/* Copyright (C) 2005 J.F.Dockes
 *   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 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 program; if not, write to the
 *   Free Software Foundation, Inc.,
 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
 */
#include "autoconfig.h"

#include <stdio.h>
#include <fcntl.h>
#include <errno.h>
#include "safeunistd.h"
#include <time.h>
#include <cstdlib>

#include <map>
#include <sstream>

#include "base64.h"
#include "cstr.h"
#include "mimehandler.h"
#include "readfile.h"
#include "transcode.h"
#include "mimeparse.h"
#include "mh_mail.h"
#include "log.h"
#include "smallut.h"
#include "mh_html.h"
#include "rclconfig.h"
#include "mimetype.h"
#include "md5ut.h"

// binc imap mime definitions
#include "mime.h"

using namespace std;

static const int maxdepth = 20;
static const string cstr_mail_charset("charset");

MimeHandlerMail::MimeHandlerMail(RclConfig *cnf, const string &id) 
    : RecollFilter(cnf, id), m_bincdoc(0), m_fd(-1), m_stream(0), m_idx(-1)
{
    // Look for additional headers to be processed as per config:
    vector<string> hdrnames = m_config->getFieldSectNames("mail");
    if (hdrnames.empty())
        return;
    for (const auto& nm : hdrnames) {
        (void)m_config->getFieldConfParam(nm, "mail", m_addProcdHdrs[nm]);
    }
    cnf->getConfParam("textfilemaxmbs", &m_maxtxtmbs);
}

MimeHandlerMail::~MimeHandlerMail() 
{
    if (m_fd >= 0) {
        close(m_fd);
        m_fd = -1;
    }
}

void MimeHandlerMail::clear_impl()
{
    delete m_bincdoc; m_bincdoc = 0;
    if (m_fd >= 0) {
        close(m_fd);
        m_fd = -1;
    }
    delete m_stream; m_stream = 0;
    m_idx = -1;
    m_startoftext = 0;
    m_subject.erase();
    for (auto attp : m_attachments) {
        delete attp;
    }
    m_attachments.clear();
}

bool MimeHandlerMail::set_document_file_impl(const string&, const string &fn)
{
    LOGDEB("MimeHandlerMail::set_document_file(" << fn << ")\n");
    if (m_fd >= 0) {
        close(m_fd);
        m_fd = -1;
    }

    if (!m_forPreview) {
        // Yes, we read the file twice. It would be possible in theory
        // to add the md5 computation to the mime analysis, but ...
        string md5, xmd5, reason;
        if (MD5File(fn, md5, &reason)) {
            m_metaData[cstr_dj_keymd5] = MD5HexPrint(md5, xmd5);
        } else {
            LOGERR("MimeHandlerMail: md5 [" << fn << "]: " << reason << "\n");
        }
    }
    m_fd = open(fn.c_str(), 0);
    if (m_fd < 0) {
        LOGERR("MimeHandlerMail::set_document_file: open(" << fn <<
               ") errno " << errno << "\n");
        return false;
    }
#if defined O_NOATIME && O_NOATIME != 0
    if (fcntl(m_fd, F_SETFL, O_NOATIME) < 0) {
        // perror("fcntl");
    }
#endif
    delete m_bincdoc;
    m_bincdoc = new Binc::MimeDocument;
    m_bincdoc->parseFull(m_fd);
    if (!m_bincdoc->isHeaderParsed() && !m_bincdoc->isAllParsed()) {
        LOGERR("MimeHandlerMail::mkDoc: mime parse error for " << fn << "\n");
        return false;
    }
    m_havedoc = true;
    return true;
}

bool MimeHandlerMail::set_document_string_impl(const string&, const string& msgtxt)
{
    LOGDEB1("MimeHandlerMail::set_document_string\n");
    LOGDEB2("Message text: [" << msgtxt << "]\n");
    delete m_stream;

