recoll / Code / Diff of /src/utils/mimeparse.h

Diff of /src/utils/mimeparse.h [211ea8] .. [385a7b]

Switch to unified view


...
 *   59 Temple Place - Suite 330, Boston, MA  02111-1307, USA.
 */
#ifndef _MIME_H_INCLUDED_
#define _MIME_H_INCLUDED_
/*
  Mime definitions RFC to 4-9-2006:

  2045 Multipurpose Internet Mail Extensions (MIME) Part One: Format of
  Internet Message Bodies. N. Freed, N. Borenstein. November 1996.
  (Format: TXT=72932 bytes) (Obsoletes RFC1521, RFC1522, RFC1590)
  (Updated by RFC2184, RFC2231) (Status: DRAFT STANDARD)

  2046 Multipurpose Internet Mail Extensions (MIME) Part Two: Media
  Types. N. Freed, N. Borenstein. November 1996. (Format: TXT=105854
  bytes) (Obsoletes RFC1521, RFC1522, RFC1590) (Updated by RFC2646,
  RFC3798) (Status: DRAFT STANDARD)

  2047 MIME (Multipurpose Internet Mail Extensions) Part Three: Message
  Header Extensions for Non-ASCII Text. K. Moore. November 1996.
  (Format: TXT=33262 bytes) (Obsoletes RFC1521, RFC1522, RFC1590)
  (Updated by RFC2184, RFC2231) (Status: DRAFT STANDARD)

  2183 Communicating Presentation Information in Internet Messages: The
  Content-Disposition Header Field. R. Troost, S. Dorner, K. Moore,
  Ed.. August 1997. (Format: TXT=23150 bytes) (Updates RFC1806)
  (Updated by RFC2184, RFC2231) (Status: PROPOSED STANDARD)

  2231 MIME Parameter Value and Encoded Word Extensions: Character Sets,
  Languages, and Continuations. N. Freed, K. Moore. November 1997.
  (Format: TXT=19280 bytes) (Obsoletes RFC2184) (Updates RFC2045,
  RFC2047, RFC2183) (Status: PROPOSED STANDARD)
*/


#include <time.h>

#include <string>
#include <map>

#include "base64.h"





/** A class to represent a MIME header value with parameters */
class MimeHeaderValue {
public:
    std::string value;
    std::map<std::string, std::string> params;
};

/** 
 * Parse MIME Content-type and Content-disposition value
 *
 * @param in the input string should be like: value; pn1=pv1; pn2=pv2. 
 *   Example: text/plain; charset="iso-8859-1" 
 */
extern bool parseMimeHeaderValue(const std::string& in, MimeHeaderValue& psd);

/** 
 * Quoted Printable decoding. 
 *
 * Doubles up as rfc2231 decoder, with the help of the hence the @param esc 
 * parameter.
 * RFC2045 Quoted Printable uses '=' , RFC2331 uses '%'. The two encodings are
 * otherwise similar.
 */
extern bool qp_decode(const std::string& in, std::string &out, char esc = '=');

/** Decode an Internet mail field value encoded according to rfc2047 
 *
 * Example input:  Some words =?iso-8859-1?Q?RE=A0=3A_Smoke_Tests?= more input
 * 
...
 * is sometimes used anyway...
 * 
 * @param in input string, ascii with rfc2047 markup
 * @return out output string encoded in utf-8
 */
extern bool rfc2047_decode(const std::string& in, std::string &out);


/** Decode RFC2822 date to unix time (gmt secs from 1970)
 *
 * @param dt date string (the part after Date: )
 * @return unix time
 */
time_t rfc2822DateToUxTime(const std::string& dt);

