]> git.ipfire.org Git - thirdparty/squid.git/blob - src/http/one/Parser.h
SourceFormat Enforcement
[thirdparty/squid.git] / src / http / one / Parser.h
1 /*
2 * Copyright (C) 1996-2017 The Squid Software Foundation and contributors
3 *
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.
7 */
8
9 #ifndef _SQUID_SRC_HTTP_ONE_PARSER_H
10 #define _SQUID_SRC_HTTP_ONE_PARSER_H
11
12 #include "anyp/ProtocolVersion.h"
13 #include "http/one/forward.h"
14 #include "http/StatusCode.h"
15 #include "sbuf/SBuf.h"
16
17 namespace Http {
18 namespace One {
19
20 // Parser states
21 enum ParseState {
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
29 };
30
31 /** HTTP/1.x protocol parser
32 *
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:
35 *
36 * \item first-line (request-line / simple-request / status-line)
37 * \item mime-header 0*( header-name ':' SP field-value CRLF)
38 */
39 class Parser : public RefCountable
40 {
41 public:
42 typedef SBuf::size_type size_type;
43
44 Parser() : parseStatusCode(Http::scNone), parsingStage_(HTTP_PARSE_NONE), hackExpectsMime_(false) {}
45 virtual ~Parser() {}
46
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;
50
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;
55
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.
59 */
60 bool needsMoreData() const {return parsingStage_!=HTTP_PARSE_DONE;}
61
62 /// size in bytes of the first line including CRLF terminator
63 virtual size_type firstLineSize() const = 0;
64
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();}
68
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();}
73
74 /// buffer containing HTTP mime headers, excluding message first-line.
75 SBuf mimeHeader() const {return mimeHeaderBlock_;}
76
77 /// the protocol label for this message
78 const AnyP::ProtocolVersion & messageProtocol() const {return msgProtocol_;}
79
80 /**
81 * Scan the mime header block (badly) for a header with the given name.
82 *
83 * BUG: omits lines when searching for headers with obs-fold or multiple entries.
84 *
85 * BUG: limits output to just 1KB when Squid accepts up to 64KB line length.
86 *
87 * \return A pointer to a field-value of the first matching field-name, or NULL.
88 */
89 char *getHeaderField(const char *name);
90
91 /// the remaining unprocessed section of buffer
92 const SBuf &remaining() const {return buf_;}
93
94 #if USE_HTTP_VIOLATIONS
95 /// the right debugs() level for parsing HTTP violation messages
96 int violationLevel() const;
97 #endif
98
99 /**
100 * HTTP status code resulting from the parse process.
101 * to be used on the invalid message handling.
102 *
103 * Http::scNone indicates incomplete parse,
104 * Http::scOkay indicates no error,
105 * other codes represent a parse error.
106 */
107 Http::StatusCode parseStatusCode;
108
109 /// the characters which are to be considered valid whitespace
110 /// (WSP / BSP / OWS)
111 static const CharacterSet &DelimiterCharacters();
112
113 protected:
114 /**
115 * detect and skip the CRLF or (if tolerant) LF line terminator
116 * consume from the tokenizer.
117 *
118 * throws if non-terminator is detected.
119 * \retval true only if line terminator found.
120 * \retval false incomplete or missing line terminator, need more data.
121 */
122 bool skipLineTerminator(Http1::Tokenizer &tok) const;
123
124 /**
125 * Scan to find the mime headers block for current message.
126 *
127 * \retval true If mime block (or a blocks non-existence) has been
128 * identified accurately within limit characters.
129 * mimeHeaderBlock_ has been updated and buf_ consumed.
130 *
131 * \retval false An error occured, or no mime terminator found within limit.
132 */
133 bool grabMimeBlock(const char *which, const size_t limit);
134
135 /// RFC 7230 section 2.6 - 7 magic octets
136 static const SBuf Http1magic;
137
138 /// bytes remaining to be parsed
139 SBuf buf_;
140
141 /// what stage the parser is currently up to
142 ParseState parsingStage_;
143
144 /// what protocol label has been found in the first line (if any)
145 AnyP::ProtocolVersion msgProtocol_;
146
147 /// buffer holding the mime headers (if any)
148 SBuf mimeHeaderBlock_;
149
150 /// Whether the invalid HTTP as HTTP/0.9 hack expects a mime header block
151 bool hackExpectsMime_;
152
153 private:
154 void cleanMimePrefix();
155 void unfoldMime();
156 };
157
158 } // namespace One
159 } // namespace Http
160
161 #endif /* _SQUID_SRC_HTTP_ONE_PARSER_H */
162