    if (!m_forPreview) {
        string md5, xmd5;
        MD5String(msgtxt, md5);
        m_metaData[cstr_dj_keymd5] = MD5HexPrint(md5, xmd5);
    }

    if ((m_stream = new stringstream(msgtxt)) == 0 || !m_stream->good()) {
        LOGERR("MimeHandlerMail::set_document_string: stream create error."
               "msgtxt.size() " << msgtxt.size() << "\n");
        return false;
    }
    delete m_bincdoc;
    if ((m_bincdoc = new Binc::MimeDocument) == 0) {
        LOGERR("MimeHandlerMail::set_doc._string: new Binc:Document failed. "
               "Out of memory?");
        return false;
    }
    m_bincdoc->parseFull(*m_stream);
    if (!m_bincdoc->isHeaderParsed() && !m_bincdoc->isAllParsed()) {
        LOGERR("MimeHandlerMail::set_document_string: mime parse error\n");
        return false;
    }
    m_havedoc = true;
    return true;
}

bool MimeHandlerMail::skip_to_document(const string& ipath) 
{
    LOGDEB("MimeHandlerMail::skip_to_document(" << ipath << ")\n");
    if (m_idx == -1) {
        // No decoding done yet. If ipath is null need do nothing
        if (ipath.empty() || ipath == "-1")
            return true;
        // ipath points to attachment: need to decode message
        if (!next_document()) {
            LOGERR("MimeHandlerMail::skip_to_doc: next_document failed\n");
            return false;
        }
    }
    m_idx = atoi(ipath.c_str());
    return true;
}

bool MimeHandlerMail::next_document()
{
    LOGDEB("MimeHandlerMail::next_document m_idx " << m_idx << " m_havedoc " <<
           m_havedoc << "\n");
    if (!m_havedoc)
        return false;
    bool res = false;

    if (m_idx == -1) {
        m_metaData[cstr_dj_keymt] = cstr_textplain;
        res = processMsg(m_bincdoc, 0);
        LOGDEB1("MimeHandlerMail::next_document: mt " << m_metaData[cstr_dj_keymt] << ", att cnt " <<
                m_attachments.size() << "\n");
        const string& txt = m_metaData[cstr_dj_keycontent];
        if ((int)(txt.size()/(1024*1024)) > m_maxtxtmbs) {
            // Guard against absurd messages. E.g. files which would be mis-identified as rfc/822
            // and actually be some non-mbox kind of concatenation e.g. happened with enormous
            // procmail backup.
            LOGERR("MimeHandlerMail:: body text size > textfilemaxmbs, not processing message.\n");
            return false;
        }
        if (m_startoftext < txt.size())
            m_metaData[cstr_dj_keyabstract] = truncate_to_word(txt.substr(m_startoftext), 250);
        if (m_attachments.size() > 0) {
            m_metaData[cstr_dj_keyanc] = "t";
        }
    } else {
        m_metaData[cstr_dj_keyabstract].clear();
        res = processAttach();
    }
    m_idx++;
    m_havedoc = m_idx < (int)m_attachments.size();
    if (!m_havedoc) {
        m_reason = "Subdocument index too high";
    }
    return res;
}

// Decode according to content transfer encoding. May actually do nothing,
// which will be indicated by the *respp argument pointing to the original 
// text on exit
static bool decodeBody(const string& cte, // Content transfer encoding
                       const string& body, // Source text
                       string& decoded,   // Decoded text if actual decoding
                       const string** respp // Decoding Indicator 
    )
{
    // By default, there is no encoding (7bit,8bit,raw). Also in case of 
    // decoding error
    *respp = &body;

    if (!stringlowercmp("quoted-printable", cte)) {
        if (!qp_decode(body, decoded)) {
            LOGERR("decodeBody: quoted-printable decoding failed !\n");
            LOGDEB("      Body: \n" << body << "\n");
            return false;
        }
        *respp = &decoded;
    } else if (!stringlowercmp("base64", cte)) {
        if (!base64_decode(body, decoded)) {
            // base64 encoding errors are actually relatively common
            LOGERR("decodeBody: base64 decoding failed !\n");
            LOGDEB("      Body: \n" << body << "\n");
            return false;
        }
        *respp = &decoded;
    }
    return true;
}

