/*
* $Id$
*
- * This is the MIME parser for Citadel. Sometimes it actually works.
+ * This is the MIME parser for Citadel.
*
- * Copyright (c) 1998-2002 by Art Cancro
- * This code is distributed under the terms of the GNU General Public License.
+ * Copyright (c) 1998-2006 by Art Cancro
+ * This code is distributed under the GNU General Public License v2.
*
*/
-#ifdef DLL_EXPORT
-#define IN_LIBCIT
-#endif
-
#include <stdlib.h>
#include <unistd.h>
#include <stdio.h>
/*
* Convert "quoted-printable" to binary. Returns number of bytes decoded.
+ * according to RFC2045 section 6.7
*/
int CtdlDecodeQuotedPrintable(char *decoded, char *encoded, int sourcelen) {
- char buf[SIZ];
- int buf_length = 0;
- int soft_line_break = 0;
- int ch;
- int decoded_length = 0;
- int i;
+ unsigned int ch;
+ int destpos = 1;
+ int sourcepos = 1;
+ int ignore_last = 0;
+ char *check;
decoded[0] = 0;
- decoded_length = 0;
- buf[0] = 0;
- buf_length = 0;
-
- for (i = 0; i < sourcelen; ++i) {
-
- buf[buf_length++] = encoded[i];
-
- if ( (encoded[i] == '\n')
- || (encoded[i] == 0)
- || (i == (sourcelen-1)) ) {
- buf[buf_length++] = 0;
-
- /*** begin -- process one line ***/
-
- if (buf[strlen(buf)-1] == '\n') {
- buf[strlen(buf)-1] = 0;
- }
- if (buf[strlen(buf)-1] == '\r') {
- buf[strlen(buf)-1] = 0;
- }
- while (isspace(buf[strlen(buf)-1])) {
- buf[strlen(buf)-1] = 0;
- }
- soft_line_break = 0;
-
- while (strlen(buf) > 0) {
- if (!strcmp(buf, "=")) {
- soft_line_break = 1;
- strcpy(buf, "");
- } else if ((strlen(buf)>=3) && (buf[0]=='=')) {
- sscanf(&buf[1], "%02x", &ch);
- decoded[decoded_length++] = ch;
- strcpy(buf, &buf[3]);
- } else {
- decoded[decoded_length++] = buf[0];
- strcpy(buf, &buf[1]);
- }
+ if (sourcelen >0)
+ decoded[0] = encoded[0];
+ while (sourcepos <= sourcelen){
+ check = &decoded[destpos];
+ decoded[destpos] = encoded[sourcepos];
+ if ((ignore_last == 0) && (decoded[destpos-1] == '='))
+ {
+ if ((*check == '\0') ||
+ (*check == '\n') ||
+ (*check == '\r'))
+ {
+ decoded[destpos - 1] = '\0';
+ destpos-=2;
}
- if (soft_line_break == 0) {
- decoded[decoded_length++] = '\r';
- decoded[decoded_length++] = '\n';
+ else if (sourcelen - sourcepos > 2)
+ {
+ sscanf(&encoded[sourcepos], "%02x", &ch);
+ decoded[destpos - 1] = ch;
+ sourcepos++;
+ destpos --;
+ ignore_last = 1;
}
- buf_length = 0;
- /*** end -- process one line ***/
}
+ else
+ ignore_last = 0;
+ destpos ++;
+ sourcepos ++;
}
- decoded[decoded_length++] = 0;
- return(decoded_length);
+ decoded[destpos] = 0;
+ return(destpos - 1);
}
+
/*
* Given a message or message-part body and a length, handle any necessary
* decoding and pass the request up the stack.
*/
void mime_decode(char *partnum,
char *part_start, size_t length,
- char *content_type, char *encoding,
+ char *content_type, char *charset, char *encoding,
char *disposition,
char *name, char *filename,
void (*CallBack)
char *cbdisp,
void *cbcontent,
char *cbtype,
+ char *cbcharset,
size_t cblength,
char *cbencoding,
void *cbuserdata),
char *cbdisp,
void *cbcontent,
char *cbtype,
+ char *cbcharset,
size_t cblength,
char *cbencoding,
void *cbuserdata),
char *cbdisp,
void *cbcontent,
char *cbtype,
+ char *cbcharset,
size_t cblength,
char *cbencoding,
void *cbuserdata),
if (CallBack != NULL) {
CallBack(name, filename, fixed_partnum(partnum),
disposition, part_start,
- content_type, length, encoding, userdata);
+ content_type, charset, length, encoding, userdata);
}
return;
}
+ /* Fail silently if we hit an unknown encoding. */
if ((strcasecmp(encoding, "base64"))
&& (strcasecmp(encoding, "quoted-printable"))) {
return;
}
+
/*
- * Allocate a buffer for the decoded data. The output buffer is the
- * same size as the input buffer; this assumes that the decoded data
- * will never be larger than the encoded data. This is a safe
- * assumption with base64, uuencode, and quoted-printable.
