view src/readpst.c @ 359:a3e674fade6c

From Jeffrey Morlan: pst_parse_block misreads Table Contexts (aka "type 2") with a multi-block Row Matrix ("ind2"). Rows are never split between blocks - every block except the last has padding at the end which should be ignored. I've only seen this affect the recipients table, but presumably it could affect attachments too. This was causing out-of-bounds memory ranges to be returned from pst_getBlockOffset and later access; patch fixes both the table reading issue and adds a missing bounds check to pst_getBlockOffset (so as not to risk a segfault if the PST is corrupted).
author Carl Byington <carl@five-ten-sg.com>
date Wed, 06 Jul 2016 10:20:12 -0700
parents 6abc3054cba2
children 3a1d25c579c6
line wrap: on
line source

/***
 * readpst.c
 * Part of the LibPST project
 * Written by David Smith
 *            dave.s@earthcorp.com
 */

#include "define.h"
#include "lzfu.h"
#include "msg.h"

#define OUTPUT_TEMPLATE "%s"
#define OUTPUT_KMAIL_DIR_TEMPLATE ".%s.directory"
#define KMAIL_INDEX ".%s.index"
#define SEP_MAIL_FILE_TEMPLATE "%i%s"

// max size of the c_time char*. It will store the date of the email
#define C_TIME_SIZE 500

struct file_ll {
    char *name;
    char *dname;
    FILE * output;
    int32_t stored_count;
    int32_t item_count;
    int32_t skip_count;
    int32_t type;
};

int       grim_reaper();
pid_t     try_fork(char* folder);
void      process(pst_item *outeritem, pst_desc_tree *d_ptr);
void      write_email_body(FILE *f, char *body);
void      removeCR(char *c);
void      usage();
void      version();
char*     mk_kmail_dir(char* fname);
int       close_kmail_dir();
char*     mk_recurse_dir(char* dir, int32_t folder_type);
int       close_recurse_dir();
char*     mk_separate_dir(char *dir);
int       close_separate_dir();
void      mk_separate_file(struct file_ll *f, char *extension, int openit);
void      close_separate_file(struct file_ll *f);
char*     my_stristr(char *haystack, char *needle);
void      check_filename(char *fname);
int       acceptable_ext(pst_item_attach* attach);
void      write_separate_attachment(char f_name[], pst_item_attach* attach, int attach_num, pst_file* pst);
void      write_embedded_message(FILE* f_output, pst_item_attach* attach, char *boundary, pst_file* pf, int save_rtf, char** extra_mime_headers);
void      write_inline_attachment(FILE* f_output, pst_item_attach* attach, char *boundary, pst_file* pst);
int       valid_headers(char *header);
void      header_has_field(char *header, char *field, int *flag);
void      header_get_subfield(char *field, const char *subfield, char *body_subfield, size_t size_subfield);
char*     header_get_field(char *header, char *field);
char*     header_end_field(char *field);
void      header_strip_field(char *header, char *field);
int       test_base64(char *body, size_t len);
void      find_html_charset(char *html, char *charset, size_t charsetlen);
void      find_rfc822_headers(char** extra_mime_headers);
void      write_body_part(FILE* f_output, pst_string *body, char *mime, char *charset, char *boundary, pst_file* pst);
void      write_schedule_part_data(FILE* f_output, pst_item* item, const char* sender, const char* method);
void      write_schedule_part(FILE* f_output, pst_item* item, const char* sender, const char* boundary);
void      write_normal_email(FILE* f_output, char f_name[], pst_item* item, int mode, int mode_MH, pst_file* pst, int save_rtf, int embedding, char** extra_mime_headers);
void      write_vcard(FILE* f_output, pst_item *item, pst_item_contact* contact, char comment[]);
int       write_extra_categories(FILE* f_output, pst_item* item);
void      write_journal(FILE* f_output, pst_item* item);
void      write_appointment(FILE* f_output, pst_item *item);
void      create_enter_dir(struct file_ll* f, pst_item *item);
void      close_enter_dir(struct file_ll *f);

const char*  prog_name;
char*  output_dir = ".";
char*  kmail_chdir = NULL;

// Normal mode just creates mbox format files in the current directory. Each file is named
// the same as the folder's name that it represents
#define MODE_NORMAL 0

// KMail mode creates a directory structure suitable for being used directly
// by the KMail application
#define MODE_KMAIL 1

// recurse mode creates a directory structure like the PST file. Each directory
// contains only one file which stores the emails in mboxrd format.
#define MODE_RECURSE 2

// separate mode creates the same directory structure as recurse. The emails are stored in
// separate files, numbering from 1 upward. Attachments belonging to the emails are
// saved as email_no-filename (e.g. 1-samplefile.doc or 1-Attachment2.zip)
#define MODE_SEPARATE 3


// Output Normal just prints the standard information about what is going on
#define OUTPUT_NORMAL 0

// Output Quiet is provided so that only errors are printed
#define OUTPUT_QUIET 1

// default mime-type for attachments that have a null mime-type
#define MIME_TYPE_DEFAULT "application/octet-stream"
#define RFC822            "message/rfc822"

// output mode for contacts
#define CMODE_VCARD 0
#define CMODE_LIST  1

// output mode for deleted items
#define DMODE_EXCLUDE 0
#define DMODE_INCLUDE 1

// Output type mode flags
#define OTMODE_EMAIL        1
#define OTMODE_APPOINTMENT  2
#define OTMODE_JOURNAL      4
#define OTMODE_CONTACT      8

// output settings for RTF bodies
// filename for the attachment
#define RTF_ATTACH_NAME "rtf-body.rtf"
// mime type for the attachment
#define RTF_ATTACH_TYPE "application/rtf"

// global settings
int         mode         = MODE_NORMAL;
int         mode_MH      = 0;   // a submode of MODE_SEPARATE
int         mode_EX      = 0;   // a submode of MODE_SEPARATE
int         mode_MSG     = 0;   // a submode of MODE_SEPARATE
int         mode_thunder = 0;   // a submode of MODE_RECURSE
int         output_mode  = OUTPUT_NORMAL;
int         contact_mode = CMODE_VCARD;
int         deleted_mode = DMODE_EXCLUDE;
int         output_type_mode = 0xff;    // Default to all.
int         contact_mode_specified = 0;
int         overwrite = 0;
int         prefer_utf8 = 0;
int         save_rtf_body = 1;
int         file_name_len = 10;     // enough room for MODE_SPEARATE file name
pst_file    pstfile;
regex_t     meta_charset_pattern;
char*       default_charset = NULL;
char*       acceptable_extensions = NULL;

int         number_processors = 1;  // number of cpus we have
int         max_children  = 0;      // based on number of cpus and command line args
int         max_child_specified = 0;// have command line arg -j
int         active_children;        // number of children of this process, cannot be larger than max_children
pid_t*      child_processes;        // setup by main(), and at the start of new child process

#ifdef HAVE_SEMAPHORE_H
int         shared_memory_id;
sem_t*      global_children = NULL;
sem_t*      output_mutex    = NULL;
#endif


int grim_reaper(int waitall)
{
    int available = 0;
#ifdef HAVE_FORK
#ifdef HAVE_SEMAPHORE_H
    if (global_children) {
        //sem_getvalue(global_children, &available);
        //printf("grim reaper %s for pid %d (parent %d) with %d children, %d available\n", (waitall) ? "all" : "", getpid(), getppid(), active_children, available);
        //fflush(stdout);
        int i,j;
        for (i=0; i<active_children; i++) {
            int status;
            pid_t child = child_processes[i];
            pid_t ch = waitpid(child, &status, ((waitall) ? 0 : WNOHANG));
            if (ch == child) {
                // check termination status
                //if (WIFEXITED(status)) {
                //    int ext = WEXITSTATUS(status);
                //    printf("Process %d exited with status  %d\n", child, ext);
                //    fflush(stdout);
                //}
                if (WIFSIGNALED(status)) {
                    int sig = WTERMSIG(status);
                    DEBUG_INFO(("Process %d terminated with signal %d\n", child, sig));
                    //printf("Process %d terminated with signal %d\n", child, sig);
                    //fflush(stdout);
                }
                if (status != 0) {
                    exit(status);
                }
                // this has terminated, remove it from the list
                for (j=i; j<active_children-1; j++) {
                    child_processes[j] = child_processes[j+1];
                }
                active_children--;
                i--;
            }
        }
        sem_getvalue(global_children, &available);
        //printf("grim reaper %s for pid %d with %d children, %d available\n", (waitall) ? "all" : "", getpid(), active_children, available);
        //fflush(stdout);
    }
#endif
#endif
    return available;
}


pid_t try_fork(char *folder)
{
#ifdef HAVE_FORK
#ifdef HAVE_SEMAPHORE_H
    int available = grim_reaper(0);
    // If children have called sem_post but not exited yet, we could have available > 0 but active_children == max_children
    if (available && active_children < max_children) {
        sem_wait(global_children);
        pid_t child = fork();
        if (child < 0) {
            // fork failed, pretend it worked and we are the child
            return 0;
        }
        else if (child == 0) {
            // fork worked, and we are the child, reinitialize *our* list of children
            active_children = 0;
            memset(child_processes, 0, sizeof(pid_t) * max_children);
            pst_reopen(&pstfile);   // close and reopen the pst file to get an independent file position pointer
        }
        else {
            // fork worked, and we are the parent, record this child that we need to wait for
            //pid_t me = getpid();
            //printf("parent %d forked child pid %d to process folder %s\n", me, child, folder);
            //fflush(stdout);
            child_processes[active_children++] = child;
        }
        return child;
    }
    else {
        return 0;   // pretend to have forked and we are the child
    }
#endif
#endif
    return 0;
}


void process(pst_item *outeritem, pst_desc_tree *d_ptr)
{
    struct file_ll ff;
    pst_item *item = NULL;

    DEBUG_ENT("process");
    memset(&ff, 0, sizeof(ff));
    create_enter_dir(&ff, outeritem);

    for (; d_ptr; d_ptr = d_ptr->next) {
        DEBUG_INFO(("New item record\n"));
        if (!d_ptr->desc) {
            ff.skip_count++;
            DEBUG_WARN(("ERROR item's desc record is NULL\n"));
            continue;
        }
        DEBUG_INFO(("Desc Email ID %#"PRIx64" [d_ptr->d_id = %#"PRIx64"]\n", d_ptr->desc->i_id, d_ptr->d_id));

        item = pst_parse_item(&pstfile, d_ptr, NULL);
        DEBUG_INFO(("About to process item\n"));

        if (!item) {
            ff.skip_count++;
            DEBUG_INFO(("A NULL item was seen\n"));
            continue;
        }

        if (item->subject.str) {
            DEBUG_INFO(("item->subject = %s\n", item->subject.str));
        }

        if (item->folder && item->file_as.str) {
            DEBUG_INFO(("Processing Folder \"%s\"\n", item->file_as.str));
            if (output_mode != OUTPUT_QUIET) {
                pst_debug_lock();
                    printf("Processing Folder \"%s\"\n", item->file_as.str);
                    fflush(stdout);
                pst_debug_unlock();
            }
            ff.item_count++;
            if (d_ptr->child && (deleted_mode == DMODE_INCLUDE || strcasecmp(item->file_as.str, "Deleted Items"))) {
                //if this is a non-empty folder other than deleted items, we want to recurse into it
                pid_t parent = getpid();
                pid_t child = try_fork(item->file_as.str);
                if (child == 0) {
                    // we are the child process, or the original parent if no children were available
                    pid_t me = getpid();
                    process(item, d_ptr->child);
#ifdef HAVE_FORK
#ifdef HAVE_SEMAPHORE_H
                    if (me != parent) {
                        // we really were a child, forked for the sole purpose of processing this folder
                        // free my child count slot before really exiting, since
                        // all I am doing here is waiting for my children to exit
                        sem_post(global_children);
                        grim_reaper(1); // wait for all my child processes to exit
                        exit(0);        // really exit
                    }
#endif
#endif
                }
            }