bool MimeHandlerMail::processAttach()
{
    LOGDEB("MimeHandlerMail::processAttach() m_idx " << m_idx << "\n");
    if (!m_havedoc)
        return false;
    if (m_idx >= (int)m_attachments.size()) {
        m_havedoc = false;
        return false;
    }
    MHMailAttach *att = m_attachments[m_idx];

    m_metaData[cstr_dj_keymt] = att->m_contentType;
    m_metaData[cstr_dj_keyorigcharset] = att->m_charset;
    m_metaData[cstr_dj_keycharset] = att->m_charset;
    m_metaData[cstr_dj_keyfn] = att->m_filename;
    m_metaData[cstr_dj_keytitle] = att->m_filename + "  (" + m_subject + ")";
    LOGDEB1("  processAttach:ct [" << att->m_contentType << "] cs [" <<
            att->m_charset << "] fn [" << att->m_filename << "]\n");

    // Erase current content and replace
    string& body = m_metaData[cstr_dj_keycontent];
    body.clear();
    att->m_part->getBody(body, 0, att->m_part->bodylength);
    {
        string decoded;
        const string *bdp;
        if (!decodeBody(att->m_contentTransferEncoding, body, decoded, &bdp)) {
            return false;
        }
        if (bdp != &body)
            body.swap(decoded);
    }

    // Special case for application/octet-stream: try to better
    // identify content, using file name if set
    if (m_metaData[cstr_dj_keymt] == "application/octet-stream" &&
        !m_metaData[cstr_dj_keyfn].empty()) {
        string mt = mimetype(m_metaData[cstr_dj_keyfn], m_config, false);
        if (!mt.empty()) 
            m_metaData[cstr_dj_keymt] = mt;
    }

    // Special case for text/plain content. Internfile should deal
    // with this but it expects text/plain to be utf-8 already, so we
    // handle the transcoding if needed. Same kind of issue for the MD5
    if (m_metaData[cstr_dj_keymt] == cstr_textplain) {
        if (!txtdcode("MimeHandlerMail::processAttach")) {
            body.clear();
        } else if (!m_forPreview) {
            string md5, xmd5;
            MD5String(body, md5);
            m_metaData[cstr_dj_keymd5] = MD5HexPrint(md5, xmd5);
        }
    }

    // Ipath
    m_metaData[cstr_dj_keyipath] = std::to_string(m_idx);

    return true;
}

// Transform a single message into a document. The subject becomes the
// title, and any simple body part with a content-type of text or html
// and content-disposition inline gets concatenated as text.
// 
// If depth is not zero, we're called recursively for an
// message/rfc822 part and we must not touch the doc fields except the
// text
bool MimeHandlerMail::processMsg(Binc::MimePart *doc, int depth)
{
    LOGDEB2("MimeHandlerMail::processMsg: depth " << depth << "\n");
    if (depth++ >= maxdepth) {
        // Have to stop somewhere
        LOGINFO("MimeHandlerMail::processMsg: maxdepth " << maxdepth <<
                " exceeded\n");
        // Return true anyway, better to index partially than not at all
        return true;
    }
        
