/*
**  Copyright (c) 2001 The OSSP Project <http://www.ossp.org/>
**  Copyright (c) 2001 Cable & Wireless Deutschland <http://www.cw.com/de/>
**
**  This file is part of OSSP lmtp2nntp, an LMTP speaking local
**  mailer which forwards mails as Usenet news articles via NNTP.
**  It can be found at http://www.ossp.org/pkg/lmtp2nntp/.
**
**  This program is free software; you can redistribute it and/or
**  modify it under the terms of the GNU General Public  License
**  as published by the Free Software Foundation; either version
**  2.0 of the License, or (at your option) any later version.
**
**  This program is distributed in the hope that it will be useful,
**  but WITHOUT ANY WARRANTY; without even the implied warranty of
**  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
**  General Public License for more details.
**
**  You should have received a copy of the GNU General Public License
**  along with this file; if not, write to the Free Software
**  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307
**  USA, or contact the OSSP project <ossp@ossp.org>.
**
**  msg.c: mail message manipulation library
*/

#include <stdlib.h>
#include <stdio.h>

#include "msg.h"
#include "str.h"
#include "argz.h"

/* third party */
#include "l2.h"

#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#if defined(HAVE_DMALLOC_H) && defined(DMALLOC)
#include "dmalloc.h"
#endif

msg_t *msg_create(void)
{
    msg_t *msg;

    if ((msg = (msg_t *)malloc(sizeof(msg_t))) == NULL)
        return NULL;

    msg->azEnvgroups = NULL;
    msg->asEnvgroups = 0;
    msg->cpMsg = NULL;
    msg->azHeaders = NULL;
    msg->asHeaders = 0;
    msg->cpFid = NULL;
    msg->cpBody = NULL;
    msg->cpMsgid = NULL;
    msg->mail_from = NULL;
    msg->azRcpt = NULL;
    msg->asRcpt = 0;
    msg->azNewsgroups = NULL;
    msg->asNewsgroups = 0;
    msg->l2 = NULL; /* this is a copy only */

    return msg;
}

void msg_destroy(msg_t *msg)
{
    if (msg == NULL)
        return;

    if (msg->azEnvgroups != NULL)
        free(msg->azEnvgroups);
    if (msg->cpMsg != NULL)
        free(msg->cpMsg);
    if (msg->azHeaders != NULL)
        free(msg->azHeaders);
    if (msg->cpFid != NULL)
        free(msg->cpFid);
    if (msg->cpBody != NULL)
        free(msg->cpBody);
    if (msg->cpMsgid != NULL)
        free(msg->cpMsgid);
    if (msg->mail_from != NULL)
        free(msg->mail_from);
    if (msg->azRcpt != NULL)
        free(msg->azRcpt);
    if (msg->azNewsgroups != NULL)
        free(msg->azNewsgroups);
    msg->l2 = NULL; /* this is a copy only, the "parent" needs to clean this up */

    free(msg);
    return;
}

msg_rc_t msg_split(msg_t *msg)
{
    char        *cpName;
    char        *cpValue;
    char        *cpRem;     /* Remainder */
    char        *cp;
    char        *cpHeaders;

    /* INPUTS
     *
     * msg->cpMsg
     * must contain the wholly RFC0822 formatted message with native
     * (unescaped) dots at the beginning of a line, the 'From ' envelope,
     * headers, double newline, body, NUL, no trailing dot;
     *
     * OUTPUTS
     *
     * msg->cpMsg
     * free()d and set to NULL
     * 
     * msg->azHeaders, msg->asHeaders contains the headers in argz format, one
     * logical NUL-terminated line per header which might be wrapped into
     * multiple '\n'-ended physical lines. The "From " envelope, "Received:",
     * "Path:", "To:" and "Cc:" headers are removed silently. The
     * "Newsgroups:" and "Message-ID" headers are removed and their values are
     * stored in separate structures (see below).
     *
     * msg->cpBody
     * contains the unmodified body of the message, NUL-terminated, no
     * trailing dot.
     *
     * msg->cpMsgid
     * contains the message id including surrounding angle brackets.
     *
     * msg->azNewsgroups, asNewsgroups
     * is a argz-type array of strings containing the Newsgroups based on the
     * header information.
     */