        } else if (item->contact && (item->type == PST_TYPE_CONTACT)) {
            DEBUG_INFO(("Processing Contact\n"));
            if (!(output_type_mode & OTMODE_CONTACT)) {
                ff.skip_count++;
                DEBUG_INFO(("skipping contact: not in output type list\n"));
            }
            else {
                if (!ff.type) ff.type = item->type;
                if ((ff.type != PST_TYPE_CONTACT) && (mode != MODE_SEPARATE)) {
                    ff.skip_count++;
                    DEBUG_INFO(("I have a contact, but the folder type %"PRIi32" isn't a contacts folder. Skipping it\n", ff.type));
                }
                else {
                    ff.item_count++;
                    if (mode == MODE_SEPARATE) mk_separate_file(&ff, (mode_EX) ? ".vcf" : "", 1);
                    if (contact_mode == CMODE_VCARD) {
                        pst_convert_utf8_null(item, &item->comment);
                        write_vcard(ff.output, item, item->contact, item->comment.str);
                    }
                    else {
                        pst_convert_utf8(item, &item->contact->fullname);
                        pst_convert_utf8(item, &item->contact->address1);
                        fprintf(ff.output, "%s <%s>\n", item->contact->fullname.str, item->contact->address1.str);
                    }
                    if (mode == MODE_SEPARATE) close_separate_file(&ff);
                }
            }

        } else if (item->email && ((item->type == PST_TYPE_NOTE) || (item->type == PST_TYPE_SCHEDULE) || (item->type == PST_TYPE_REPORT))) {
            DEBUG_INFO(("Processing Email\n"));
            if (!(output_type_mode & OTMODE_EMAIL)) {
                ff.skip_count++;
                DEBUG_INFO(("skipping email: not in output type list\n"));
            }
            else {
                if (!ff.type) ff.type = item->type;
                if ((ff.type != PST_TYPE_NOTE) && (ff.type != PST_TYPE_SCHEDULE) && (ff.type != PST_TYPE_REPORT) && (mode != MODE_SEPARATE)) {
                    ff.skip_count++;
                    DEBUG_INFO(("I have an email type %"PRIi32", but the folder type %"PRIi32" isn't an email folder. Skipping it\n", item->type, ff.type));
                }
                else {
                    char *extra_mime_headers = NULL;
                    ff.item_count++;
                    if (mode == MODE_SEPARATE) {
                        // process this single email message, possibly forking
                        pid_t parent = getpid();
                        pid_t child = try_fork(item->file_as.str);
                        if (child == 0) {
                            // we are the child process, or the original parent if no children were available
                            pid_t me = getpid();
                            mk_separate_file(&ff, (mode_EX) ? ".eml" : "", 1);
                            write_normal_email(ff.output, ff.name, item, mode, mode_MH, &pstfile, save_rtf_body, 0, &extra_mime_headers);
                            close_separate_file(&ff);
                            if (mode_MSG) {
                                mk_separate_file(&ff, ".msg", 0);
                                write_msg_email(ff.name, item, &pstfile);
                            }
#ifdef HAVE_FORK
#ifdef HAVE_SEMAPHORE_H
                            if (me != parent) {
                                // we really were a child, forked for the sole purpose of processing this message
                                // free my child count slot before really exiting, since
                                // all I am doing here is waiting for my children to exit
                                sem_post(global_children);
                                grim_reaper(1); // wait for all my child processes to exit - there should not be any
                                exit(0);        // really exit
                            }
#endif
#endif
                        }
                    }
                    else {
                        // process this single email message, cannot fork since not separate mode
                        write_normal_email(ff.output, ff.name, item, mode, mode_MH, &pstfile, save_rtf_body, 0, &extra_mime_headers);
                    }
                }
            }

        } else if (item->journal && (item->type == PST_TYPE_JOURNAL)) {
            DEBUG_INFO(("Processing Journal Entry\n"));
            if (!(output_type_mode & OTMODE_JOURNAL)) {
                ff.skip_count++;
                DEBUG_INFO(("skipping journal entry: not in output type list\n"));
            }
            else {
                if (!ff.type) ff.type = item->type;
                if ((ff.type != PST_TYPE_JOURNAL) && (mode != MODE_SEPARATE)) {
                    ff.skip_count++;
                    DEBUG_INFO(("I have a journal entry, but the folder type %"PRIi32" isn't a journal folder. Skipping it\n", ff.type));
                }
                else {
                    ff.item_count++;
                    if (mode == MODE_SEPARATE) mk_separate_file(&ff, (mode_EX) ? ".ics" : "", 1);
                    write_journal(ff.output, item);
                    fprintf(ff.output, "\n");
                    if (mode == MODE_SEPARATE) close_separate_file(&ff);
                }
            }

        } else if (item->appointment && (item->type == PST_TYPE_APPOINTMENT)) {
            DEBUG_INFO(("Processing Appointment Entry\n"));
            if (!(output_type_mode & OTMODE_APPOINTMENT)) {
                ff.skip_count++;
                DEBUG_INFO(("skipping appointment: not in output type list\n"));
            }
            else {
                if (!ff.type) ff.type = item->type;
                if ((ff.type != PST_TYPE_APPOINTMENT) && (mode != MODE_SEPARATE)) {
                    ff.skip_count++;
                    DEBUG_INFO(("I have an appointment, but the folder type %"PRIi32" isn't an appointment folder. Skipping it\n", ff.type));
                }
                else {
                    ff.item_count++;
                    if (mode == MODE_SEPARATE) mk_separate_file(&ff, (mode_EX) ? ".ics" : "", 1);
                    write_schedule_part_data(ff.output, item, NULL, NULL);
                    fprintf(ff.output, "\n");
                    if (mode == MODE_SEPARATE) close_separate_file(&ff);
                }
            }

        } else if (item->message_store) {
            // there should only be one message_store, and we have already done it
            ff.skip_count++;
            DEBUG_WARN(("item with message store content, type %i %s folder type %i, skipping it\n", item->type, item->ascii_type, ff.type));

        } else {
            ff.skip_count++;
            DEBUG_WARN(("Unknown item type %i (%s) name (%s)\n",
                        item->type, item->ascii_type, item->file_as.str));
        }
        pst_freeItem(item);
    }
    close_enter_dir(&ff);
    DEBUG_RET();
}



int main(int argc, char* const* argv) {
    pst_item *item = NULL;
    pst_desc_tree *d_ptr;
    char * fname = NULL;
    char *d_log  = NULL;
    int c,x;
    char *temp = NULL;               //temporary char pointer
    prog_name = argv[0];

    time_t now = time(NULL);
    srand((unsigned)now);

    if (regcomp(&meta_charset_pattern, "<meta[^>]*content=\"[^>]*charset=([^>\";]*)[\";]", REG_ICASE | REG_EXTENDED)) {
        printf("cannot compile regex pattern to find content charset in html bodies\n");
        exit(3);
    }

    // command-line option handling
    while ((c = getopt(argc, argv, "a:bC:c:Dd:emhj:kMo:qrSt:uVwL:8"))!= -1) {
        switch (c) {
        case 'a':
            if (optarg) {
                int n = strlen(optarg);
                acceptable_extensions = (char*)pst_malloc(n+2);
                strcpy(acceptable_extensions, optarg);
                acceptable_extensions[n+1] = '\0';  // double null terminates array of non-empty null terminated strings.
                char *p = acceptable_extensions;
                while (*p) {
                    if (*p == ',') *p = '\0';
                    p++;
                }
            }
            break;
        case 'b':
            save_rtf_body = 0;
            break;
        case 'C':
            if (optarg) {
                default_charset = optarg;
            }
            else {
                usage();
                exit(0);
            }
            break;
        case 'c':
            if (optarg && optarg[0]=='v') {
                contact_mode=CMODE_VCARD;
                contact_mode_specified = 1;
            }
            else if (optarg && optarg[0]=='l') {
                contact_mode=CMODE_LIST;
                contact_mode_specified = 1;
            }
            else {
                usage();
                exit(0);
            }
            break;
        case 'D':
            deleted_mode = DMODE_INCLUDE;
            break;
        case 'd':
            d_log = optarg;
            break;
        case 'h':
            usage();
            exit(0);
            break;
        case 'j':
            max_children = atoi(optarg);
            max_child_specified = 1;
            break;
        case 'k':
            mode = MODE_KMAIL;
            break;
        case 'M':
            mode = MODE_SEPARATE;
            mode_MH  = 1;
            mode_EX  = 0;
            mode_MSG = 0;
            break;
        case 'e':
            mode = MODE_SEPARATE;
            mode_MH  = 1;
            mode_EX  = 1;
            mode_MSG = 0;
            file_name_len = 14;
            break;
        case 'L':
            pst_debug_setlevel(atoi(optarg));
            break;
        case 'm':
            mode = MODE_SEPARATE;
            mode_MH  = 1;
            mode_EX  = 1;
            mode_MSG = 1;
            file_name_len = 14;
            break;
        case 'o':
            output_dir = optarg;
            break;
        case 'q':
            output_mode = OUTPUT_QUIET;
            break;
        case 'r':
            mode = MODE_RECURSE;
            mode_thunder = 0;
            break;
        case 'S':
            mode = MODE_SEPARATE;
            mode_MH  = 0;
            mode_EX  = 0;
            mode_MSG = 0;
            break;
        case 't':
            // email, appointment, contact, other
            if (!optarg) {
                usage();
                exit(0);
            }
            temp = optarg;
            output_type_mode = 0;
            while (*temp > 0) {
              switch (temp[0]) {
                case 'e':
                    output_type_mode |= OTMODE_EMAIL;
                    break;
                case 'a':
                    output_type_mode |= OTMODE_APPOINTMENT;
                    break;
                case 'j':
                    output_type_mode |= OTMODE_JOURNAL;
                    break;
                case 'c':
                    output_type_mode |= OTMODE_CONTACT;
                    break;
                default:
                    usage();
                    exit(0);
                    break;
              }
              temp++;
            }
            break;
        case 'u':
            mode = MODE_RECURSE;
            mode_thunder = 1;
            break;
        case 'V':
            version();
            exit(0);
            break;
        case 'w':
            overwrite = 1;
            break;
        case '8':
            prefer_utf8 = 1;
            break;
        default:
            usage();
            exit(1);
            break;
        }
    }

    if (argc > optind) {
        fname = argv[optind];
    } else {
        usage();
        exit(2);
    }

#ifdef _SC_NPROCESSORS_ONLN
    number_processors =  sysconf(_SC_NPROCESSORS_ONLN);
#endif
    max_children    = (max_child_specified) ? max_children : number_processors * 4;
    active_children = 0;
    child_processes = (pid_t *)pst_malloc(sizeof(pid_t) * max_children);
    memset(child_processes, 0, sizeof(pid_t) * max_children);

#ifdef HAVE_SEMAPHORE_H
    if (max_children) {
        shared_memory_id = shmget(IPC_PRIVATE, sizeof(sem_t)*2, 0777);
        if (shared_memory_id >= 0) {
            global_children = (sem_t *)shmat(shared_memory_id, NULL, 0);
            if (global_children == (sem_t *)-1) global_children = NULL;
            if (global_children) {
                output_mutex = &(global_children[1]);
                sem_init(global_children, 1, max_children);
                sem_init(output_mutex, 1, 1);
            }
            shmctl(shared_memory_id, IPC_RMID, NULL);
        }
    }
#endif

