imap_fetchstructure

Anonymous (2013-06-20 22:52:59)

I improved the code from this post (
http://www.php.net/manual/en/function.imap-fetchstructure.php#51497) a bit, now it handles multi embedded email correctly, and gives you a content with not marking character, like the '=' (soft line break).

<?php
//return is structured as
// $mail = array ( 'body' => ...
//          'attachment' => array() );
function _get_body_attach($mbox, $mid) {
    $struct = imap_fetchstructure($mbox, $mid);

    $parts = $struct->parts;
    $i = 0;
    if (!$parts) { /* Simple message, only 1 piece */
        $attachment = array(); /* No attachments */
        $content = imap_body($mbox, $mid);
    } else { /* Complicated message, multiple parts */

        $endwhile = false;

        $stack = array(); /* Stack while parsing message */
        $content = "";    /* Content of message */
        $attachment = array(); /* Attachments */

        while (!$endwhile) {
            if (!$parts[$i]) {
                if (count($stack) > 0) {
                    $parts = $stack[count($stack)-1]["p"];
                    $i     = $stack[count($stack)-1]["i"] + 1;
                    array_pop($stack);
                } else {
                    $endwhile = true;
                }
            }

            if (!$endwhile) {
                /* Create message part first (example '1.2.3') */
                $partstring = "";
                foreach ($stack as $s) {
                    $partstring .= ($s["i"]+1) . ".";
                }
                $partstring .= ($i+1);

                if (strtoupper($parts[$i]->disposition) == "ATTACHMENT" || strtoupper($parts[$i]->disposition) == "INLINE") { /* Attachment or inline images */
                    $filedata = imap_fetchbody($mbox, $mid, $partstring);
                    if ( $filedata != '' ) {
                        // handles base64 encoding or plain text
                        $decoded_data = base64_decode($filedata);
                        if ( $decoded_data == false ) {
                            $attachment[] = array("filename" => $parts[$i]->parameters[0]->value,
                                "filedata" => $filedata);
                        } else {
                            $attachment[] = array("filename" => $parts[$i]->parameters[0]->value,
                                "filedata" => $decoded_data);
                        }
                    }
                } elseif (strtoupper($parts[$i]->subtype) == "PLAIN" && strtoupper($parts[$i+1]->subtype) != "HTML") { /* plain text message */
                    $content .= imap_fetchbody($mbox, $mid, $partstring);
                } elseif ( strtoupper($parts[$i]->subtype) == "HTML" ) {
                    /* HTML message takes priority */
                    $content .= imap_fetchbody($mbox, $mid, $partstring);
                }
            }

            if ($parts[$i]->parts) {
                if ( $parts[$i]->subtype != 'RELATED' ) {
                    // a glitch: embedded email message have one additional stack in the structure with subtype 'RELATED', but this stack is not present when using imap_fetchbody() to fetch parts.
                    $stack[] = array("p" => $parts, "i" => $i);
                }
                $parts = $parts[$i]->parts;
                $i = 0;
            } else {
                $i++;
            }
        } /* while */
    } /* complicated message */

    $ret = array();
    $ret['body'] = quoted_printable_decode($content);
    $ret['attachment'] = $attachment;
    return $ret;
}
?>

raulggonzalez at gmail dot com (2010-09-21 05:54:18)

Hello eveybody,

After reading plenty of posts (some good, some others not so good) I want to post a function to find out whether an email is coming with attachments

The function can be called passing the object returned by imap_fetchstructure() or any of its parts (recursively)

<?php
$inbox = imap_open($hostname,$username,$password) or die('Cannot connect to mail: ' . imap_last_error());
$struct = imap_fetchstructure($inbox,$uid,FT_UID);
$existAttachments = existAttachment($struct);

function existAttachment($part){
    if (isset($part->parts)){
        foreach ($part->parts as $partOfPart){
            existAttachment($partOfPart);
        }
    }
    else{
        if (isset($part->disposition)){
            if ($part->disposition == 'attachment'){
                echo '<p>' . $part->dparameters[0]->value . '</p>';
// here you can create a link to the file whose name is  $part->dparameters[0]->value to download it
                return true;
            }
        }
    }
}
?>

Hopefully somebody can use it as well

burninleo at gmx dot net (2010-08-03 09:42:25)

Use imap_errors() to get rid of sticky error messages about incorrect MIME headers.