    // Handle some headers. 
    string& text = m_metaData[cstr_dj_keycontent];
    Binc::HeaderItem hi;
    string decoded;
    if (doc->h.getFirstHeader("From", hi)) {
        rfc2047_decode(hi.getValue(), decoded);
        if (preview())
            text += string("From: ");
        text += decoded + cstr_newline;
        if (depth == 1) {
            m_metaData[cstr_dj_keyauthor] = decoded;
        }
    }
    if (doc->h.getFirstHeader("To", hi)) {
        rfc2047_decode(hi.getValue(), decoded);
        if (preview())
            text += string("To: ");
        text += decoded + cstr_newline;
        if (depth == 1) {
            m_metaData[cstr_dj_keyrecipient] = decoded;
        }
    }
    if (doc->h.getFirstHeader("Cc", hi)) {
        rfc2047_decode(hi.getValue(), decoded);
        if (preview())
            text += string("Cc: ");
        text += decoded + cstr_newline;
        if (depth == 1) {
            m_metaData[cstr_dj_keyrecipient] += " " + decoded;
        }
    }
#if 0 // 2022-10: this does not appear to be used anywhere
// cstr.h DEF_CSTR(dj_keymsgid, "msgid");
    if (doc->h.getFirstHeader("Message-Id", hi)) {
        if (depth == 1) {
            m_metaData[cstr_dj_keymsgid] =  hi.getValue();
            trimstring(m_metaData[cstr_dj_keymsgid], "<>");
        }
    }
#endif
    if (doc->h.getFirstHeader("Date", hi)) {
        rfc2047_decode(hi.getValue(), decoded);
        if (depth == 1) {
            time_t t = rfc2822DateToUxTime(decoded);
            if (t != (time_t)-1) {
                m_metaData[cstr_dj_keymd] = std::to_string(t);
            } else {
                // Leave mtime field alone, ftime will be used instead.
                LOGDEB("rfc2822Date...: failed: [" << decoded << "]\n");
            }
        }
        if (preview())
            text += string("Date: ");
        text += decoded + cstr_newline;
    }
    if (doc->h.getFirstHeader("Subject", hi)) {
        rfc2047_decode(hi.getValue(), decoded);
        if (depth == 1) {
            m_metaData[cstr_dj_keytitle] = decoded;
            m_subject = decoded;
        }
        if (preview())
            text += string("Subject: ");
        text += decoded + cstr_newline;
    }

    // Check for the presence of configured additional headers and possibly
    // add them to the metadata (with appropriate field name).
    if (!m_addProcdHdrs.empty()) {
        for (auto& it : m_addProcdHdrs) {
            if (!it.second.empty() && doc->h.getFirstHeader(it.first, hi)) {
                rfc2047_decode(hi.getValue(), m_metaData[it.second]);
            }
        }
    }

    text += '\n';
    m_startoftext = text.size();
    LOGDEB2("MimeHandlerMail::processMsg:ismultipart " <<
            doc->isMultipart() << " mime subtype '"<<doc->getSubType()<< "'\n");
    walkmime(doc, depth);

    LOGDEB2("MimeHandlerMail::processMsg:text:[" << m_metaData[cstr_dj_keycontent] << "]\n");
    return true;
}

// Recursively walk the message mime parts and concatenate all the
// inline html or text that we find anywhere.  
//
// RFC2046 reminder: 
// Top level media types: 
//      Simple:    text, image, audio, video, application, 
//      Composite: multipart, message.
// 
// multipart can be mixed, signed, alternative, parallel, digest.
// message/rfc822 may also be of interest.
void MimeHandlerMail::walkmime(Binc::MimePart* doc, int depth)
{
    LOGDEB2("MimeHandlerMail::walkmime: depth " << depth << "\n");
    if (depth++ >= maxdepth) {
        LOGINFO("walkmime: max depth (" << maxdepth << ") exceeded\n");
        return;
    }

    string& out = m_metaData[cstr_dj_keycontent];

    if (doc->isMultipart()) {
        LOGDEB2("walkmime: ismultipart " << doc->isMultipart() <<
                " subtype '" << doc->getSubType() << "'\n");
        // We only handle alternative, related and mixed (no digests). 
        std::vector<Binc::MimePart>::iterator it;

        if (!stringicmp("mixed", doc->getSubType()) || 
            !stringicmp("signed", doc->getSubType()) ||
            !stringicmp("related", doc->getSubType())) {
            // Multipart mixed and related:  process each part.
            for (it = doc->members.begin(); it != doc->members.end();it++) {
                walkmime(&(*it), depth);
            }