    #ifdef DEBUG_ALL
        // force a log file
        if (!d_log) d_log = "readpst.log";
    #endif // defined DEBUG_ALL
    #ifdef HAVE_SEMAPHORE_H
        DEBUG_INIT(d_log, output_mutex);
    #else
        DEBUG_INIT(d_log, NULL);
    #endif
    DEBUG_ENT("main");

    if (output_mode != OUTPUT_QUIET) printf("Opening PST file and indexes...\n");
    RET_DERROR(pst_open(&pstfile, fname, default_charset), 1, ("Error opening File\n"));
    RET_DERROR(pst_load_index(&pstfile), 2, ("Index Error\n"));

    pst_load_extended_attributes(&pstfile);

    if (chdir(output_dir)) {
        x = errno;
        pst_close(&pstfile);
        DEBUG_RET();
        DIE(("Cannot change to output dir %s: %s\n", output_dir, strerror(x)));
    }

    d_ptr = pstfile.d_head; // first record is main record
    item  = pst_parse_item(&pstfile, d_ptr, NULL);
    if (!item || !item->message_store) {
        DEBUG_RET();
        DIE(("Could not get root record\n"));
    }

    // default the file_as to the same as the main filename if it doesn't exist
    if (!item->file_as.str) {
        if (!(temp = strrchr(fname, '/')))
            if (!(temp = strrchr(fname, '\\')))
                temp = fname;
            else
                temp++; // get past the "\\"
        else
            temp++; // get past the "/"
        item->file_as.str = (char*)pst_malloc(strlen(temp)+1);
        strcpy(item->file_as.str, temp);
        item->file_as.is_utf8 = 1;
        DEBUG_INFO(("file_as was blank, so am using %s\n", item->file_as.str));
    }
    DEBUG_INFO(("Root Folder Name: %s\n", item->file_as.str));

    d_ptr = pst_getTopOfFolders(&pstfile, item);
    if (!d_ptr) {
        DEBUG_RET();
        DIE(("Top of folders record not found. Cannot continue\n"));
    }

    process(item, d_ptr->child);    // do the children of TOPF
    grim_reaper(1); // wait for all child processes

    pst_freeItem(item);
    pst_close(&pstfile);
    DEBUG_RET();

#ifdef HAVE_SEMAPHORE_H
    if (global_children) {
        sem_destroy(global_children);
        sem_destroy(output_mutex);
        shmdt(global_children);
    }
#endif

    regfree(&meta_charset_pattern);
    return 0;
}


void write_email_body(FILE *f, char *body) {
    char *n = body;
    DEBUG_ENT("write_email_body");
    if (mode != MODE_SEPARATE) {
        while (n) {
            char *p = body;
            while (*p == '>') p++;
            if (strncmp(p, "From ", 5) == 0) fprintf(f, ">");
            if ((n = strchr(body, '\n'))) {
                n++;
                pst_fwrite(body, n-body, 1, f); //write just a line
                body = n;
            }
        }
    }
    pst_fwrite(body, strlen(body), 1, f);
    DEBUG_RET();
}


void removeCR (char *c) {
    // converts \r\n to \n
    char *a, *b;
    DEBUG_ENT("removeCR");
    a = b = c;
    while (*a != '\0') {
        *b = *a;
        if (*a != '\r') b++;
        a++;
    }
    *b = '\0';
    DEBUG_RET();
}


void usage() {
    DEBUG_ENT("usage");
    version();
    printf("Usage: %s [OPTIONS] {PST FILENAME}\n", prog_name);
    printf("OPTIONS:\n");
    printf("\t-V\t- Version. Display program version\n");
    printf("\t-C charset\t- character set for items with an unspecified character set\n");
    printf("\t-D\t- Include deleted items in output\n");
    printf("\t-L <level> \t- Set debug level; 1=debug,2=info,3=warn.\n");
    printf("\t-M\t- Write emails in the MH (rfc822) format\n");
    printf("\t-S\t- Separate. Write emails in the separate format\n");
    printf("\t-a <attachment-extension-list>\t- Discard any attachment without an extension on the list\n");
    printf("\t-b\t- Don't save RTF-Body attachments\n");
    printf("\t-c[v|l]\t- Set the Contact output mode. -cv = VCard, -cl = EMail list\n");
    printf("\t-d <filename> \t- Debug to file.\n");
    printf("\t-e\t- As with -M, but include extensions on output files\n");
    printf("\t-h\t- Help. This screen\n");
    printf("\t-j <integer>\t- Number of parallel jobs to run\n");
    printf("\t-k\t- KMail. Output in kmail format\n");
    printf("\t-m\t- As with -e, but write .msg files also\n");
    printf("\t-o <dirname>\t- Output directory to write files to. CWD is changed *after* opening pst file\n");
    printf("\t-q\t- Quiet. Only print error messages\n");
    printf("\t-r\t- Recursive. Output in a recursive format\n");
    printf("\t-t[eajc]\t- Set the output type list. e = email, a = attachment, j = journal, c = contact\n");
    printf("\t-u\t- Thunderbird mode. Write two extra .size and .type files\n");
    printf("\t-w\t- Overwrite any output mbox files\n");
    printf("\t-8\t- Output bodies in UTF-8, rather than original encoding, if UTF-8 version is available\n");
    printf("\n");
    printf("Only one of -M -S -e -k -m -r should be specified\n");
    DEBUG_RET();
}


void version() {
    DEBUG_ENT("version");
    printf("ReadPST / LibPST v%s\n", VERSION);
#if BYTE_ORDER == BIG_ENDIAN
    printf("Big Endian implementation being used.\n");
#elif BYTE_ORDER == LITTLE_ENDIAN
    printf("Little Endian implementation being used.\n");
#else
#  error "Byte order not supported by this library"
#endif
    DEBUG_RET();
}


char *mk_kmail_dir(char *fname) {
    //change to that directory
    //make a directory based on OUTPUT_KMAIL_DIR_TEMPLATE
    //allocate space for OUTPUT_TEMPLATE and form a char* with fname
    //return that value
    char *dir, *out_name, *index;
    int x;
    DEBUG_ENT("mk_kmail_dir");
    if (kmail_chdir && chdir(kmail_chdir)) {
        x = errno;
        DIE(("mk_kmail_dir: Cannot change to directory %s: %s\n", kmail_chdir, strerror(x)));
    }
    dir = pst_malloc(strlen(fname)+strlen(OUTPUT_KMAIL_DIR_TEMPLATE)+1);
    sprintf(dir, OUTPUT_KMAIL_DIR_TEMPLATE, fname);
    check_filename(dir);
    if (D_MKDIR(dir)) {
        if (errno != EEXIST) {  // not an error because it exists
            x = errno;
            DIE(("mk_kmail_dir: Cannot create directory %s: %s\n", dir, strerror(x)));
        }
    }
    kmail_chdir = pst_realloc(kmail_chdir, strlen(dir)+1);
    strcpy(kmail_chdir, dir);
    free (dir);

    //we should remove any existing indexes created by KMail, cause they might be different now
    index = pst_malloc(strlen(fname)+strlen(KMAIL_INDEX)+1);
    sprintf(index, KMAIL_INDEX, fname);
    unlink(index);
    free(index);

    out_name = pst_malloc(strlen(fname)+strlen(OUTPUT_TEMPLATE)+1);
    sprintf(out_name, OUTPUT_TEMPLATE, fname);
    DEBUG_RET();
    return out_name;
}


int close_kmail_dir() {
    // change ..
    int x;
    DEBUG_ENT("close_kmail_dir");
    if (kmail_chdir) { //only free kmail_chdir if not NULL. do not change directory
        free(kmail_chdir);
        kmail_chdir = NULL;
    } else {
        if (chdir("..")) {
            x = errno;
            DIE(("close_kmail_dir: Cannot move up dir (..): %s\n", strerror(x)));
        }
    }
    DEBUG_RET();
    return 0;
}


// this will create a directory by that name,
// then make an mbox file inside that directory.
char *mk_recurse_dir(char *dir, int32_t folder_type) {
    int x;
    char *out_name;
    DEBUG_ENT("mk_recurse_dir");
    check_filename(dir);
    if (D_MKDIR (dir)) {
        if (errno != EEXIST) {  // not an error because it exists
            x = errno;
            DIE(("mk_recurse_dir: Cannot create directory %s: %s\n", dir, strerror(x)));
        }
    }
    if (chdir (dir)) {
        x = errno;
        DIE(("mk_recurse_dir: Cannot change to directory %s: %s\n", dir, strerror(x)));
    }
    switch (folder_type) {
        case PST_TYPE_APPOINTMENT:
            out_name = strdup("calendar");
            break;
        case PST_TYPE_CONTACT:
            out_name = strdup("contacts");
            break;
        case PST_TYPE_JOURNAL:
            out_name = strdup("journal");
            break;
        case PST_TYPE_STICKYNOTE:
        case PST_TYPE_TASK:
        case PST_TYPE_NOTE:
        case PST_TYPE_OTHER:
        case PST_TYPE_REPORT:
        default:
            out_name = strdup("mbox");
            break;
    }
    DEBUG_RET();
    return out_name;
}


int close_recurse_dir() {
    int x;
    DEBUG_ENT("close_recurse_dir");
    if (chdir("..")) {
        x = errno;
        DIE(("close_recurse_dir: Cannot go up dir (..): %s\n", strerror(x)));
    }
    DEBUG_RET();
    return 0;
}


char *mk_separate_dir(char *dir) {
    size_t dirsize = strlen(dir) + 10;
    char dir_name[dirsize];
    int x = 0, y = 0;

    DEBUG_ENT("mk_separate_dir");
    do {
        if (y == 0)
            snprintf(dir_name, dirsize, "%s", dir);
        else
            snprintf(dir_name, dirsize, "%s" SEP_MAIL_FILE_TEMPLATE, dir, y, ""); // enough for 9 digits allocated above

        check_filename(dir_name);
        DEBUG_INFO(("about to try creating %s\n", dir_name));
        if (D_MKDIR(dir_name)) {
            if (errno != EEXIST) { // if there is an error, and it doesn't already exist
                x = errno;
                DIE(("mk_separate_dir: Cannot create directory %s: %s\n", dir, strerror(x)));
            }
        } else {
            break;
        }
        y++;
    } while (overwrite == 0);

    if (chdir(dir_name)) {
        x = errno;
        DIE(("mk_separate_dir: Cannot change to directory %s: %s\n", dir, strerror(x)));
    }

    if (overwrite) {
        // we should probably delete all files from this directory
#if !defined(WIN32) && !defined(__CYGWIN__)
        DIR * sdir = NULL;
        struct dirent *dirent = NULL;
        struct stat filestat;
        if (!(sdir = opendir("./"))) {
            DEBUG_WARN(("mk_separate_dir: Cannot open dir \"%s\" for deletion of old contents\n", "./"));
        } else {
            while ((dirent = readdir(sdir))) {
                if (lstat(dirent->d_name, &filestat) != -1)
                    if (S_ISREG(filestat.st_mode)) {
                        if (unlink(dirent->d_name)) {
                            y = errno;
                            DIE(("mk_separate_dir: unlink returned error on file %s: %s\n", dirent->d_name, strerror(y)));
                        }
                    }
            }
            closedir(sdir);     // cppcheck detected leak
        }
#endif
    }