+ * Allocate a buffer for the decoded data. The output buffer is slightly
+ * larger than the input buffer; this assumes that the decoded data
+ * will never be significantly larger than the encoded data. This is a
+ * safe assumption with base64, uuencode, and quoted-printable.
*/
- decoded = mallok(length+2048);
+ decoded = malloc(length + 32768);
if (decoded == NULL) {
return;
}
bytes_decoded = CtdlDecodeBase64(decoded, part_start, length);
}
else if (!strcasecmp(encoding, "quoted-printable")) {
- bytes_decoded = CtdlDecodeQuotedPrintable(decoded,
- part_start, length);
+ bytes_decoded = CtdlDecodeQuotedPrintable(decoded, part_start, length);
}
if (bytes_decoded > 0) if (CallBack != NULL) {
CallBack(name, filename, fixed_partnum(partnum),
disposition, decoded,
- content_type, bytes_decoded, "binary", userdata);
+ content_type, charset, bytes_decoded, "binary", userdata);
}
- phree(decoded);
+ free(decoded);
}
/*
char *cbdisp,
void *cbcontent,
char *cbtype,
+ char *cbcharset,
size_t cblength,
char *cbencoding,
void *cbuserdata),
char *cbdisp,
void *cbcontent,
char *cbtype,
+ char *cbcharset,
size_t cblength,
char *cbencoding,
void *cbuserdata),
char *cbdisp,
void *cbcontent,
char *cbtype,
+ char *cbcharset,
size_t cblength,
char *cbencoding,
void *cbuserdata),
{
char *ptr;
+ char *srch = NULL;
char *part_start, *part_end = NULL;
char buf[SIZ];
char *header;
char *boundary;
char *startary;
+ size_t startary_len = 0;
char *endary;
+ char *next_boundary;
char *content_type;
+ char *charset;
size_t content_length;
char *encoding;
char *disposition;
ptr = content_start;
content_length = 0;
- boundary = mallok(SIZ);
+ boundary = malloc(SIZ);
memset(boundary, 0, SIZ);
- startary = mallok(SIZ);
+ startary = malloc(SIZ);
memset(startary, 0, SIZ);
- endary = mallok(SIZ);
+ endary = malloc(SIZ);
memset(endary, 0, SIZ);
- header = mallok(SIZ);
+ header = malloc(SIZ);
memset(header, 0, SIZ);
- content_type = mallok(SIZ);
+ content_type = malloc(SIZ);
memset(content_type, 0, SIZ);
- encoding = mallok(SIZ);
+ charset = malloc(SIZ);
+ memset(charset, 0, SIZ);
+
+ encoding = malloc(SIZ);
memset(encoding, 0, SIZ);
- content_type_name = mallok(SIZ);
+ content_type_name = malloc(SIZ);
memset(content_type_name, 0, SIZ);
- content_disposition_name = mallok(SIZ);
+ content_disposition_name = malloc(SIZ);
memset(content_disposition_name, 0, SIZ);
- filename = mallok(SIZ);
+ filename = malloc(SIZ);
memset(filename, 0, SIZ);
- disposition = mallok(SIZ);
+ disposition = malloc(SIZ);
memset(disposition, 0, SIZ);
/* If the caller didn't supply an endpointer, generate one by measure */
goto end_parser;
}
- for (i = 0; i < strlen(buf); ++i)
- if (isspace(buf[i]))
+ for (i = 0; i < strlen(buf); ++i) {
+ if (isspace(buf[i])) {
buf[i] = ' ';
+ }
+ }
+
if (!isspace(buf[0])) {
if (!strncasecmp(header, "Content-type: ", 14)) {
strcpy(content_type, &header[14]);
extract_key(content_type_name, content_type, "name");
+ extract_key(charset, content_type, "charset");
/* Deal with weird headers */
if (strchr(content_type, ' '))
*(strchr(content_type, ' ')) = '\0';
/* Tell the client about this message's multipartedness */
if (PreMultiPartCallBack != NULL) {
PreMultiPartCallBack("", "", partnum, "",
- NULL, content_type,
+ NULL, content_type, charset,
0, encoding, userdata);
}
/* Figure out where the boundaries are */
snprintf(startary, SIZ, "--%s", boundary);
snprintf(endary, SIZ, "--%s--", boundary);
+ startary_len = strlen(startary);
+
+ part_start = NULL;
do {
- if ( (!strncasecmp(ptr, startary, strlen(startary)))
- || (!strncasecmp(ptr, endary, strlen(endary))) ) {
- if (part_start != NULL) {
- if (strlen(partnum) > 0) {
- snprintf(nested_partnum,
- sizeof nested_partnum,
- "%s.%d", partnum,
- ++part_seq);
- }
- else {
- snprintf(nested_partnum,
- sizeof nested_partnum,
- "%d", ++part_seq);
- }
- the_mime_parser(nested_partnum,
- part_start, part_end,
- CallBack,
- PreMultiPartCallBack,
- PostMultiPartCallBack,
- userdata,
- dont_decode);
+ next_boundary = NULL;
+ for (srch=ptr; srch<content_end; ++srch) {
+ if (!memcmp(srch, startary, startary_len)) {
+ next_boundary = srch;
+ srch = content_end;
}
- ptr = memreadline(ptr, buf, SIZ);
- part_start = ptr;
}
- else {
- part_end = ptr;
- ++ptr;
+
+ if ( (part_start != NULL) && (next_boundary != NULL) ) {
+ part_end = next_boundary;
+ --part_end;
+
+ if (strlen(partnum) > 0) {
+ snprintf(nested_partnum,
+ sizeof nested_partnum,
+ "%s.%d", partnum,
+ ++part_seq);
+ }
+ else {
+ snprintf(nested_partnum,
+ sizeof nested_partnum,
+ "%d", ++part_seq);
+ }
+ the_mime_parser(nested_partnum,
+ part_start, part_end,
+ CallBack,
+ PreMultiPartCallBack,
+ PostMultiPartCallBack,
+ userdata,
+ dont_decode);
}
- /* If we pass out of scope in the MIME multipart (by
- * hitting the end boundary), force the pointer out
- * of scope so this loop ends.