    log0(msg, DEBUG, "split message into header and body");
    if (str_parse(msg->cpMsg, "m/((?:.*?)\\n)\\n(.*)$/s", &cpHeaders, &msg->cpBody) <= 0)
        return MSG_ERR_SPLITHEADBODY;

    free(msg->cpMsg);
    msg->cpMsg = NULL;

    log0(msg, DEBUG, "replace envelope From w/o colon by X-F: pseudotag");
    /* This eliminates the special case of having one header, which is really
     * an embedded envelope, not ending with a colon while all others do.
     * After splitting headers into name and value pairs this envelope ist
     * stripped off.
     */
    if (strncasecmp(cpHeaders, "From", 4) == 0)
        memcpy(cpHeaders, "X-F:", 4);

    log0(msg, DEBUG, "unwrap header lines");
    /* poor man's s///g simulator as current str library doesn't support global substitution */
    while (str_parse(cpHeaders, "s/(.*?)\\n[ \\t]+(.*)/$1 $2/s", &cpRem) > 0) {
        free(cpHeaders);
        cpHeaders = cpRem;
    }

    log0(msg, DEBUG, "split header lines into names and values");
    while (str_parse(cpHeaders, "m/^[> \\t]*([\\x21-\\x7e]+?:)[ \\t]*([^\\n]*?)[ \\t]*\\n(.*)/s", &cpName, &cpValue, &cpRem) > 0) {
        free(cpHeaders);
        cpHeaders = cpRem;
        argz_add(&msg->azHeaders, &msg->asHeaders, cpName);
        argz_add(&msg->azHeaders, &msg->asHeaders, cpValue);
        free(cpName);
        free(cpValue);
    }

    log0(msg, DEBUG, "check for headers we care about and do whatever neccessary");
    msg->cpMsgid = NULL;
    msg->azNewsgroups = NULL;
    msg->asNewsgroups = 0;
    cp = msg->azHeaders;
    while (cp != NULL) {
        log1(msg, DEBUG, "processing header \"%s\"", cp);
        if (strcasecmp("X-F:", cp) == 0) {
            argz_delete(&msg->azHeaders, &msg->asHeaders, cp);             /* del  name  */
            argz_delete(&msg->azHeaders, &msg->asHeaders, cp);             /* del  value */
            continue;
        }
        if (strcasecmp("Path:", cp) == 0) {
            argz_delete(&msg->azHeaders, &msg->asHeaders, cp);             /* del  name  */
            argz_delete(&msg->azHeaders, &msg->asHeaders, cp);             /* del  value */
            continue;
        }
        if (strcasecmp("Received:", cp) == 0) {
            argz_delete(&msg->azHeaders, &msg->asHeaders, cp);             /* del  name  */
            if ((msg->cpFid == NULL) &&
                (str_parse(cp, "m/\\sid\\s+<?([\\w\\d]{1,30})/i", &msg->cpFid) > 0))
                    log1(msg, DEBUG, "found foreign-ID \"%s\" for logging", msg->cpFid);
            argz_delete(&msg->azHeaders, &msg->asHeaders, cp);             /* del  value */
            continue;
        }
        if (strcasecmp("To:", cp) == 0) {
            argz_delete(&msg->azHeaders, &msg->asHeaders, cp);             /* del  name  */
            argz_delete(&msg->azHeaders, &msg->asHeaders, cp);             /* del  value */
            continue;
        }
        if (strcasecmp("Cc:", cp) == 0) {
            argz_delete(&msg->azHeaders, &msg->asHeaders, cp);             /* del  name  */
            argz_delete(&msg->azHeaders, &msg->asHeaders, cp);             /* del  value */
            continue;
        }
        if (strcasecmp("Message-ID:", cp) == 0) {
            if (msg->cpMsgid != NULL)
                return MSG_ERR_SPLITIDMULTI;
            argz_delete(&msg->azHeaders, &msg->asHeaders, cp);             /* del  name  */
            if ((cp == NULL) || (strlen(cp) == 0))                         /* get  value */
                return MSG_ERR_SPLITIDEMPTY;
            if ((msg->cpMsgid = strdup(cp)) == NULL)
                return MSG_ERR_MEM;
            argz_delete(&msg->azHeaders, &msg->asHeaders, cp);             /* del  value */
            continue;
        }
        if (strcasecmp("Newsgroups:", cp) == 0) {
            argz_delete(&msg->azHeaders, &msg->asHeaders, cp);             /* del  name  */
            if (argz_add(&msg->azNewsgroups, &msg->asNewsgroups, cp) != 0) /* get  value */
                return MSG_ERR_MEM;
            argz_delete(&msg->azHeaders, &msg->asHeaders, cp);             /* del  value */
            continue;
        }
        if ((cp = argz_next(msg->azHeaders, msg->asHeaders, cp)) == NULL)  /* next value */
            break;
        if ((cp = argz_next(msg->azHeaders, msg->asHeaders, cp)) == NULL)  /* next name  */
            break;
    }