    // we don't return a filename here cause it isn't necessary.
    DEBUG_RET();
    return NULL;
}


int close_separate_dir() {
    int x;
    DEBUG_ENT("close_separate_dir");
    if (chdir("..")) {
        x = errno;
        DIE(("close_separate_dir: Cannot go up dir (..): %s\n", strerror(x)));
    }
    DEBUG_RET();
    return 0;
}


void mk_separate_file(struct file_ll *f, char *extension, int openit) {
    DEBUG_ENT("mk_separate_file");
    DEBUG_INFO(("opening next file to save email\n"));
    if (f->item_count > 999999999) { // bigger than nine 9's
        DIE(("mk_separate_file: The number of emails in this folder has become too high to handle\n"));
    }
    sprintf(f->name, SEP_MAIL_FILE_TEMPLATE, f->item_count, extension);
    check_filename(f->name);
    if (openit) {
        if (!(f->output = fopen(f->name, "w"))) {
            DIE(("mk_separate_file: Cannot open file to save email \"%s\"\n", f->name));
        }
    }
    DEBUG_RET();
}


void close_separate_file(struct file_ll *f) {
    DEBUG_ENT("close_separate_file");
    if (f->output) {
        struct stat st;
        fclose(f->output);
        stat(f->name, &st);
        if (!st.st_size) {
            DEBUG_WARN(("removing empty output file %s\n", f->name));
            remove(f->name);
        }
        f->output = NULL;
    }
    DEBUG_RET();
}


char *my_stristr(char *haystack, char *needle) {
    // my_stristr varies from strstr in that its searches are case-insensitive
    char *x=haystack, *y=needle, *z = NULL;
    if (!haystack || !needle) {
        return NULL;
    }
    while (*y != '\0' && *x != '\0') {
        if (tolower(*y) == tolower(*x)) {
            // move y on one
            y++;
            if (!z) {
                z = x; // store first position in haystack where a match is made
            }
        } else {
            y = needle; // reset y to the beginning of the needle
            z = NULL; // reset the haystack storage point
        }
        x++; // advance the search in the haystack
    }
    // If the haystack ended before our search finished, it's not a match.
    if (*y != '\0') return NULL;
    return z;
}


void check_filename(char *fname) {
    char *t = fname;
    DEBUG_ENT("check_filename");
    if (!t) {
        DEBUG_RET();
        return;
    }
    while ((t = strpbrk(t, "/\\:"))) {
        // while there are characters in the second string that we don't want
        *t = '_'; //replace them with an underscore
    }
    DEBUG_RET();
}


/**
 * check if the file name extension is acceptable. If not, the attachment
 * will be discarded
 * @param attach  pst attachment object
 * @return        true if the attachment filename contains an extension that we want.
 */
int  acceptable_ext(pst_item_attach* attach)
{
    if (!acceptable_extensions || *acceptable_extensions == '\0') return 1;     // acceptable list missing or empty
    char *attach_filename = (attach->filename2.str) ? attach->filename2.str
                                                    : attach->filename1.str;
    if (!attach_filename) return 1; // attachment with no name is always acceptable
    char *e = strrchr(attach_filename, '.');
    if (!e) return 1;               // attachment with no extension is always acceptable.
    DEBUG_ENT("acceptable_ext");
    DEBUG_INFO(("attachment extension %s\n", e));
    int rc = 0;
    char *a = acceptable_extensions;
    while (*a) {
        if (pst_stricmp(a, e) == 0) {
            rc = 1;
            break;
        }
        a += strlen(a) + 1;
    }
    DEBUG_INFO(("attachment acceptable returns %d\n", rc));
    DEBUG_RET();
    return rc;
}


void write_separate_attachment(char f_name[], pst_item_attach* attach, int attach_num, pst_file* pst)
{
    FILE *fp = NULL;
    int x = 0;
    char *temp = NULL;

    // If there is a long filename (filename2) use that, otherwise
    // use the 8.3 filename (filename1)
    char *attach_filename = (attach->filename2.str) ? attach->filename2.str
                                                    : attach->filename1.str;
    DEBUG_ENT("write_separate_attachment");
    DEBUG_INFO(("Attachment %s Size is %#"PRIx64", data = %#"PRIxPTR", id %#"PRIx64"\n", attach_filename, (uint64_t)attach->data.size, attach->data.data, attach->i_id));

    if (!attach->data.data) {
        // make sure we can fetch data from the id
        pst_index_ll *ptr = pst_getID(pst, attach->i_id);
        if (!ptr) {
            DEBUG_WARN(("Couldn't find i_id %#"PRIx64". Cannot save attachment to file\n", attach->i_id));
            DEBUG_RET();
            return;
        }
    }

    check_filename(f_name);
    if (!attach_filename) {
        // generate our own (dummy) filename for the attachement
        temp = pst_malloc(strlen(f_name)+15);
        sprintf(temp, "%s-attach%i", f_name, attach_num);
    } else {
        // have an attachment name, make sure it's unique
        temp = pst_malloc(strlen(f_name)+strlen(attach_filename)+15);
        do {
            if (fp) fclose(fp);
            if (x == 0)
                sprintf(temp, "%s-%s", f_name, attach_filename);
            else
                sprintf(temp, "%s-%s-%i", f_name, attach_filename, x);
        } while ((fp = fopen(temp, "r")) && ++x < 99999999);
        if (x > 99999999) {
            DIE(("error finding attachment name. exhausted possibilities to %s\n", temp));
        }
    }
    DEBUG_INFO(("Saving attachment to %s\n", temp));
    if (!(fp = fopen(temp, "w"))) {
        DEBUG_WARN(("write_separate_attachment: Cannot open attachment save file \"%s\"\n", temp));
    } else {
        (void)pst_attach_to_file(pst, attach, fp);
        fclose(fp);
    }
    if (temp) free(temp);
    DEBUG_RET();
}


void write_embedded_message(FILE* f_output, pst_item_attach* attach, char *boundary, pst_file* pf, int save_rtf, char** extra_mime_headers)
{
    pst_index_ll *ptr;
    DEBUG_ENT("write_embedded_message");
    ptr = pst_getID(pf, attach->i_id);

    pst_desc_tree d_ptr;
    d_ptr.d_id        = 0;
    d_ptr.parent_d_id = 0;
    d_ptr.assoc_tree  = NULL;
    d_ptr.desc        = ptr;
    d_ptr.no_child    = 0;
    d_ptr.prev        = NULL;
    d_ptr.next        = NULL;
    d_ptr.parent      = NULL;
    d_ptr.child       = NULL;
    d_ptr.child_tail  = NULL;

    pst_item *item = pst_parse_item(pf, &d_ptr, attach->id2_head);
    // It appears that if the embedded message contains an appointment/
    // calendar item, pst_parse_item returns NULL due to the presence of
    // an unexpected reference type of 0x1048, which seems to represent
    // an array of GUIDs representing a CLSID. It's likely that this is
    // a reference to an internal Outlook COM class.
    //      Log the skipped item and continue on.
    if (!item) {
        DEBUG_WARN(("write_embedded_message: pst_parse_item was unable to parse the embedded message in attachment ID %llu", attach->i_id));
    } else {
        if (!item->email) {
            DEBUG_WARN(("write_embedded_message: pst_parse_item returned type %d, not an email message", item->type));
        } else {
            fprintf(f_output, "\n--%s\n", boundary);
            fprintf(f_output, "Content-Type: %s\n\n", attach->mimetype.str);
            write_normal_email(f_output, "", item, MODE_NORMAL, 0, pf, save_rtf, 1, extra_mime_headers);
        }
        pst_freeItem(item);
    }

    DEBUG_RET();
}


void write_inline_attachment(FILE* f_output, pst_item_attach* attach, char *boundary, pst_file* pst)
{
    DEBUG_ENT("write_inline_attachment");
    DEBUG_INFO(("Attachment Size is %#"PRIx64", data = %#"PRIxPTR", id %#"PRIx64"\n", (uint64_t)attach->data.size, attach->data.data, attach->i_id));

    if (!attach->data.data) {
        // make sure we can fetch data from the id
        pst_index_ll *ptr = pst_getID(pst, attach->i_id);
        if (!ptr) {
            DEBUG_WARN(("Couldn't find ID pointer. Cannot save attachment to file\n"));
            DEBUG_RET();
            return;
        }
    }

    fprintf(f_output, "\n--%s\n", boundary);
    if (!attach->mimetype.str) {
        fprintf(f_output, "Content-Type: %s\n", MIME_TYPE_DEFAULT);
    } else {
        fprintf(f_output, "Content-Type: %s\n", attach->mimetype.str);
    }
    fprintf(f_output, "Content-Transfer-Encoding: base64\n");

    if (attach->content_id.str) {
        fprintf(f_output, "Content-ID: <%s>\n", attach->content_id.str);
    }

    if (attach->filename2.str) {
        // use the long filename, converted to proper encoding if needed.
        // it is already utf8
        pst_rfc2231(&attach->filename2);
        fprintf(f_output, "Content-Disposition: attachment; \n        filename*=%s\n\n", attach->filename2.str);
    }
    else if (attach->filename1.str) {
        // short filename never needs encoding
        fprintf(f_output, "Content-Disposition: attachment; filename=\"%s\"\n\n", attach->filename1.str);
    }
    else {
        // no filename is inline
        fprintf(f_output, "Content-Disposition: inline\n\n");
    }

    (void)pst_attach_to_file_base64(pst, attach, f_output);
    fprintf(f_output, "\n\n");
    DEBUG_RET();
}


int  valid_headers(char *header)
{
    // headers are sometimes really bogus - they seem to be fragments of the
    // message body, so we only use them if they seem to be real rfc822 headers.
    // this list is composed of ones that we have seen in real pst files.
    // there are surely others. the problem is - given an arbitrary character
    // string, is it a valid (or even reasonable) set of rfc822 headers?
    if (header) {
        if ((strncasecmp(header, "X-Barracuda-URL: ", 17) == 0) ||
            (strncasecmp(header, "X-ASG-Debug-ID: ",  16) == 0) ||
            (strncasecmp(header, "Return-Path: ",     13) == 0) ||
            (strncasecmp(header, "Received: ",        10) == 0) ||
            (strncasecmp(header, "Subject: ",          9) == 0) ||
            (strncasecmp(header, "Date: ",             6) == 0) ||
            (strncasecmp(header, "From: ",             6) == 0) ||
            (strncasecmp(header, "X-x: ",              5) == 0) ||
            (strncasecmp(header, "Microsoft Mail Internet Headers", 31) == 0)) {
            return 1;
        }
        else {
            if (strlen(header) > 2) {
                DEBUG_INFO(("Ignore bogus headers = %s\n", header));
            }
            return 0;
        }
    }
    else return 0;
}


void header_has_field(char *header, char *field, int *flag)
{
    DEBUG_ENT("header_has_field");
    if (my_stristr(header, field) || (strncasecmp(header, field+1, strlen(field)-1) == 0)) {
        DEBUG_INFO(("header block has %s header\n", field+1));
        *flag = 1;
    }
    DEBUG_RET();
}


void header_get_subfield(char *field, const char *subfield, char *body_subfield, size_t size_subfield)
{
    if (!field) return;
    DEBUG_ENT("header_get_subfield");
    char search[60];
    snprintf(search, sizeof(search), " %s=", subfield);
    field++;
    char *n = header_end_field(field);
    char *s = my_stristr(field, search);
    if (n && s && (s < n)) {
        char *e, *f, save;
        s += strlen(search);    // skip over subfield=
        if (*s == '"') {
            s++;
            e = strchr(s, '"');
        }
        else {
            e = strchr(s, ';');
            f = strchr(s, '\n');
            if (e && f && (f < e)) e = f;
        }
        if (!e || (e > n)) e = n;   // use the trailing lf as terminator if nothing better
        save = *e;
        *e = '\0';
        snprintf(body_subfield, size_subfield, "%s", s);  // copy the subfield to our buffer
        *e = save;
        DEBUG_INFO(("body %s %s from headers\n", subfield, body_subfield));
    }
    DEBUG_RET();
}

char* header_get_field(char *header, char *field)
{
    char *t = my_stristr(header, field);
    if (!t && (strncasecmp(header, field+1, strlen(field)-1) == 0)) t = header;
    return t;
}