#endif /* _MIME_H_INCLUDED_ */

	a/src/utils/mimeparse.h		b/src/utils/mimeparse.h
	...		...
15	* 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.	15	* 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
16	*/	16	*/
17	#ifndef _MIME_H_INCLUDED_	17	#ifndef _MIME_H_INCLUDED_
18	#define _MIME_H_INCLUDED_	18	#define _MIME_H_INCLUDED_
19	/*	19	/*
20	Mime definitions RFC to 4-9-2006:	20	Mime definitions RFC to 4-9-2006:
21		21
22	2045 Multipurpose Internet Mail Extensions (MIME) Part One: Format of	22	2045 Multipurpose Internet Mail Extensions (MIME) Part One: Format of
23	Internet Message Bodies. N. Freed, N. Borenstein. November 1996.	23	Internet Message Bodies. N. Freed, N. Borenstein. November 1996.
24	(Format: TXT=72932 bytes) (Obsoletes RFC1521, RFC1522, RFC1590)	24	(Format: TXT=72932 bytes) (Obsoletes RFC1521, RFC1522, RFC1590)
25	(Updated by RFC2184, RFC2231) (Status: DRAFT STANDARD)	25	(Updated by RFC2184, RFC2231) (Status: DRAFT STANDARD)
26		26
27	2046 Multipurpose Internet Mail Extensions (MIME) Part Two: Media	27	2046 Multipurpose Internet Mail Extensions (MIME) Part Two: Media
28	Types. N. Freed, N. Borenstein. November 1996. (Format: TXT=105854	28	Types. N. Freed, N. Borenstein. November 1996. (Format: TXT=105854
29	bytes) (Obsoletes RFC1521, RFC1522, RFC1590) (Updated by RFC2646,	29	bytes) (Obsoletes RFC1521, RFC1522, RFC1590) (Updated by RFC2646,
30	RFC3798) (Status: DRAFT STANDARD)	30	RFC3798) (Status: DRAFT STANDARD)
31		31
32	2047 MIME (Multipurpose Internet Mail Extensions) Part Three: Message	32	2047 MIME (Multipurpose Internet Mail Extensions) Part Three: Message
33	Header Extensions for Non-ASCII Text. K. Moore. November 1996.	33	Header Extensions for Non-ASCII Text. K. Moore. November 1996.
34	(Format: TXT=33262 bytes) (Obsoletes RFC1521, RFC1522, RFC1590)	34	(Format: TXT=33262 bytes) (Obsoletes RFC1521, RFC1522, RFC1590)
35	(Updated by RFC2184, RFC2231) (Status: DRAFT STANDARD)	35	(Updated by RFC2184, RFC2231) (Status: DRAFT STANDARD)
36		36
37	2183 Communicating Presentation Information in Internet Messages: The	37	2183 Communicating Presentation Information in Internet Messages: The
38	Content-Disposition Header Field. R. Troost, S. Dorner, K. Moore,	38	Content-Disposition Header Field. R. Troost, S. Dorner, K. Moore,
39	Ed.. August 1997. (Format: TXT=23150 bytes) (Updates RFC1806)	39	Ed.. August 1997. (Format: TXT=23150 bytes) (Updates RFC1806)
40	(Updated by RFC2184, RFC2231) (Status: PROPOSED STANDARD)	40	(Updated by RFC2184, RFC2231) (Status: PROPOSED STANDARD)
41		41
42	2231 MIME Parameter Value and Encoded Word Extensions: Character Sets,	42	2231 MIME Parameter Value and Encoded Word Extensions: Character Sets,
43	Languages, and Continuations. N. Freed, K. Moore. November 1997.	43	Languages, and Continuations. N. Freed, K. Moore. November 1997.
44	(Format: TXT=19280 bytes) (Obsoletes RFC2184) (Updates RFC2045,	44	(Format: TXT=19280 bytes) (Obsoletes RFC2184) (Updates RFC2045,
45	RFC2047, RFC2183) (Status: PROPOSED STANDARD)	45	RFC2047, RFC2183) (Status: PROPOSED STANDARD)
46	*/	46	*/
47		47
48		48
49	#include <time.h>	49	#include <time.h>
50		50
51	#include <string>	51	#include <string>
52	#include <map>	52	#include <map>
53		53
54	#include "base64.h"	54	#include "base64.h"
55		55
56	#ifndef NO_NAMESPACES
57	using std::string;
58	#endif
59
60	/** A class to represent a MIME header value with parameters */	56	/** A class to represent a MIME header value with parameters */
61	class MimeHeaderValue {	57	class MimeHeaderValue {
62	public:	58	public:
63	string value;	59	std::string value;
64	std::map<string, string> params;	60	std::map<std::string, std::string> params;
65	};	61	};
66		62
67	/**	63	/**
68	* Parse MIME Content-type and Content-disposition value	64	* Parse MIME Content-type and Content-disposition value
69	*	65	*
70	* @param in the input string should be like: value; pn1=pv1; pn2=pv2.	66	* @param in the input string should be like: value; pn1=pv1; pn2=pv2.
71	* Example: text/plain; charset="iso-8859-1"	67	* Example: text/plain; charset="iso-8859-1"
72	*/	68	*/
73	extern bool parseMimeHeaderValue(const string& in, MimeHeaderValue& psd);	69	extern bool parseMimeHeaderValue(const std::string& in, MimeHeaderValue& psd);
74		70
75	/**	71	/**
76	* Quoted printable decoding. Doubles up as rfc2231 decoder, hence the esc	72	* Quoted Printable decoding.
		73	*
		74	* Doubles up as rfc2231 decoder, with the help of the hence the @param esc
		75	* parameter.
77	* RFC2045 Quoted printable uses '=' , rfc2331 uses '%'. The two encodings are	76	* RFC2045 Quoted Printable uses '=' , RFC2331 uses '%'. The two encodings are
78	* otherwise similar.	77	* otherwise similar.
79	*/	78	*/
80	extern bool qp_decode(const string& in, string &out, char esc = '=');	79	extern bool qp_decode(const std::string& in, std::string &out, char esc = '=');
81		80
82	/** Decode an Internet mail field value encoded according to rfc2047	81	/** Decode an Internet mail field value encoded according to rfc2047
83	*	82	*
84	* Example input: Some words =?iso-8859-1?Q?RE=A0=3A_Smoke_Tests?= more input	83	* Example input: Some words =?iso-8859-1?Q?RE=A0=3A_Smoke_Tests?= more input
85	*	84	*
	...		...
88	* is sometimes used anyway...	87	* is sometimes used anyway...
89	*	88	*
90	* @param in input string, ascii with rfc2047 markup	89	* @param in input string, ascii with rfc2047 markup
91	* @return out output string encoded in utf-8	90	* @return out output string encoded in utf-8
92	*/	91	*/
93	extern bool rfc2047_decode(const string& in, string &out);	92	extern bool rfc2047_decode(const std::string& in, std::string &out);
94		93
95		94
96	/** Decode RFC2822 date to unix time (gmt secs from 1970	95	/** Decode RFC2822 date to unix time (gmt secs from 1970)
97	*	96	*
98	* @param dt date string (the part after Date: )	97	* @param dt date string (the part after Date: )
99	* @return unix time	98	* @return unix time
100	*/	99	*/
101	time_t rfc2822DateToUxTime(const string& dt);	100	time_t rfc2822DateToUxTime(const std::string& dt);
102		101
103	#endif /* _MIME_H_INCLUDED_ */	102	#endif /* _MIME_H_INCLUDED_ */