    log0(msg, DEBUG, "checking Message-ID");
    if (msg->cpMsgid == NULL)
        return MSG_ERR_SPLITIDNONE;

    log0(msg, DEBUG, "checking Newsgroups");
    if (msg->azNewsgroups != NULL) {
        argz_stringify(msg->azNewsgroups, msg->asNewsgroups, ',');
        if (argz_create_sep(msg->azNewsgroups, ',', &msg->azNewsgroups, &msg->asNewsgroups) != 0)
            return MSG_ERR_MEM;
    }

    log0(msg, DEBUG, "adding mandatory Path: header");
    argz_add(&msg->azHeaders, &msg->asHeaders, "Path:");
    argz_add(&msg->azHeaders, &msg->asHeaders, "lmtp2nntp!not-for-mail");

    log0(msg, DEBUG, "split complete");
    return MSG_OK;
}

msg_rc_t msg_join(msg_t *msg)
{
    char        *cp;
    char        *cpRem;
    char       **aHeaders;
    int          i;
    int          o;
    char        *cpCut;
    char        *cpWrap;
    char         c;
    char         cOld;
    int          n;
    char        *cpHeaders;
    char        *azNewheaders;
    size_t       asNewheaders;

    log0(msg, DEBUG, "verify Newsgroups");
    if (msg->azNewsgroups == NULL)
        return MSG_ERR_JOINGROUPNONE;
    argz_stringify(msg->azNewsgroups, msg->asNewsgroups, ',');
    if (strlen(msg->azNewsgroups) == 0)
        return MSG_ERR_JOINGROUPEMPTY;
    argz_add(&msg->azHeaders, &msg->asHeaders, "Newsgroups:");
    argz_add(&msg->azHeaders, &msg->asHeaders, msg->azNewsgroups);

    log0(msg, DEBUG, "verify Message-ID");
    if (msg->cpMsgid == NULL)
        return MSG_ERR_JOINIDNONE;
    if (strlen(msg->cpMsgid) == 0)
        return MSG_ERR_JOINIDEMPTY;
    argz_add(&msg->azHeaders, &msg->asHeaders, "Message-ID:");
    argz_add(&msg->azHeaders, &msg->asHeaders, msg->cpMsgid);

    log0(msg, DEBUG, "merge name/value pairs into single string");
    argz_add(&msg->azHeaders, &msg->asHeaders, ""); /* append empty string */
    if ((aHeaders = (char **)malloc((argz_count(msg->azHeaders, msg->asHeaders) + 1) * sizeof(char *))) == NULL)
        return MSG_ERR_MEM;
    argz_extract(msg->azHeaders, msg->asHeaders, aHeaders);
    /* replace the trailing NUL, which is *(cp-1) of the predecessor, with a
     * space at every second string. Break action when terminating NULL string
     * is detected */
    i=0;
    while(1) {
        if ((cp = aHeaders[++i]) == NULL)
            break;
        *(cp-1) = ' ';
        if ((cp = aHeaders[++i]) == NULL)
            break;
    }
    free(aHeaders);