// return pointer to \n at the end of this header field,
// or NULL if this field goes to the end of the string.
char *header_end_field(char *field)
{
    char *e = strchr(field+1, '\n');
    while (e && ((e[1] == ' ') || (e[1] == '\t'))) {
        e = strchr(e+1, '\n');
    }
    return e;
}


void header_strip_field(char *header, char *field)
{
    char *t;
    while ((t = header_get_field(header, field))) {
        char *e = header_end_field(t);
        if (e) {
            if (t == header) e++;   // if *t is not \n, we don't want to keep the \n at *e either.
            while (*e != '\0') {
                *t = *e;
                t++;
                e++;
            }
            *t = '\0';
        }
        else {
            // this was the last header field, truncate the headers
            *t = '\0';
        }
    }
}


int  test_base64(char *body, size_t len)
{
    int b64 = 0;
    uint8_t *b = (uint8_t *)body;
    DEBUG_ENT("test_base64");
    while (len--) {
        if ((*b < 32) && (*b != 9) && (*b != 10)) {
            DEBUG_INFO(("found base64 byte %d\n", (int)*b));
            DEBUG_HEXDUMPC(body, strlen(body), 0x10);
            b64 = 1;
            break;
        }
        b++;
    }
    DEBUG_RET();
    return b64;
}


void find_html_charset(char *html, char *charset, size_t charsetlen)
{
    const int  index = 1;
    const int nmatch = index+1;
    regmatch_t match[nmatch];
    DEBUG_ENT("find_html_charset");
    int rc = regexec(&meta_charset_pattern, html, nmatch, match, 0);
    if (rc == 0) {
        int s = match[index].rm_so;
        int e = match[index].rm_eo;
        if (s != -1) {
            char save = html[e];
            html[e] = '\0';
                snprintf(charset, charsetlen, "%s", html+s);    // copy the html charset
            html[e] = save;
            DEBUG_INFO(("charset %s from html text\n", charset));
        }
        else {
            DEBUG_INFO(("matching %d %d %d %d\n", match[0].rm_so, match[0].rm_eo, match[1].rm_so, match[1].rm_eo));
            DEBUG_HEXDUMPC(html, strlen(html), 0x10);
        }
    }
    else {
        DEBUG_INFO(("regexec returns %d\n", rc));
    }
    DEBUG_RET();
}


void find_rfc822_headers(char** extra_mime_headers)
{
    DEBUG_ENT("find_rfc822_headers");
    char *headers = *extra_mime_headers;
    if (headers) {
        char *temp, *t;
        while ((temp = strstr(headers, "\n\n"))) {
            temp[1] = '\0';
            t = header_get_field(headers, "\nContent-Type:");
            if (t) {
                t++;
                DEBUG_INFO(("found content type header\n"));
                char *n = strchr(t, '\n');
                char *s = strstr(t, ": ");
                char *e = strchr(t, ';');
                if (!e || (e > n)) e = n;
                if (s && (s < e)) {
                    s += 2;
                    if (!strncasecmp(s, RFC822, e-s)) {
                        headers = temp+2;   // found rfc822 header
                        DEBUG_INFO(("found 822 headers\n%s\n", headers));
                        break;
                    }
                }
            }
            //DEBUG_INFO(("skipping to next block after\n%s\n", headers));
            headers = temp+2;   // skip to next chunk of headers
        }
        *extra_mime_headers = headers;
    }
    DEBUG_RET();
}


void write_body_part(FILE* f_output, pst_string *body, char *mime, char *charset, char *boundary, pst_file* pst)
{
    DEBUG_ENT("write_body_part");
    removeCR(body->str);
    size_t body_len = strlen(body->str);

    if (body->is_utf8 && (strcasecmp("utf-8", charset))) {
        if (prefer_utf8) {
            charset = "utf-8";
        } else {
            // try to convert to the specified charset since the target
            // is not utf-8, and the data came from a unicode (utf16) field
            // and is now in utf-8.
            size_t rc;
            DEBUG_INFO(("Convert %s utf-8 to %s\n", mime, charset));
            pst_vbuf *newer = pst_vballoc(2);
            rc = pst_vb_utf8to8bit(newer, body->str, body_len, charset);
            if (rc == (size_t)-1) {
                // unable to convert, change the charset to utf8
                free(newer->b);
                DEBUG_INFO(("Failed to convert %s utf-8 to %s\n", mime, charset));
                charset = "utf-8";
            } else {
                // null terminate the output string
                pst_vbgrow(newer, 1);
                newer->b[newer->dlen] = '\0';
                free(body->str);
                body->str = newer->b;
                body_len = newer->dlen;
            }
            free(newer);
        }
    }
    int base64 = test_base64(body->str, body_len);
    fprintf(f_output, "\n--%s\n", boundary);
    fprintf(f_output, "Content-Type: %s; charset=\"%s\"\n", mime, charset);
    if (base64) fprintf(f_output, "Content-Transfer-Encoding: base64\n");
    fprintf(f_output, "\n");
    // Any body that uses an encoding with NULLs, e.g. UTF16, will be base64-encoded here.
    if (base64) {
        char *enc = pst_base64_encode(body->str, body_len);
        if (enc) {
            write_email_body(f_output, enc);
            fprintf(f_output, "\n");
            free(enc);
        }
    }
    else {
        write_email_body(f_output, body->str);
    }
    DEBUG_RET();
}


void write_schedule_part_data(FILE* f_output, pst_item* item, const char* sender, const char* method)
{
    fprintf(f_output, "BEGIN:VCALENDAR\n");
    fprintf(f_output, "VERSION:2.0\n");
    fprintf(f_output, "PRODID:LibPST v%s\n", VERSION);
    if (method) fprintf(f_output, "METHOD:%s\n", method);
    fprintf(f_output, "BEGIN:VEVENT\n");
    if (sender) {
        if (item->email->outlook_sender_name.str) {
            fprintf(f_output, "ORGANIZER;CN=\"%s\":MAILTO:%s\n", item->email->outlook_sender_name.str, sender);
        } else {
            fprintf(f_output, "ORGANIZER;CN=\"\":MAILTO:%s\n", sender);
        }
    }
    write_appointment(f_output, item);
    fprintf(f_output, "END:VCALENDAR\n");
}


void write_schedule_part(FILE* f_output, pst_item* item, const char* sender, const char* boundary)
{
    const char* method  = "REQUEST";
    const char* charset = "utf-8";
    char fname[30];
    if (!item->appointment) return;

    // inline appointment request
    fprintf(f_output, "\n--%s\n", boundary);
    fprintf(f_output, "Content-Type: %s; method=\"%s\"; charset=\"%s\"\n\n", "text/calendar", method, charset);
    write_schedule_part_data(f_output, item, sender, method);
    fprintf(f_output, "\n");

    // attachment appointment request
    snprintf(fname, sizeof(fname), "i%i.ics", rand());
    fprintf(f_output, "\n--%s\n", boundary);
    fprintf(f_output, "Content-Type: %s; charset=\"%s\"; name=\"%s\"\n", "text/calendar", "utf-8", fname);
    fprintf(f_output, "Content-Disposition: attachment; filename=\"%s\"\n\n", fname);
    write_schedule_part_data(f_output, item, sender, method);
    fprintf(f_output, "\n");
}


void write_normal_email(FILE* f_output, char f_name[], pst_item* item, int mode, int mode_MH, pst_file* pst, int save_rtf, int embedding, char** extra_mime_headers)
{
    char boundary[60];
    char altboundary[66];
    char *altboundaryp = NULL;
    char body_charset[30];
    char buffer_charset[30];
    char body_report[60];
    char sender[60];
    int  sender_known = 0;
    char *temp = NULL;
    time_t em_time;
    char *c_time;
    char *headers = NULL;
    int has_from, has_subject, has_to, has_cc, has_date, has_msgid;
    has_from = has_subject = has_to = has_cc = has_date = has_msgid = 0;
    DEBUG_ENT("write_normal_email");

    pst_convert_utf8_null(item, &item->email->header);
    headers = valid_headers(item->email->header.str) ? item->email->header.str :
              valid_headers(*extra_mime_headers)     ? *extra_mime_headers     :
              NULL;

    // setup default body character set and report type
    strncpy(body_charset, pst_default_charset(item, sizeof(buffer_charset), buffer_charset), sizeof(body_charset));
    body_charset[sizeof(body_charset)-1] = '\0';
    strncpy(body_report, "delivery-status", sizeof(body_report));
    body_report[sizeof(body_report)-1] = '\0';

    // setup default sender
    pst_convert_utf8(item, &item->email->sender_address);
    if (item->email->sender_address.str && strchr(item->email->sender_address.str, '@')) {
        temp = item->email->sender_address.str;
        sender_known = 1;
    }
    else {
        temp = "MAILER-DAEMON";
    }
    strncpy(sender, temp, sizeof(sender));
    sender[sizeof(sender)-1] = '\0';

    // convert the sent date if it exists, or set it to a fixed date
    if (item->email->sent_date) {
        em_time = pst_fileTimeToUnixTime(item->email->sent_date);
        c_time = ctime(&em_time);
        if (c_time)
            c_time[strlen(c_time)-1] = '\0'; //remove end \n
        else
            c_time = "Thu Jan 1 00:00:00 1970";
    } else
        c_time = "Thu Jan 1 00:00:00 1970";

    // create our MIME boundaries here.
    snprintf(boundary, sizeof(boundary), "--boundary-LibPST-iamunique-%i_-_-", rand());
    snprintf(altboundary, sizeof(altboundary), "alt-%s", boundary);

    // we will always look at the headers to discover some stuff
    if (headers ) {
        char *t;
        removeCR(headers);

        temp = strstr(headers, "\n\n");
        if (temp) {
            // cut off our real rfc822 headers here
            temp[1] = '\0';
            // pointer to all the embedded MIME headers.
            // we use these to find the actual rfc822 headers for embedded message/rfc822 mime parts
            // but only for the outermost message
            if (!*extra_mime_headers) *extra_mime_headers = temp+2;
            DEBUG_INFO(("Found extra mime headers\n%s\n", temp+2));
        }

        // Check if the headers have all the necessary fields
        header_has_field(headers, "\nFrom:",        &has_from);
        header_has_field(headers, "\nTo:",          &has_to);
        header_has_field(headers, "\nSubject:",     &has_subject);
        header_has_field(headers, "\nDate:",        &has_date);
        header_has_field(headers, "\nCC:",          &has_cc);
        header_has_field(headers, "\nMessage-Id:",  &has_msgid);

        // look for charset and report-type in Content-Type header
        t = header_get_field(headers, "\nContent-Type:");
        header_get_subfield(t, "charset", body_charset, sizeof(body_charset));
        header_get_subfield(t, "report-type", body_report, sizeof(body_report));

        // derive a proper sender email address
        if (!sender_known) {
            t = header_get_field(headers, "\nFrom:");
            if (t) {
                // assume address is on the first line, rather than on a continuation line
                t++;
                char *n = strchr(t, '\n');
                char *s = strchr(t, '<');
                char *e = strchr(t, '>');
                if (s && e && n && (s < e) && (e < n)) {
                char save = *e;
                *e = '\0';
                    snprintf(sender, sizeof(sender), "%s", s+1);
                *e = save;
                }
            }
        }

        // Strip out the mime headers and some others that we don't want to emit
        header_strip_field(headers, "\nMicrosoft Mail Internet Headers");
        header_strip_field(headers, "\nMIME-Version:");
        header_strip_field(headers, "\nContent-Type:");
        header_strip_field(headers, "\nContent-Transfer-Encoding:");
        header_strip_field(headers, "\nContent-class:");
        header_strip_field(headers, "\nX-MimeOLE:");
        header_strip_field(headers, "\nX-From_:");
    }

