
mimeHeader() const {return mimeHeaderBlock_
;}
77 /// the protocol label for this message
78 const AnyP::ProtocolVersion
& messageProtocol() const {return msgProtocol_
;}
81 * Scan the mime header block (badly) for a header with the given name.
83 * BUG: omits lines when searching for headers with obs-fold or multiple entries.
85 * BUG: limits output to just 1KB when Squid accepts up to 64KB line length.
87 * \return A pointer to a field-value of the first matching field-name, or NULL.
89 char *getHeaderField(const char *name
);
91 /// the remaining unprocessed section of buffer
92 const SBuf
&remaining() const {return buf_
;}
95 * HTTP status code resulting from the parse process.
96 * to be used on the invalid message handling.
98 * Http::scNone indicates incomplete parse,
99 * Http::scOkay indicates no error,
100 * other codes represent a parse error.
102 Http::StatusCode parseStatusCode
;
105 /// detect and skip the CRLF or (if tolerant) LF line terminator
106 /// consume from the tokenizer and return true only if found
107 bool skipLineTerminator(Http1::Tokenizer
&tok
) const;
110 * Scan to find the mime headers block for current message.
112 * \retval true If mime block (or a blocks non-existence) has been
113 * identified accurately within limit characters.
114 * mimeHeaderBlock_ has been updated and buf_ consumed.
116 * \retval false An error occured, or no mime terminator found within limit.
118 bool grabMimeBlock(const char *which
, const size_t limit
);
120 /// RFC 7230 section 2.6 - 7 magic octets
121 static const SBuf Http1magic
;
123 /// bytes remaining to be parsed
126 /// what stage the parser is currently up to
127 ParseState parsingStage_
;
129 /// what protocol label has been found in the first line (if any)
130 AnyP::ProtocolVersion msgProtocol_
;
132 /// buffer holding the mime headers (if any)
133 SBuf mimeHeaderBlock_
;
135 /// Whether the invalid HTTP as HTTP/0.9 hack expects a mime header block
136 bool hackExpectsMime_
;
142 #endif /* _SQUID_SRC_HTTP_ONE_PARSER_H */