    log0(msg, DEBUG, "fold headers");
    /* A logical line is split into one or more physical '\n'-terminated
     * lines. The physical line is never longer than WRAPAT characters. This
     * includes the folded data and the header name + colon + space for the
     * first line and WRAPUSING string prefix for all other lines. Leading and
     * trailing blanks of folded lines are removed while blanks inside the
     * line are preserved.  The header is never left alone in a physical line.
     * Fragments exceeding WRAPAT characters without having a blank as a
     * splitting point are forcibly cut at a non-blank character.
     */
    azNewheaders = NULL;
    asNewheaders = 0;
    cp = NULL;
    while ((cp = argz_next(msg->azHeaders, msg->asHeaders, cp)) != NULL) {
        if (strlen(cp) > WRAPAT) {
            cpRem = cp;
            cpWrap = NULL;
            for (o = 0; (cpRem[o] != ':') && (cpRem[o] != NUL); o++); /* offset name so at least one char of value remains in first line */
            o += 2; /* skip ": " */
            while ((strlen(cpRem) + (cpWrap == NULL ? 0 : strlen(WRAPUSING))) > WRAPAT) {
                for (i = WRAPAT - 1 - (cpWrap == NULL ? 0 :
                    strlen(WRAPUSING)); (i >= o) && !isspace((int)cpRem[i]); i--);
                if (i < o)
                    i = WRAPAT - 1 - (cpWrap == NULL ? 0 : strlen(WRAPUSING) - 1); /* sorry, forced cut at non-blank */
                cpCut = cpRem;
                cpRem += i;
                for (; (isspace((int)*cpRem) && (*cpRem != NUL)); cpRem++); /* skip next lines leading blanks */
                for (; (i >= o) && isspace((int)cpCut[i-1]); i--); /* chop off this lines trailing blanks */
                if (i >= o) { /* only keep line fragment if some non-blanks inside */
                    if (cpWrap == NULL) {
                        if ((cpWrap = (char *)malloc(i+strlen(WRAPUSING)+1)) == NULL)
                            return MSG_ERR_MEM;
                        *cpWrap = NUL;
                        o = 1;
                    }
                    else {
                        if ((cpWrap = (char *)realloc(cpWrap, strlen(cpWrap)+i+strlen(WRAPUSING)+1)) == NULL)
                            return MSG_ERR_MEM;
                        strcat(cpWrap, WRAPUSING);
                    }
                    strncat(cpWrap, cpCut, i);
                }
            }
            if (strlen(cpRem) > 0) {
                if ((cpWrap = (char *)realloc(cpWrap, strlen(cpWrap)+strlen(cpRem)+strlen(WRAPUSING)+1)) == NULL)
                        return MSG_ERR_MEM;
                strcat(cpWrap, WRAPUSING);
                strcat(cpWrap, cpRem);
            }
            argz_add(&azNewheaders, &asNewheaders, cpWrap);
            log2(msg, DEBUG, "a folded header \"%{text}D\"", cpWrap, strlen(cpWrap));
            free(cpWrap);
        }
        else {
            argz_add(&azNewheaders, &asNewheaders, cp);
            log2(msg, DEBUG, "verbatim header \"%{text}D\"", cp, strlen(cp));
        }
    }
    free(msg->azHeaders);
    msg->azHeaders = azNewheaders;
    msg->asHeaders = asNewheaders;

    log0(msg, DEBUG, "strigify headers");
    argz_stringify(msg->azHeaders, msg->asHeaders, '\n');
    cpHeaders = msg->azHeaders;