    DEBUG_INFO(("About to print Header\n"));

    if (item && item->subject.str) {
        pst_convert_utf8(item, &item->subject);
        DEBUG_INFO(("item->subject = %s\n", item->subject.str));
    }

    if (mode != MODE_SEPARATE) {
        // most modes need this separator line.
        // procmail produces this separator without the quotes around the
        // sender email address, but apparently some Mac email client needs
        // those quotes, and they don't seem to cause problems for anyone else.
        char *quo = (embedding) ? ">" : "";
        fprintf(f_output, "%sFrom \"%s\" %s\n", quo, sender, c_time);
    }

    // print the supplied email headers
    if (headers) {
        int len = strlen(headers);
        if (len > 0) {
            fprintf(f_output, "%s", headers);
            // make sure the headers end with a \n
            if (headers[len-1] != '\n') fprintf(f_output, "\n");
            //char *h = headers;
            //while (*h) {
            //    char *e = strchr(h, '\n');
            //    int   d = 1;    // normally e points to trailing \n
            //    if (!e) {
            //        e = h + strlen(h);  // e points to trailing null
            //        d = 0;
            //    }
            //    // we could do rfc2047 encoding here if needed
            //    fprintf(f_output, "%.*s\n", (int)(e-h), h);
            //    h = e + d;
            //}
        }
    }

    // record read status
    if ((item->flags & PST_FLAG_READ) == PST_FLAG_READ) {
        fprintf(f_output, "Status: RO\n");
    }

    // create required header fields that are not already written

    if (!has_from) {
        if (item->email->outlook_sender_name.str){
            pst_rfc2047(item, &item->email->outlook_sender_name, 1);
            fprintf(f_output, "From: %s <%s>\n", item->email->outlook_sender_name.str, sender);
        } else {
            fprintf(f_output, "From: <%s>\n", sender);
        }
    }

    if (!has_subject) {
        if (item->subject.str) {
            pst_rfc2047(item, &item->subject, 0);
            fprintf(f_output, "Subject: %s\n", item->subject.str);
        } else {
            fprintf(f_output, "Subject: \n");
        }
    }

    if (!has_to && item->email->sentto_address.str) {
        pst_rfc2047(item, &item->email->sentto_address, 0);
        fprintf(f_output, "To: %s\n", item->email->sentto_address.str);
    }

    if (!has_cc && item->email->cc_address.str) {
        pst_rfc2047(item, &item->email->cc_address, 0);
        fprintf(f_output, "Cc: %s\n", item->email->cc_address.str);
    }

    if (!has_date && item->email->sent_date) {
        char c_time[C_TIME_SIZE];
        struct tm stm;
        gmtime_r(&em_time, &stm);
        strftime(c_time, C_TIME_SIZE, "%a, %d %b %Y %H:%M:%S %z", &stm);
        fprintf(f_output, "Date: %s\n", c_time);
    }

    if (!has_msgid && item->email->messageid.str) {
        pst_convert_utf8(item, &item->email->messageid);
        fprintf(f_output, "Message-Id: %s\n", item->email->messageid.str);
    }

    // add forensic headers to capture some .pst stuff that is not really
    // needed or used by mail clients
    pst_convert_utf8_null(item, &item->email->sender_address);
    if (item->email->sender_address.str && !strchr(item->email->sender_address.str, '@')
                                        && strcmp(item->email->sender_address.str, ".")
                                        && (strlen(item->email->sender_address.str) > 0)) {
        fprintf(f_output, "X-libpst-forensic-sender: %s\n", item->email->sender_address.str);
    }

    if (item->email->bcc_address.str) {
        pst_convert_utf8(item, &item->email->bcc_address);
        fprintf(f_output, "X-libpst-forensic-bcc: %s\n", item->email->bcc_address.str);
    }

    // add our own mime headers
    fprintf(f_output, "MIME-Version: 1.0\n");
    if (item->type == PST_TYPE_REPORT) {
        // multipart/report for DSN/MDN reports
        fprintf(f_output, "Content-Type: multipart/report; report-type=%s;\n\tboundary=\"%s\"\n", body_report, boundary);
    }
    else {
        fprintf(f_output, "Content-Type: multipart/mixed;\n\tboundary=\"%s\"\n", boundary);
    }
    fprintf(f_output, "\n");    // end of headers, start of body

    // now dump the body parts
    if ((item->type == PST_TYPE_REPORT) && (item->email->report_text.str)) {
        write_body_part(f_output, &item->email->report_text, "text/plain", body_charset, boundary, pst);
        fprintf(f_output, "\n");
    }

    if (item->body.str && item->email->htmlbody.str) {
        // start the nested alternative part
        fprintf(f_output, "\n--%s\n", boundary);
        fprintf(f_output, "Content-Type: multipart/alternative;\n\tboundary=\"%s\"\n", altboundary);
        altboundaryp = altboundary;
    }
    else {
        altboundaryp = boundary;
    }

    if (item->body.str) {
        write_body_part(f_output, &item->body, "text/plain", body_charset, altboundaryp, pst);
    }

    if (item->email->htmlbody.str) {
        find_html_charset(item->email->htmlbody.str, body_charset, sizeof(body_charset));
        write_body_part(f_output, &item->email->htmlbody, "text/html", body_charset, altboundaryp, pst);
    }

    if (item->body.str && item->email->htmlbody.str) {
        // end the nested alternative part
        fprintf(f_output, "\n--%s--\n", altboundary);
    }

    if (item->email->rtf_compressed.data && save_rtf) {
        pst_item_attach* attach = (pst_item_attach*)pst_malloc(sizeof(pst_item_attach));
        DEBUG_INFO(("Adding RTF body as attachment\n"));
        memset(attach, 0, sizeof(pst_item_attach));
        attach->next = item->attach;
        item->attach = attach;
        attach->data.data         = pst_lzfu_decompress(item->email->rtf_compressed.data, item->email->rtf_compressed.size, &attach->data.size);
        attach->filename2.str     = strdup(RTF_ATTACH_NAME);
        attach->filename2.is_utf8 = 1;
        attach->mimetype.str      = strdup(RTF_ATTACH_TYPE);
        attach->mimetype.is_utf8  = 1;
    }

    if (item->email->encrypted_body.data) {
        pst_item_attach* attach = (pst_item_attach*)pst_malloc(sizeof(pst_item_attach));
        DEBUG_INFO(("Adding encrypted text body as attachment\n"));
        memset(attach, 0, sizeof(pst_item_attach));
        attach->next = item->attach;
        item->attach = attach;
        attach->data.data = item->email->encrypted_body.data;
        attach->data.size = item->email->encrypted_body.size;
        item->email->encrypted_body.data = NULL;
    }

    if (item->email->encrypted_htmlbody.data) {
        pst_item_attach* attach = (pst_item_attach*)pst_malloc(sizeof(pst_item_attach));
        DEBUG_INFO(("Adding encrypted HTML body as attachment\n"));
        memset(attach, 0, sizeof(pst_item_attach));
        attach->next = item->attach;
        item->attach = attach;
        attach->data.data = item->email->encrypted_htmlbody.data;
        attach->data.size = item->email->encrypted_htmlbody.size;
        item->email->encrypted_htmlbody.data = NULL;
    }

    if (item->type == PST_TYPE_SCHEDULE) {
        write_schedule_part(f_output, item, sender, boundary);
    }

    // other attachments
    {
        pst_item_attach* attach;
        int attach_num = 0;
        for (attach = item->attach; attach; attach = attach->next) {
            pst_convert_utf8_null(item, &attach->filename1);
            pst_convert_utf8_null(item, &attach->filename2);
            pst_convert_utf8_null(item, &attach->mimetype);
            DEBUG_INFO(("Attempting Attachment encoding\n"));
            if (attach->method == PST_ATTACH_EMBEDDED) {
                DEBUG_INFO(("have an embedded rfc822 message attachment\n"));
                if (attach->mimetype.str) {
                    DEBUG_INFO(("which already has a mime-type of %s\n", attach->mimetype.str));
                    free(attach->mimetype.str);
                }
                attach->mimetype.str = strdup(RFC822);
                attach->mimetype.is_utf8 = 1;
                find_rfc822_headers(extra_mime_headers);
                write_embedded_message(f_output, attach, boundary, pst, save_rtf, extra_mime_headers);
            }
            else if (attach->data.data || attach->i_id) {
                if (acceptable_ext(attach)) {
                    if (mode == MODE_SEPARATE && !mode_MH)
                        write_separate_attachment(f_name, attach, ++attach_num, pst);
                    else
                        write_inline_attachment(f_output, attach, boundary, pst);
                }
            }
        }
    }

    fprintf(f_output, "\n--%s--\n\n", boundary);
    DEBUG_RET();
}


void write_vcard(FILE* f_output, pst_item* item, pst_item_contact* contact, char comment[])
{
    char*  result = NULL;
    size_t resultlen = 0;
    char   time_buffer[30];
    // We can only call rfc escape once per printf, since the second call
    // may free the buffer returned by the first call.
    // I had tried to place those into a single printf - Carl.

    DEBUG_ENT("write_vcard");

    // make everything utf8
    pst_convert_utf8_null(item, &contact->fullname);
    pst_convert_utf8_null(item, &contact->surname);
    pst_convert_utf8_null(item, &contact->first_name);
    pst_convert_utf8_null(item, &contact->middle_name);
    pst_convert_utf8_null(item, &contact->display_name_prefix);
    pst_convert_utf8_null(item, &contact->suffix);
    pst_convert_utf8_null(item, &contact->nickname);
    pst_convert_utf8_null(item, &contact->address1);
    pst_convert_utf8_null(item, &contact->address2);
    pst_convert_utf8_null(item, &contact->address3);
    pst_convert_utf8_null(item, &contact->home_po_box);
    pst_convert_utf8_null(item, &contact->home_street);
    pst_convert_utf8_null(item, &contact->home_city);
    pst_convert_utf8_null(item, &contact->home_state);
    pst_convert_utf8_null(item, &contact->home_postal_code);
    pst_convert_utf8_null(item, &contact->home_country);
    pst_convert_utf8_null(item, &contact->home_address);
    pst_convert_utf8_null(item, &contact->business_po_box);
    pst_convert_utf8_null(item, &contact->business_street);
    pst_convert_utf8_null(item, &contact->business_city);
    pst_convert_utf8_null(item, &contact->business_state);
    pst_convert_utf8_null(item, &contact->business_postal_code);
    pst_convert_utf8_null(item, &contact->business_country);
    pst_convert_utf8_null(item, &contact->business_address);
    pst_convert_utf8_null(item, &contact->other_po_box);
    pst_convert_utf8_null(item, &contact->other_street);
    pst_convert_utf8_null(item, &contact->other_city);
    pst_convert_utf8_null(item, &contact->other_state);
    pst_convert_utf8_null(item, &contact->other_postal_code);
    pst_convert_utf8_null(item, &contact->other_country);
    pst_convert_utf8_null(item, &contact->other_address);
    pst_convert_utf8_null(item, &contact->business_fax);
    pst_convert_utf8_null(item, &contact->business_phone);
    pst_convert_utf8_null(item, &contact->business_phone2);
    pst_convert_utf8_null(item, &contact->car_phone);
    pst_convert_utf8_null(item, &contact->home_fax);
    pst_convert_utf8_null(item, &contact->home_phone);
    pst_convert_utf8_null(item, &contact->home_phone2);
    pst_convert_utf8_null(item, &contact->isdn_phone);
    pst_convert_utf8_null(item, &contact->mobile_phone);
    pst_convert_utf8_null(item, &contact->other_phone);
    pst_convert_utf8_null(item, &contact->pager_phone);
    pst_convert_utf8_null(item, &contact->primary_fax);
    pst_convert_utf8_null(item, &contact->primary_phone);
    pst_convert_utf8_null(item, &contact->radio_phone);
    pst_convert_utf8_null(item, &contact->telex);
    pst_convert_utf8_null(item, &contact->job_title);
    pst_convert_utf8_null(item, &contact->profession);
    pst_convert_utf8_null(item, &contact->assistant_name);
    pst_convert_utf8_null(item, &contact->assistant_phone);
    pst_convert_utf8_null(item, &contact->company_name);
    pst_convert_utf8_null(item, &item->body);