Rui Torre rui_sharkye at hotmail dot com (2008-11-02 02:34:29)

I've made this two functions to decode mime headers and so far they are working just fine decoding all headers. Please let me know if you find any error. Hope they are useful for you too.

<?php

function decode_headers($cabecalho){
        while (preg_match('/^(.*)=\?([^?]*)\?(Q|B)\?(.*)$/Ui', $cabecalho)){
            $cabecalho = preg_replace("/(\t|\r|\n)/", "", $cabecalho); 
            $cabecalho = preg_replace("/\?(Q|B)\?=/Ui", "?\\1? =", $cabecalho); 
               
            $partes_cabecalho = explode("?=", $cabecalho);
            
            $resultado = "";
            foreach ($partes_cabecalho as $texto){
                $texto = preg_replace("/\?(Q|B)\?\s=/Ui", "?\\1?=", $texto);
                $texto = preg_replace("/\?(Q|B)\?/Ui", "=?\\1=?", $texto);
                
                if (preg_match('/^(.*)=\?([^?]*)=\?(Q|B)=\?(.*)$/Ui', $texto)){
                    $partes_texto = explode ("=?", $texto);  
                    
                    $parte = descodificar_parte($partes_texto[0], $partes_texto[1]);
                    $parte .= descodificar_parte($partes_texto[3], $partes_texto[2]);
                } else { 
                    $parte = $texto;  
                }  
                
                if (strtoupper($partes_texto[1]) == "UTF-8"){ $parte = utf8_decode($parte);}  
                $resultado .= $parte;  
            }
            $cabecalho = $resultado;
        } 
        
        return trim($cabecalho);       
    }
    
    function descodificar_parte($texto, $codificacao){
        if (trim($texto) == ""){ return "";}
        
        if (ucfirst($codificacao) == "B"){
            $texto = base64_decode($texto);
            $texto = htmlspecialchars($texto);
        } else if (ucfirst($codificacao) == "Q"){
            $texto = str_replace("_", " ", $texto); 
            $texto = quoted_printable_decode($texto);  
        } 
        return $texto;
    }

?>

Rui Torre rui_sharkye at hotmail dot com (2008-10-29 10:38:42)

I'm programing an webmail program and trying to make the less code possible, so I've made this function to get all mail parts

Hope it helps someone or if you manage to improve this function please post me.

<?php
$caixa_mail = @imap_open ("{localhost:143/notls}INBOX", $user, $password) or die ("It was not possible to establish the imap connection.");

$uid_msg = 1;

$partes_mensagem = obter_partes($caixa_mail, $uid_msg);

function obter_partes($imap, $uid, $estrutura = false, $nparte = 1){
    if (!$estrutura){ $estrutura = imap_fetchstructure($imap, $uid, FT_UID);}        
    
    //This shows the headers part number (I need to work a little more on this part)
    
    if ($estrutura->parts > 0){    
        $decimas = explode(".", $nparte);
        $decimas[count($decimas)-1] -= 1;
        $np = implode(".", $decimas); 
    } else {
        $np = $nparte;
    }
    
