]> code.citadel.org Git - citadel.git/blobdiff - webcit/mime_parser.c
* Header file adjustments to make it work on FreeBSD
[citadel.git] / webcit / mime_parser.c
index 021d5e52be11fba11822e472a13a33b7611ee25c..7067bb1a8d2d684ee470260cda225e7629d582b3 100644 (file)
@@ -3,22 +3,42 @@
  *
  * This is the MIME parser for Citadel.  Sometimes it actually works.
  *
- * Copyright (c) 1998-2001 by Art Cancro
+ * Copyright (c) 1998-2003 by Art Cancro
  * This code is distributed under the terms of the GNU General Public License.
  *
  */
 
+#include <ctype.h>
 #include <stdlib.h>
+#ifdef HAVE_UNISTD_H
 #include <unistd.h>
+#endif
 #include <stdio.h>
+#ifdef HAVE_FCNTL_H
+#include <fcntl.h>
+#endif
 #include <signal.h>
 #include <sys/types.h>
-#include <ctype.h>
+#include <sys/wait.h>
+#include <sys/socket.h>
+#ifdef HAVE_SYS_TIME_H
+#include <sys/time.h>
+#endif
+#ifdef HAVE_LIMITS_H
+#include <limits.h>
+#endif
+#include <netinet/in.h>
+#include <arpa/inet.h>
+#include <netdb.h>
 #include <string.h>
-#include <sys/stat.h>
+#include <pwd.h>
 #include <errno.h>
+#include <stdarg.h>
 #include <pthread.h>
+#include <signal.h>
 #include "webcit.h"
+#include "webserver.h"
+
 #include "mime_parser.h"
 
 
@@ -58,7 +78,7 @@ char *fixed_partnum(char *supplied_partnum) {
 /*
  * Convert "quoted-printable" to binary.  Returns number of bytes decoded.
  */
-int decode_quoted_printable(char *decoded, char *encoded, int sourcelen) {
+int CtdlDecodeQuotedPrintable(char *decoded, char *encoded, int sourcelen) {
        char buf[SIZ];
        int buf_length = 0;
        int soft_line_break = 0;
@@ -183,29 +203,27 @@ void mime_decode(char *partnum,
                        }
                return;
        }
+       
        if ((strcasecmp(encoding, "base64"))
            && (strcasecmp(encoding, "quoted-printable"))) {
-               fprintf(stderr, "ERROR: unknown MIME encoding '%s'\n", encoding);
                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.  Just to
-        * be safe, we still pad the buffer a bit.
+        * assumption with base64, uuencode, and quoted-printable.
         */
-       decoded = malloc(length + 1024);
+       decoded = mallok(length+2048);
        if (decoded == NULL) {
-               fprintf(stderr, "ERROR: cannot allocate memory.\n");
                return;
        }
 
        if (!strcasecmp(encoding, "base64")) {
-               bytes_decoded = decode_base64(decoded, part_start);
+               bytes_decoded = CtdlDecodeBase64(decoded, part_start, length);
        }
        else if (!strcasecmp(encoding, "quoted-printable")) {
-               bytes_decoded = decode_quoted_printable(decoded,
+               bytes_decoded = CtdlDecodeQuotedPrintable(decoded,
                                                        part_start, length);
        }
 
@@ -215,12 +233,9 @@ void mime_decode(char *partnum,
                        content_type, bytes_decoded, "binary", userdata);
        }
 
-       free(decoded);
+       phree(decoded);
 }
 
-
-
-
 /*
  * Break out the components of a multipart message
  * (This function expects to be fed HEADERS + CONTENT)
@@ -275,7 +290,9 @@ void the_mime_parser(char *partnum,
        size_t content_length;
        char *encoding;
        char *disposition;
-       char *name;
+       char *name = NULL;
+       char *content_type_name;
+       char *content_disposition_name;
        char *filename;
        int is_multipart;
        int part_seq = 0;
@@ -283,35 +300,37 @@ void the_mime_parser(char *partnum,
        size_t length;
        char nested_partnum[SIZ];
 