- */
- if (ptr < content_end) {
- if (!strcasecmp(ptr, endary)) {
- ptr = content_end++;
+
+ if (next_boundary != NULL) {
+ /* If we pass out of scope, don't attempt to
+ * read past the end boundary. */
+ if (!strcmp(next_boundary, endary)) {
+ ptr = content_end;
}
+ else {
+ /* Set up for the next part. */
+ part_start = strstr(next_boundary, "\n");
+ ++part_start;
+ ptr = part_start;
+ }
+ }
+ else {
+ /* Invalid end of multipart. Bail out! */
+ ptr = content_end;
}
- } while (ptr <= content_end);
+ } while ( (ptr < content_end) && (next_boundary != NULL) );
+
if (PostMultiPartCallBack != NULL) {
PostMultiPartCallBack("", "", partnum, "", NULL,
- content_type, 0, encoding, userdata);
+ content_type, charset, 0, encoding, userdata);
}
goto end_parser;
}
++length;
}
part_end = content_end;
+ /* fix an off-by-one error */
+ --part_end;
+ --length;
/* Truncate if the header told us to */
if ( (content_length > 0) && (length > content_length) ) {
else {
name = content_type_name;
}
-
+
+ /* lprintf(CTDL_DEBUG, "mime_decode part=%s, len=%d, type=%s, charset=%s, encoding=%s\n",
+ partnum, length, content_type, charset, encoding); */
+
+ /* Ok, we've got a non-multipart part here, so do something with it.
+ */
mime_decode(partnum,
- part_start, length,
- content_type, encoding, disposition,
- name, filename,
- CallBack, NULL, NULL,
- userdata, dont_decode);
+ part_start, length,
+ content_type, charset, encoding, disposition,
+ name, filename,
+ CallBack, NULL, NULL,
+ userdata, dont_decode
+ );
+
+ /*
+ * Now if it's an encapsulated message/rfc822 then we have to recurse into it
+ */
+ if (!strcasecmp(content_type, "message/rfc822")) {
+
+ if (PreMultiPartCallBack != NULL) {
+ PreMultiPartCallBack("", "", partnum, "",
+ NULL, content_type, charset,
+ 0, encoding, userdata);
+ }
+ if (CallBack != NULL) {
+ if (strlen(partnum) > 0) {
+ snprintf(nested_partnum,
+ sizeof nested_partnum,
+ "%s.%d", partnum,
+ ++part_seq);
+ }
+ else {
+ snprintf(nested_partnum,
+ sizeof nested_partnum,
+ "%d", ++part_seq);
+ }
+ the_mime_parser(nested_partnum,
+ part_start, part_end,
+ CallBack,
+ PreMultiPartCallBack,
+ PostMultiPartCallBack,
+ userdata,
+ dont_decode
+ );
+ }
+ if (PostMultiPartCallBack != NULL) {
+ PostMultiPartCallBack("", "", partnum, "", NULL,
+ content_type, charset, 0, encoding, userdata);
+ }
+
+
+ }
+
}
end_parser: /* free the buffers! end the oppression!! */
- phree(boundary);
- phree(startary);
- phree(endary);
- phree(header);
- phree(content_type);
- phree(encoding);
- phree(content_type_name);
- phree(content_disposition_name);
- phree(filename);
- phree(disposition);
+ free(boundary);
+ free(startary);
+ free(endary);
+ free(header);
+ free(content_type);
+ free(charset);
+ free(encoding);
+ free(content_type_name);
+ free(content_disposition_name);
+ free(filename);
+ free(disposition);
}
char *cbdisp,
void *cbcontent,
char *cbtype,
+ char *cbcharset,
size_t cblength,
char *cbencoding,
void *cbuserdata),
char *cbdisp,
void *cbcontent,
char *cbtype,
+ char *cbcharset,
size_t cblength,
char *cbencoding,
void *cbuserdata),
char *cbdisp,
void *cbcontent,
char *cbtype,
+ char *cbcharset,
size_t cblength,
char *cbencoding,
void *cbuserdata),