    /********************************************************************
     * header + CRLF + body + '.' + CRLF + NUL, replacing NL with CRLF *
     ********************************************************************/

    log0(msg, DEBUG, "assemble header and body");
    n = 0;
    /* count size of headers, reserve space for NL to CRLF conversion */
    for (i = 0; ((c = cpHeaders[i]) != NUL); i++) {
        if (c == '\n')
            n++;
        n++;
    }
    /* if headers don't end with NL, reserve space for CRLF */
    if (i >= 0 && cpHeaders[i - 1] != '\n')
        n+=2;
    /* reserve space for CRLF between headers and body */
    n+=2;
    /* count size of body, reserve space for NL-DOT escape and NL to CRLF conversion */
    cOld = '\n';
    for (i = 0; ((c = msg->cpBody[i]) != NUL); i++) {
        if (c == '\n')
            n++;
        if (c == '.' && cOld == '\n')
            n++;
        n++;
        cOld = c;
    }
    /* if body doesn't end with NL, reserve space for CRLF */
    if (i >= 0 && msg->cpBody[i - 1] != '\n')
        n+=2;
    /* reserve space for terminating '.'-CRLF-NUL at the end of the message */
    n+=4;

    if ((msg->cpMsg = (char *)malloc(n)) == NULL)
        return MSG_ERR_MEM;

    n = 0;
    /* copy headers, do NL to CRLF conversion */
    for (i = 0; ((c = cpHeaders[i]) != NUL); i++) {
        if (c == '\n')
            msg->cpMsg[n++] = '\r';
        msg->cpMsg[n++] = c;
    }
    /* if headers don't end with NL, append CRLF */
    if (i >= 0 && cpHeaders[i - 1] != '\n') {
        msg->cpMsg[n++] = '\r';
        msg->cpMsg[n++] = '\n';
    }
    /* add CRLF between headers and body */
    msg->cpMsg[n++] = '\r';
    msg->cpMsg[n++] = '\n';
    /* copy body, do NL-DOT escape and NL to CRLF conversion */
    cOld = '\n';
    for (i = 0; ((c = msg->cpBody[i]) != NUL); i++) {
        if (c == '\n')
            msg->cpMsg[n++] = '\r';
        if (c == '.' && cOld == '\n')
            msg->cpMsg[n++] = '.';
        msg->cpMsg[n++] = c;
        cOld = c;
    }
    /* if body doesn't end with NL, append CRLF */
    if (i >= 0 && msg->cpBody[i - 1] != '\n') {
        msg->cpMsg[n++] = '\r';
        msg->cpMsg[n++] = '\n';
    }
    /* add terminating '.'-CRLF-NUL at the end of the message */
    msg->cpMsg[n++] =  '.';
    msg->cpMsg[n++] = '\r';
    msg->cpMsg[n++] = '\n';
    msg->cpMsg[n]   = NUL;

    log0(msg, DEBUG, "join complete");
    return MSG_OK;
}

char *msg_error(msg_rc_t rc)
{
    char *str;
                                              str = "MSG: no description";
    if      (rc == MSG_OK                   ) str = "MSG: no error";
    else if (rc == MSG_ERR_MEM              ) str = "MSG: memory";
    else if (rc == MSG_ERR_SPLITHEADBODY    ) str = "MSG: split into header and body failed";
    else if (rc == MSG_ERR_SPLITLEN         ) str = "MSG: header is too short";
    else if (rc == MSG_ERR_SPLITMISSINGFROM ) str = "MSG: header is missing 'From ' envelope";
    else if (rc == MSG_ERR_SPLITIDNONE      ) str = "MSG: header is missing 'Message-ID'";
    else if (rc == MSG_ERR_SPLITIDEMPTY     ) str = "MSG: header has empty 'Message-ID'";
    else if (rc == MSG_ERR_SPLITIDMULTI     ) str = "MSG: header has multiple 'Message-ID's";
    else if (rc == MSG_ERR_JOINGROUPNONE    ) str = "MSG: join with no 'Newsgroup'";
    else if (rc == MSG_ERR_JOINGROUPEMPTY   ) str = "MSG: join with empty 'Newsgroup'";
    else if (rc == MSG_ERR_JOINIDNONE       ) str = "MSG: join with no 'Message-ID'";
    else if (rc == MSG_ERR_JOINIDEMPTY      ) str = "MSG: join with empty 'Message-ID'";
    return str;
}