-       fprintf(stderr, "the_mime_parser() called\n");
        ptr = content_start;
        content_length = 0;
 
-       boundary = malloc(SIZ);
+       boundary = mallok(SIZ);
        memset(boundary, 0, SIZ);
 
-       startary = malloc(SIZ);
+       startary = mallok(SIZ);
        memset(startary, 0, SIZ);
 
-       endary = malloc(SIZ);
+       endary = mallok(SIZ);
        memset(endary, 0, SIZ);
 
-       header = malloc(SIZ);
+       header = mallok(SIZ);
        memset(header, 0, SIZ);
 
-       content_type = malloc(SIZ);
+       content_type = mallok(SIZ);
        memset(content_type, 0, SIZ);
 
-       encoding = malloc(SIZ);
+       encoding = mallok(SIZ);
        memset(encoding, 0, SIZ);
 
-       name = malloc(SIZ);
-       memset(name, 0, SIZ);
+       content_type_name = mallok(SIZ);
+       memset(content_type_name, 0, SIZ);
+
+       content_disposition_name = mallok(SIZ);
+       memset(content_disposition_name, 0, SIZ);
 
-       filename = malloc(SIZ);
+       filename = mallok(SIZ);
        memset(filename, 0, SIZ);
 
-       disposition = malloc(SIZ);
+       disposition = mallok(SIZ);
        memset(disposition, 0, SIZ);
 
        /* If the caller didn't supply an endpointer, generate one by measure */
@@ -322,7 +341,7 @@ void the_mime_parser(char *partnum,
        /* Learn interesting things from the headers */
        strcpy(header, "");
        do {
-               ptr = memreadline(ptr, buf, sizeof buf);
+               ptr = memreadline(ptr, buf, SIZ);
                if (ptr >= content_end) {
                        goto end_parser;
                }
@@ -333,12 +352,16 @@ void the_mime_parser(char *partnum,
                if (!isspace(buf[0])) {
                        if (!strncasecmp(header, "Content-type: ", 14)) {
                                strcpy(content_type, &header[14]);
-                               extract_key(name, content_type, "name");
-                               fprintf(stderr, "Extracted content-type <%s>\n",
-                                       content_type);
+                               extract_key(content_type_name, content_type, "name");
+                               /* Deal with weird headers */
+                               if (strchr(content_type, ' '))
+                                       *(strchr(content_type, ' ')) = '\0';
+                               if (strchr(content_type, ';'))
+                                       *(strchr(content_type, ';')) = '\0';
                        }
                        if (!strncasecmp(header, "Content-Disposition: ", 21)) {
                                strcpy(disposition, &header[21]);
+                               extract_key(content_disposition_name, disposition, "name");
                                extract_key(filename, disposition, "filename");
                        }
                        if (!strncasecmp(header, "Content-length: ", 16)) {
@@ -351,20 +374,16 @@ void the_mime_parser(char *partnum,
                                extract_key(boundary, header, "boundary");
                        strcpy(header, "");
                }
-               if ((strlen(header) + strlen(buf) + 2) < sizeof(header))
+               if ((strlen(header) + strlen(buf) + 2) < SIZ)
                        strcat(header, buf);
        } while ((strlen(buf) > 0) && (*ptr != 0));
 
-       for (i = 0; i < strlen(disposition); ++i)
-               if (disposition[i] == ';')
-                       disposition[i] = 0;
-       while (isspace(disposition[0]))
-               strcpy(disposition, &disposition[1]);
-       for (i = 0; i < strlen(content_type); ++i)
-               if (content_type[i] == ';')
-                       content_type[i] = 0;
-       while (isspace(content_type[0]))
-               strcpy(content_type, &content_type[1]);
+       if (strchr(disposition, ';'))
+               *(strchr(disposition, ';')) = '\0';
+       striplt(disposition);
+       if (strchr(content_type, ';'))
+               *(strchr(content_type, ';')) = '\0';
+       striplt(content_type);
 
        if (strlen(boundary) > 0) {
                is_multipart = 1;
@@ -372,9 +391,6 @@ void the_mime_parser(char *partnum,
                is_multipart = 0;
        }
 