        } else if (!stringicmp("alternative", doc->getSubType())) {
            // Multipart/alternative: look for a text/plain part, then html.
            // Process if found
            std::vector<Binc::MimePart>::iterator ittxt, ithtml;
            ittxt = ithtml = doc->members.end();
            int i = 1;
            for (it = doc->members.begin(); 
                 it != doc->members.end(); it++, i++) {
                // Get and parse content-type header
                Binc::HeaderItem hi;
                if (!it->h.getFirstHeader("Content-Type", hi)) {
                    LOGDEB("walkmime:no ctent-type header for part "<<i<< "\n");
                    continue;
                }
                MimeHeaderValue content_type;
                parseMimeHeaderValue(hi.getValue(), content_type);
                LOGDEB2("walkmime: C-type: " << content_type.value << "\n");
                if (!stringlowercmp(cstr_textplain, content_type.value))
                    ittxt = it;
                else if (!stringlowercmp("text/html", content_type.value)) 
                    ithtml = it;
            }
            if (ittxt != doc->members.end() && ittxt->bodylength) {
                LOGDEB2("walkmime: alternative: choose text/plain. Size: " <<
                        ittxt->bodylength << endl);
                walkmime(&(*ittxt), depth);
            } else if (ithtml != doc->members.end()) {
                LOGDEB2("walkmime: alternative: choose text/html.\n");
                walkmime(&(*ithtml), depth);
            }
        }
        return;
    } 
    
    // Part is not multipart: it must be either simple or message. Take
    // a look at interesting headers and a possible filename parameter

    // Get and parse content-type header.
    Binc::HeaderItem hi;
    string ctt = cstr_textplain;
    if (doc->h.getFirstHeader("Content-Type", hi)) {
        ctt = hi.getValue();
    }
    LOGDEB2("walkmime:content-type: " << ctt << "\n");
    MimeHeaderValue content_type;
    parseMimeHeaderValue(ctt, content_type);
            
    // Get and parse Content-Disposition header
    string ctd = "inline";
    if (doc->h.getFirstHeader("Content-Disposition", hi)) {
        ctd = hi.getValue();
    }
    MimeHeaderValue content_disposition;
    parseMimeHeaderValue(ctd, content_disposition);
    LOGDEB2("Content_disposition:[" << content_disposition.value << "]\n");
    string dispindic;
    if (stringlowercmp("inline", content_disposition.value))
        dispindic = "Attachment";
    else 
        dispindic = "Inline";

    // See if we have a filename.
    string filename;
    map<string,string>::const_iterator it;
    it = content_disposition.params.find(string("filename"));
    if (it != content_disposition.params.end())
        filename = it->second;
    if (filename.empty()) {
        it = content_type.params.find(string("name"));
        if (it != content_type.params.end())
            filename = it->second;
    }

    // Note: I have never seen anything useful in the Content-Description field.
    // So it's not processed. Use rfc2047 to decode if we ever do it.
        
    if (doc->isMessageRFC822()) {
        LOGDEB2("walkmime: message/RFC822 part\n");
        
        // The first part is the already parsed message.  Call
        // processMsg instead of walkmime so that mail headers get
        // printed. The depth will tell it what to do
        if (doc->members.empty()) {
            //??
            return;
        }
        out += "\n";
        if (m_forPreview)
            out += "[" + dispindic + " " + content_type.value + ": ";
        out += filename;
        if (m_forPreview)
            out += "]";
        out += "\n\n";
        processMsg(&doc->members[0], depth);
        return;
    }