    // the specification I am following is (hopefully) RFC2426 vCard Mime Directory Profile
    fprintf(f_output, "BEGIN:VCARD\n");
    fprintf(f_output, "FN:%s\n", pst_rfc2426_escape(contact->fullname.str, &result, &resultlen));

    //fprintf(f_output, "N:%s;%s;%s;%s;%s\n",
    fprintf(f_output, "N:%s;", (!contact->surname.str)             ? "" : pst_rfc2426_escape(contact->surname.str, &result, &resultlen));
    fprintf(f_output, "%s;",   (!contact->first_name.str)          ? "" : pst_rfc2426_escape(contact->first_name.str, &result, &resultlen));
    fprintf(f_output, "%s;",   (!contact->middle_name.str)         ? "" : pst_rfc2426_escape(contact->middle_name.str, &result, &resultlen));
    fprintf(f_output, "%s;",   (!contact->display_name_prefix.str) ? "" : pst_rfc2426_escape(contact->display_name_prefix.str, &result, &resultlen));
    fprintf(f_output, "%s\n",  (!contact->suffix.str)              ? "" : pst_rfc2426_escape(contact->suffix.str, &result, &resultlen));

    if (contact->nickname.str)
        fprintf(f_output, "NICKNAME:%s\n", pst_rfc2426_escape(contact->nickname.str, &result, &resultlen));
    if (contact->address1.str)
        fprintf(f_output, "EMAIL:%s\n", pst_rfc2426_escape(contact->address1.str, &result, &resultlen));
    if (contact->address2.str)
        fprintf(f_output, "EMAIL:%s\n", pst_rfc2426_escape(contact->address2.str, &result, &resultlen));
    if (contact->address3.str)
        fprintf(f_output, "EMAIL:%s\n", pst_rfc2426_escape(contact->address3.str, &result, &resultlen));
    if (contact->birthday)
        fprintf(f_output, "BDAY:%s\n", pst_rfc2425_datetime_format(contact->birthday, sizeof(time_buffer), time_buffer));

    if (contact->home_address.str) {
        //fprintf(f_output, "ADR;TYPE=home:%s;%s;%s;%s;%s;%s;%s\n",
        fprintf(f_output, "ADR;TYPE=home:%s;",  (!contact->home_po_box.str)      ? "" : pst_rfc2426_escape(contact->home_po_box.str, &result, &resultlen));
        fprintf(f_output, "%s;",                ""); // extended Address
        fprintf(f_output, "%s;",                (!contact->home_street.str)      ? "" : pst_rfc2426_escape(contact->home_street.str, &result, &resultlen));
        fprintf(f_output, "%s;",                (!contact->home_city.str)        ? "" : pst_rfc2426_escape(contact->home_city.str, &result, &resultlen));
        fprintf(f_output, "%s;",                (!contact->home_state.str)       ? "" : pst_rfc2426_escape(contact->home_state.str, &result, &resultlen));
        fprintf(f_output, "%s;",                (!contact->home_postal_code.str) ? "" : pst_rfc2426_escape(contact->home_postal_code.str, &result, &resultlen));
        fprintf(f_output, "%s\n",               (!contact->home_country.str)     ? "" : pst_rfc2426_escape(contact->home_country.str, &result, &resultlen));
        fprintf(f_output, "LABEL;TYPE=home:%s\n", pst_rfc2426_escape(contact->home_address.str, &result, &resultlen));
    }

    if (contact->business_address.str) {
        //fprintf(f_output, "ADR;TYPE=work:%s;%s;%s;%s;%s;%s;%s\n",
        fprintf(f_output, "ADR;TYPE=work:%s;",  (!contact->business_po_box.str)      ? "" : pst_rfc2426_escape(contact->business_po_box.str, &result, &resultlen));
        fprintf(f_output, "%s;",                ""); // extended Address
        fprintf(f_output, "%s;",                (!contact->business_street.str)      ? "" : pst_rfc2426_escape(contact->business_street.str, &result, &resultlen));
        fprintf(f_output, "%s;",                (!contact->business_city.str)        ? "" : pst_rfc2426_escape(contact->business_city.str, &result, &resultlen));
        fprintf(f_output, "%s;",                (!contact->business_state.str)       ? "" : pst_rfc2426_escape(contact->business_state.str, &result, &resultlen));
        fprintf(f_output, "%s;",                (!contact->business_postal_code.str) ? "" : pst_rfc2426_escape(contact->business_postal_code.str, &result, &resultlen));
        fprintf(f_output, "%s\n",               (!contact->business_country.str)     ? "" : pst_rfc2426_escape(contact->business_country.str, &result, &resultlen));
        fprintf(f_output, "LABEL;TYPE=work:%s\n", pst_rfc2426_escape(contact->business_address.str, &result, &resultlen));
    }

    if (contact->other_address.str) {
        //fprintf(f_output, "ADR;TYPE=postal:%s;%s;%s;%s;%s;%s;%s\n",
        fprintf(f_output, "ADR;TYPE=postal:%s;",(!contact->other_po_box.str)       ? "" : pst_rfc2426_escape(contact->other_po_box.str, &result, &resultlen));
        fprintf(f_output, "%s;",                ""); // extended Address
        fprintf(f_output, "%s;",                (!contact->other_street.str)       ? "" : pst_rfc2426_escape(contact->other_street.str, &result, &resultlen));
        fprintf(f_output, "%s;",                (!contact->other_city.str)         ? "" : pst_rfc2426_escape(contact->other_city.str, &result, &resultlen));
        fprintf(f_output, "%s;",                (!contact->other_state.str)        ? "" : pst_rfc2426_escape(contact->other_state.str, &result, &resultlen));
        fprintf(f_output, "%s;",                (!contact->other_postal_code.str)  ? "" : pst_rfc2426_escape(contact->other_postal_code.str, &result, &resultlen));
        fprintf(f_output, "%s\n",               (!contact->other_country.str)      ? "" : pst_rfc2426_escape(contact->other_country.str, &result, &resultlen));
        fprintf(f_output, "LABEL;TYPE=postal:%s\n", pst_rfc2426_escape(contact->other_address.str, &result, &resultlen));
    }

    if (contact->business_fax.str)      fprintf(f_output, "TEL;TYPE=work,fax:%s\n",         pst_rfc2426_escape(contact->business_fax.str, &result, &resultlen));
    if (contact->business_phone.str)    fprintf(f_output, "TEL;TYPE=work,voice:%s\n",       pst_rfc2426_escape(contact->business_phone.str, &result, &resultlen));
    if (contact->business_phone2.str)   fprintf(f_output, "TEL;TYPE=work,voice:%s\n",       pst_rfc2426_escape(contact->business_phone2.str, &result, &resultlen));
    if (contact->car_phone.str)         fprintf(f_output, "TEL;TYPE=car,voice:%s\n",        pst_rfc2426_escape(contact->car_phone.str, &result, &resultlen));
    if (contact->home_fax.str)          fprintf(f_output, "TEL;TYPE=home,fax:%s\n",         pst_rfc2426_escape(contact->home_fax.str, &result, &resultlen));
    if (contact->home_phone.str)        fprintf(f_output, "TEL;TYPE=home,voice:%s\n",       pst_rfc2426_escape(contact->home_phone.str, &result, &resultlen));
    if (contact->home_phone2.str)       fprintf(f_output, "TEL;TYPE=home,voice:%s\n",       pst_rfc2426_escape(contact->home_phone2.str, &result, &resultlen));
    if (contact->isdn_phone.str)        fprintf(f_output, "TEL;TYPE=isdn:%s\n",             pst_rfc2426_escape(contact->isdn_phone.str, &result, &resultlen));
    if (contact->mobile_phone.str)      fprintf(f_output, "TEL;TYPE=cell,voice:%s\n",       pst_rfc2426_escape(contact->mobile_phone.str, &result, &resultlen));
    if (contact->other_phone.str)       fprintf(f_output, "TEL;TYPE=msg:%s\n",              pst_rfc2426_escape(contact->other_phone.str, &result, &resultlen));
    if (contact->pager_phone.str)       fprintf(f_output, "TEL;TYPE=pager:%s\n",            pst_rfc2426_escape(contact->pager_phone.str, &result, &resultlen));
    if (contact->primary_fax.str)       fprintf(f_output, "TEL;TYPE=fax,pref:%s\n",         pst_rfc2426_escape(contact->primary_fax.str, &result, &resultlen));
    if (contact->primary_phone.str)     fprintf(f_output, "TEL;TYPE=phone,pref:%s\n",       pst_rfc2426_escape(contact->primary_phone.str, &result, &resultlen));
    if (contact->radio_phone.str)       fprintf(f_output, "TEL;TYPE=pcs:%s\n",              pst_rfc2426_escape(contact->radio_phone.str, &result, &resultlen));
    if (contact->telex.str)             fprintf(f_output, "TEL;TYPE=bbs:%s\n",              pst_rfc2426_escape(contact->telex.str, &result, &resultlen));
    if (contact->job_title.str)         fprintf(f_output, "TITLE:%s\n",                     pst_rfc2426_escape(contact->job_title.str, &result, &resultlen));
    if (contact->profession.str)        fprintf(f_output, "ROLE:%s\n",                      pst_rfc2426_escape(contact->profession.str, &result, &resultlen));
    if (contact->assistant_name.str || contact->assistant_phone.str) {
        fprintf(f_output, "AGENT:BEGIN:VCARD\n");
        if (contact->assistant_name.str)    fprintf(f_output, "FN:%s\n",                    pst_rfc2426_escape(contact->assistant_name.str, &result, &resultlen));
        if (contact->assistant_phone.str)   fprintf(f_output, "TEL:%s\n",                   pst_rfc2426_escape(contact->assistant_phone.str, &result, &resultlen));
    }
    if (contact->company_name.str)      fprintf(f_output, "ORG:%s\n",                       pst_rfc2426_escape(contact->company_name.str, &result, &resultlen));
    if (comment)                        fprintf(f_output, "NOTE:%s\n",                      pst_rfc2426_escape(comment, &result, &resultlen));
    if (item->body.str)                 fprintf(f_output, "NOTE:%s\n",                      pst_rfc2426_escape(item->body.str, &result, &resultlen));

    write_extra_categories(f_output, item);

    fprintf(f_output, "VERSION: 3.0\n");
    fprintf(f_output, "END:VCARD\n\n");
    if (result) free(result);
    DEBUG_RET();
}


/**
 * write extra vcard or vcalendar categories from the extra keywords fields
 *
 * @param f_output open file pointer
 * @param item     pst item containing the keywords
 * @return         true if we write a categories line
 */
int write_extra_categories(FILE* f_output, pst_item* item)
{
    char*  result = NULL;
    size_t resultlen = 0;
    pst_item_extra_field *ef = item->extra_fields;
    const char *fmt = "CATEGORIES:%s";
    int category_started = 0;
    while (ef) {
        if (strcmp(ef->field_name, "Keywords") == 0) {
            fprintf(f_output, fmt, pst_rfc2426_escape(ef->value, &result, &resultlen));
            fmt = ", %s";
            category_started = 1;
        }
        ef = ef->next;
    }
    if (category_started) fprintf(f_output, "\n");
    if (result) free(result);
    return category_started;
}


void write_journal(FILE* f_output, pst_item* item)
{
    char*  result = NULL;
    size_t resultlen = 0;
    char   time_buffer[30];
    pst_item_journal* journal = item->journal;