-       fprintf(stderr, "is_multipart=%d, boundary=<%s>\n",
-               is_multipart, boundary);
-
        /* If this is a multipart message, then recursively process it */
        part_start = NULL;
        if (is_multipart) {
@@ -387,20 +403,22 @@ void the_mime_parser(char *partnum,
                }
 
                /* Figure out where the boundaries are */
-               sprintf(startary, "--%s", boundary);
-               sprintf(endary, "--%s--", boundary);
+               snprintf(startary, SIZ, "--%s", boundary);
+               snprintf(endary, SIZ, "--%s--", boundary);
                do {
                        if ( (!strncasecmp(ptr, startary, strlen(startary)))
                           || (!strncasecmp(ptr, endary, strlen(endary))) ) {
-                               fprintf(stderr, "hit boundary!\n");
                                if (part_start != NULL) {
                                        if (strlen(partnum) > 0) {
-                                               sprintf(nested_partnum, "%s.%d",
-                                                       partnum, ++part_seq);
+                                               snprintf(nested_partnum,
+                                                        sizeof nested_partnum,
+                                                        "%s.%d", partnum,
+                                                        ++part_seq);
                                        }
                                        else {
-                                               sprintf(nested_partnum, "%d",
-                                                       ++part_seq);
+                                               snprintf(nested_partnum,
+                                                        sizeof nested_partnum,
+                                                        "%d", ++part_seq);
                                        }
                                        the_mime_parser(nested_partnum,
                                                    part_start, part_end,
@@ -410,14 +428,23 @@ void the_mime_parser(char *partnum,
                                                        userdata,
                                                        dont_decode);
                                }
-                               ptr = memreadline(ptr, buf, sizeof(buf));
+                               ptr = memreadline(ptr, buf, SIZ);
                                part_start = ptr;
                        }
                        else {
                                part_end = ptr;
                                ++ptr;
                        }
-               } while ( (strcasecmp(ptr, endary)) && (ptr <= content_end) );
+                       /* 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++;
+                               }
+                       }
+               } while (ptr <= content_end);
                if (PostMultiPartCallBack != NULL) {
                        PostMultiPartCallBack("", "", partnum, "", NULL,
                                content_type, 0, encoding, userdata);
@@ -427,7 +454,6 @@ void the_mime_parser(char *partnum,
 
        /* If it's not a multipart message, then do something with it */
        if (!is_multipart) {
-               fprintf(stderr, "doing non-multipart thing\n");
                part_start = ptr;
                length = 0;
                while (ptr < content_end) {
@@ -435,11 +461,24 @@ void the_mime_parser(char *partnum,
                        ++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) ) {
                        length = content_length;
-                       fprintf(stderr, "truncated to %ld\n", (long)content_length);
+               }
+
+               /* Sometimes the "name" field is tacked on to Content-type,
+                * and sometimes it's tacked on to Content-disposition.  Use
+                * whichever one we have.
+                */
+               if (strlen(content_disposition_name) > strlen(content_type_name)) {
+                       name = content_disposition_name;
+               }
+               else {
+                       name = content_type_name;
                }
                
                mime_decode(partnum,
@@ -451,23 +490,20 @@ void the_mime_parser(char *partnum,
        }
 
 end_parser:    /* free the buffers!  end the oppression!! */
-       free(boundary);
-       free(startary);
-       free(endary);   
-       free(header);
-       free(content_type);
-       free(encoding);
-       free(name);
-       free(filename);
-       free(disposition);
+       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);
 }
 
 
 
-
-
-
-
 /*
  * Entry point for the MIME parser.
  * (This function expects to be fed HEADERS + CONTENT)
@@ -515,10 +551,10 @@ void mime_parser(char *content_start,
 )
 {
 
-       fprintf(stderr, "mime_parser() called\n");
        the_mime_parser("", content_start, content_end,
                        CallBack,
                        PreMultiPartCallBack,
                        PostMultiPartCallBack,
                        userdata, dont_decode);
 }
+