    // "Simple" part. 
    LOGDEB2("walkmime: simple  part\n");
    // Normally the default charset is us-ascii. But it happens that 8
    // bit chars exist in a message that is stated as us-ascii. Ie the
    // mailer used by yahoo support ('KANA') does this. We could
    // convert to iso-8859 only if the transfer-encoding is 8 bit, or
    // test for actual 8 bit chars, but what the heck, le'ts use
    // 8859-1 (actually CP1252 which is compatible, but with more
    // useful chars) as default.
    string charset;
    it = content_type.params.find(cstr_mail_charset);
    if (it != content_type.params.end())
        charset = it->second;
    if (charset.empty() || 
        !stringlowercmp("us-ascii", charset) || 
        !stringlowercmp("default", charset) || 
        !stringlowercmp("x-user-defined", charset) || 
        !stringlowercmp("x-unknown", charset) || 
        !stringlowercmp("unknown", charset) ) {
        if (!m_config->getConfParam("maildefcharset", charset))
            charset = "CP1252";
    }

    // Content transfer encoding
    string cte = "7bit";
    if (doc->h.getFirstHeader("Content-Transfer-Encoding", hi)) {
        cte = hi.getValue();
    } 

    // If the Content-Disposition is not inline, we treat it as
    // attachment, as per rfc2183. 
    // If it is inline but not text or html, same thing.
    // Some early MIME msgs have "text" instead of "text/plain" as type
    if (stringlowercmp("inline", content_disposition.value) ||
        (stringlowercmp(cstr_textplain, content_type.value) && 
         stringlowercmp("text", content_type.value) && 
         stringlowercmp("text/html", content_type.value)) ) {
        if (!filename.empty()) {
            out += "\n";
            if (m_forPreview)
                out += "[" + dispindic + " " + content_type.value + ": ";
            out += filename;
            if (m_forPreview)
                out += "]";
            out += "\n\n";
        }
        MHMailAttach *att = new MHMailAttach;
        if (att == 0) {
            LOGERR("Out of memory\n");
            return;
        }
        att->m_contentType = content_type.value;
        stringtolower(att->m_contentType);
        att->m_filename = filename;
        att->m_charset = charset;
        att->m_contentTransferEncoding = cte;
        att->m_part = doc;
        LOGDEB("walkmime: attachmnt: ct [" << att->m_contentType <<
               "] cte [" << att->m_contentTransferEncoding << "] cs [" <<
               att->m_charset << "] fn [" << filename << "]\n");
        m_attachments.push_back(att);
        return;
    }

    // We are dealing with an inline part of text/plain or text/html
    // type. We can't just return a text or html subdoc and let the
    // filter stack work: this would create another subdocument, but
    // we want instead to decode a body part of this message document.

    LOGDEB2("walkmime: final: body start offset " <<
            doc->getBodyStartOffset()<<", length "<<doc->getBodyLength()<<"\n");
    string body;
    doc->getBody(body, 0, doc->bodylength);
    {
        string decoded;
        const string *bdp;
        if (!decodeBody(cte, body, decoded, &bdp)) {
            LOGERR("MimeHandlerMail::walkmime: failed decoding body\n");
        }
        if (bdp != &body)
            body.swap(decoded);
    }

    // Handle html stripping and transcoding to utf8
    if (!stringlowercmp("text/html", content_type.value)) {
        MimeHandlerHtml mh(m_config, "1234");
        mh.set_property(Dijon::Filter::OPERATING_MODE, 
                        m_forPreview ? "view" : "index");
        mh.set_property(Dijon::Filter::DEFAULT_CHARSET, charset);
        mh.set_document_string("text/html", body);
        mh.next_document();
        map<string, string>::const_iterator it = 
            mh.get_meta_data().find(cstr_dj_keycontent);
        if (it != mh.get_meta_data().end())
            out += it->second;
    } else {
        string utf8;
        // Transcode to utf-8 
        LOGDEB1("walkmime: transcoding from " << charset << " to UTF-8\n");
        if (!transcode(body, utf8, charset, cstr_utf8)) {
            LOGERR("walkmime: transcode failed from cs '" << charset <<
                   "' to UTF-8\n");
            out += body;
        } else {
            out += utf8;
        }
    }

    if (out.length() && out[out.length()-1] != '\n')
        out += '\n';
    
    LOGDEB2("walkmime: out now: [" << out << "]\n");
}