    // make everything utf8
    pst_convert_utf8_null(item, &item->subject);
    pst_convert_utf8_null(item, &item->body);

    fprintf(f_output, "BEGIN:VJOURNAL\n");
    fprintf(f_output, "DTSTAMP:%s\n",                     pst_rfc2445_datetime_format_now(sizeof(time_buffer), time_buffer));
    if (item->create_date)
        fprintf(f_output, "CREATED:%s\n",                 pst_rfc2445_datetime_format(item->create_date, sizeof(time_buffer), time_buffer));
    if (item->modify_date)
        fprintf(f_output, "LAST-MOD:%s\n",                pst_rfc2445_datetime_format(item->modify_date, sizeof(time_buffer), time_buffer));
    if (item->subject.str)
        fprintf(f_output, "SUMMARY:%s\n",                 pst_rfc2426_escape(item->subject.str, &result, &resultlen));
    if (item->body.str)
        fprintf(f_output, "DESCRIPTION:%s\n",             pst_rfc2426_escape(item->body.str, &result, &resultlen));
    if (journal && journal->start)
        fprintf(f_output, "DTSTART;VALUE=DATE-TIME:%s\n", pst_rfc2445_datetime_format(journal->start, sizeof(time_buffer), time_buffer));
    fprintf(f_output, "END:VJOURNAL\n");
    if (result) free(result);
}


void write_appointment(FILE* f_output, pst_item* item)
{
    char*  result = NULL;
    size_t resultlen = 0;
    char   time_buffer[30];
    pst_item_appointment* appointment = item->appointment;

    // make everything utf8
    pst_convert_utf8_null(item, &item->subject);
    pst_convert_utf8_null(item, &item->body);
    pst_convert_utf8_null(item, &appointment->location);

    fprintf(f_output, "UID:%#"PRIx64"\n", item->block_id);
    fprintf(f_output, "DTSTAMP:%s\n",                     pst_rfc2445_datetime_format_now(sizeof(time_buffer), time_buffer));
    if (item->create_date)
        fprintf(f_output, "CREATED:%s\n",                 pst_rfc2445_datetime_format(item->create_date, sizeof(time_buffer), time_buffer));
    if (item->modify_date)
        fprintf(f_output, "LAST-MOD:%s\n",                pst_rfc2445_datetime_format(item->modify_date, sizeof(time_buffer), time_buffer));
    if (item->subject.str)
        fprintf(f_output, "SUMMARY:%s\n",                 pst_rfc2426_escape(item->subject.str, &result, &resultlen));
    if (item->body.str)
        fprintf(f_output, "DESCRIPTION:%s\n",             pst_rfc2426_escape(item->body.str, &result, &resultlen));
    if (appointment && appointment->start)
        fprintf(f_output, "DTSTART;VALUE=DATE-TIME:%s\n", pst_rfc2445_datetime_format(appointment->start, sizeof(time_buffer), time_buffer));
    if (appointment && appointment->end)
        fprintf(f_output, "DTEND;VALUE=DATE-TIME:%s\n",   pst_rfc2445_datetime_format(appointment->end, sizeof(time_buffer), time_buffer));
    if (appointment && appointment->location.str)
        fprintf(f_output, "LOCATION:%s\n",                pst_rfc2426_escape(appointment->location.str, &result, &resultlen));
    if (appointment) {
        switch (appointment->showas) {
            case PST_FREEBUSY_TENTATIVE:
                fprintf(f_output, "STATUS:TENTATIVE\n");
                break;
            case PST_FREEBUSY_FREE:
                // mark as transparent and as confirmed
                fprintf(f_output, "TRANSP:TRANSPARENT\n");
            case PST_FREEBUSY_BUSY:
            case PST_FREEBUSY_OUT_OF_OFFICE:
                fprintf(f_output, "STATUS:CONFIRMED\n");
                break;
        }
        if (appointment->is_recurring) {
            const char* rules[] = {"DAILY", "WEEKLY", "MONTHLY", "YEARLY"};
            const char* days[]  = {"SU", "MO", "TU", "WE", "TH", "FR", "SA"};
            pst_recurrence *rdata = pst_convert_recurrence(appointment);
            fprintf(f_output, "RRULE:FREQ=%s", rules[rdata->type]);
            if (rdata->count)       fprintf(f_output, ";COUNT=%u",      rdata->count);
            if ((rdata->interval != 1) &&
                (rdata->interval))  fprintf(f_output, ";INTERVAL=%u",   rdata->interval);
            if (rdata->dayofmonth)  fprintf(f_output, ";BYMONTHDAY=%d", rdata->dayofmonth);
            if (rdata->monthofyear) fprintf(f_output, ";BYMONTH=%d",    rdata->monthofyear);
            if (rdata->position)    fprintf(f_output, ";BYSETPOS=%d",   rdata->position);
            if (rdata->bydaymask) {
                char byday[40];
                int  empty = 1;
                int i=0;
                memset(byday, 0, sizeof(byday));
                for (i=0; i<6; i++) {
                    int bit = 1 << i;
                    if (bit & rdata->bydaymask) {
                        char temp[40];
                        snprintf(temp, sizeof(temp), "%s%s%s", byday, (empty) ? ";BYDAY=" : ";", days[i]);
                        strcpy(byday, temp);
                        empty = 0;
                    }
                }
                fprintf(f_output, "%s", byday);
            }
            fprintf(f_output, "\n");
            pst_free_recurrence(rdata);
        }
        switch (appointment->label) {
            case PST_APP_LABEL_NONE:
                if (!write_extra_categories(f_output, item)) fprintf(f_output, "CATEGORIES:NONE\n");
                break;
            case PST_APP_LABEL_IMPORTANT:
                fprintf(f_output, "CATEGORIES:IMPORTANT\n");
                break;
            case PST_APP_LABEL_BUSINESS:
                fprintf(f_output, "CATEGORIES:BUSINESS\n");
                break;
            case PST_APP_LABEL_PERSONAL:
                fprintf(f_output, "CATEGORIES:PERSONAL\n");
                break;
            case PST_APP_LABEL_VACATION:
                fprintf(f_output, "CATEGORIES:VACATION\n");
                break;
            case PST_APP_LABEL_MUST_ATTEND:
                fprintf(f_output, "CATEGORIES:MUST-ATTEND\n");
                break;
            case PST_APP_LABEL_TRAVEL_REQ:
                fprintf(f_output, "CATEGORIES:TRAVEL-REQUIRED\n");
                break;
            case PST_APP_LABEL_NEEDS_PREP:
                fprintf(f_output, "CATEGORIES:NEEDS-PREPARATION\n");
                break;
            case PST_APP_LABEL_BIRTHDAY:
                fprintf(f_output, "CATEGORIES:BIRTHDAY\n");
                break;
            case PST_APP_LABEL_ANNIVERSARY:
                fprintf(f_output, "CATEGORIES:ANNIVERSARY\n");
                break;
            case PST_APP_LABEL_PHONE_CALL:
                fprintf(f_output, "CATEGORIES:PHONE-CALL\n");
                break;
        }
        // ignore bogus alarms
        if (appointment->alarm && (appointment->alarm_minutes >= 0) && (appointment->alarm_minutes < 1440)) {
            fprintf(f_output, "BEGIN:VALARM\n");
            fprintf(f_output, "TRIGGER:-PT%dM\n", appointment->alarm_minutes);
            fprintf(f_output, "ACTION:DISPLAY\n");
            fprintf(f_output, "DESCRIPTION:Reminder\n");
            fprintf(f_output, "END:VALARM\n");
        }
    }
    fprintf(f_output, "END:VEVENT\n");
    if (result) free(result);
}


void create_enter_dir(struct file_ll* f, pst_item *item)
{
    pst_convert_utf8(item, &item->file_as);
    f->type         = item->type;
    f->stored_count = (item->folder) ? item->folder->item_count : 0;

    DEBUG_ENT("create_enter_dir");
    if (mode == MODE_KMAIL)
        f->name = mk_kmail_dir(item->file_as.str);
    else if (mode == MODE_RECURSE) {
        f->name = mk_recurse_dir(item->file_as.str, f->type);
        if (mode_thunder) {
            FILE *type_file = fopen(".type", "w");
            fprintf(type_file, "%d\n", item->type);
            fclose(type_file);
        }
    } else if (mode == MODE_SEPARATE) {
        // do similar stuff to recurse here.
        mk_separate_dir(item->file_as.str);
        f->name = (char*) pst_malloc(file_name_len);
        memset(f->name, 0, file_name_len);
    } else {
        f->name = (char*) pst_malloc(strlen(item->file_as.str)+strlen(OUTPUT_TEMPLATE)+1);
        sprintf(f->name, OUTPUT_TEMPLATE, item->file_as.str);
    }

    f->dname = (char*) pst_malloc(strlen(item->file_as.str)+1);
    strcpy(f->dname, item->file_as.str);

    if (overwrite != 1) {
        int x = 0;
        char *temp = (char*) pst_malloc (strlen(f->name)+10); //enough room for 10 digits

        sprintf(temp, "%s", f->name);
        check_filename(temp);
        while ((f->output = fopen(temp, "r"))) {
            DEBUG_INFO(("need to increase filename because one already exists with that name\n"));
            DEBUG_INFO(("- increasing it to %s%d\n", f->name, x));
            x++;
            sprintf(temp, "%s%08d", f->name, x);
            DEBUG_INFO(("- trying \"%s\"\n", f->name));
            if (x == 99999999) {
                DIE(("create_enter_dir: Why can I not create a folder %s? I have tried %i extensions...\n", f->name, x));
            }
            fclose(f->output);
        }
        if (x > 0) { //then the f->name should change
            free (f->name);
            f->name = temp;
        } else {
            free(temp);
        }
    }

    DEBUG_INFO(("f->name = %s\nitem->folder_name = %s\n", f->name, item->file_as.str));
    if (mode != MODE_SEPARATE) {
        check_filename(f->name);
        if (!(f->output = fopen(f->name, "w"))) {
            DIE(("create_enter_dir: Could not open file \"%s\" for write\n", f->name));
        }
    }
    DEBUG_RET();
}


void close_enter_dir(struct file_ll *f)
{
    DEBUG_INFO(("processed item count for folder %s is %i, skipped %i, total %i \n",
                f->dname, f->item_count, f->skip_count, f->stored_count));
    if (output_mode != OUTPUT_QUIET) {
        pst_debug_lock();
            printf("\t\"%s\" - %i items done, %i items skipped.\n", f->dname, f->item_count, f->skip_count);
            fflush(stdout);
        pst_debug_unlock();
    }
    if (f->output) {
        if (mode == MODE_SEPARATE) DEBUG_WARN(("close_enter_dir finds open separate file\n"));
        struct stat st;
        fclose(f->output);
        stat(f->name, &st);
        if (!st.st_size) {
            DEBUG_WARN(("removing empty output file %s\n", f->name));
            remove(f->name);
        }
        f->output = NULL;
    }
    free(f->name);
    free(f->dname);

    if (mode == MODE_KMAIL)
        close_kmail_dir();
    else if (mode == MODE_RECURSE) {
        if (mode_thunder) {
            FILE *type_file = fopen(".size", "w");
            fprintf(type_file, "%i %i\n", f->item_count, f->stored_count);
            fclose(type_file);
        }
        close_recurse_dir();
    } else if (mode == MODE_SEPARATE)
        close_separate_dir();
}