    //Its a possible message parte (plain or html)
    if ($estrutura->type == 0){ print "> mensagem: ";}  
    print "$np - ".$estrutura->subtype."<br>";
 
         
    if (count($estrutura->parts) > 0){            
        foreach ($estrutura->parts as $n => $parte){
            if ($n >= 1){
                $arr_decimas = explode(".", $nparte);
                $arr_decimas[count($arr_decimas)-1] += 1;
                $nparte = implode(".", $arr_decimas);
            }
                
            if ($parte->type != 1){
                if ($parte->type == 0){ print "mensagem: ";}
                print "$nparte - ".$parte->subtype."<br>";
            }     
                
            if (count($parte->parts) > 0){
                if ($parte->type == 1){
                    obter_partes($imap, $uid, $parte, $nparte.".".($n+1));
                } else {
                    foreach ($parte->parts as $idx => $prt){
                        obter_partes($imap, $uid, $prt, $nparte.".".($idx+1));
                    }    
                }
            } 
        }
    } 
}
?>

david at hundsness dot com (2008-09-12 01:00:47)

[EDITORS: This post corrects mistakes in my previous post, so please delete #85486 on 02-Sep-2008 02:31]

Here is code to parse and decode all types of messages, including attachments. I've been using something like this for a while now, so it's pretty robust.

<?
function getmsg($mbox,$mid) {
    // input $mbox = IMAP stream, $mid = message id
    // output all the following:
    global $htmlmsg,$plainmsg,$charset,$attachments;
    // the message may in $htmlmsg, $plainmsg, or both
    $htmlmsg = $plainmsg = $charset = '';
    $attachments = array();

    // HEADER
    $h = imap_header($mbox,$mid);
    // add code here to get date, from, to, cc, subject...

    // BODY
    $s = imap_fetchstructure($mbox,$mid);
    if (!$s->parts)  // not multipart
        getpart($mbox,$mid,$s,0);  // no part-number, so pass 0
    else {  // multipart: iterate through each part
        foreach ($s->parts as $partno0=>$p)
            getpart($mbox,$mid,$p,$partno0+1);
    }
}

function getpart($mbox,$mid,$p,$partno) {
    // $partno = '1', '2', '2.1', '2.1.3', etc if multipart, 0 if not multipart
    global $htmlmsg,$plainmsg,$charset,$attachments;

    // DECODE DATA
    $data = ($partno)?
        imap_fetchbody($mbox,$mid,$partno):  // multipart
        imap_body($mbox,$mid);  // not multipart
    // Any part may be encoded, even plain text messages, so check everything.
    if ($p->encoding==4)
        $data = quoted_printable_decode($data);
    elseif ($p->encoding==3)
        $data = base64_decode($data);
    // no need to decode 7-bit, 8-bit, or binary

    // PARAMETERS
    // get all parameters, like charset, filenames of attachments, etc.
    $params = array();
    if ($p->parameters)
        foreach ($p->parameters as $x)
            $params[ strtolower( $x->attribute ) ] = $x->value;
    if ($p->dparameters)
        foreach ($p->dparameters as $x)
            $params[ strtolower( $x->attribute ) ] = $x->value;

    // ATTACHMENT
    // Any part with a filename is an attachment,
    // so an attached text file (type 0) is not mistaken as the message.
    if ($params['filename'] || $params['name']) {
        // filename may be given as 'Filename' or 'Name' or both
        $filename = ($params['filename'])? $params['filename'] : $params['name'];
        // filename may be encoded, so see imap_mime_header_decode()
        $attachments[$filename] = $data;  // this is a problem if two files have same name
    }

    // TEXT
    elseif ($p->type==0 && $data) {
        // Messages may be split in different parts because of inline attachments,
        // so append parts together with blank row.
        if (strtolower($p->subtype)=='plain')
            $plainmsg .= trim($data) ."\n\n";
        else
            $htmlmsg .= $data ."<br><br>";
        $charset = $params['charset'];  // assume all parts are same charset
    }

    // EMBEDDED MESSAGE
    // Many bounce notifications embed the original message as type 2,
    // but AOL uses type 1 (multipart), which is not handled here.
    // There are no PHP functions to parse embedded messages,
    // so this just appends the raw source to the main message.
    elseif ($p->type==2 && $data) {
        $plainmsg .= trim($data) ."\n\n";
    }

    // SUBPART RECURSION
    if ($p->parts) {
        foreach ($p->parts as $partno0=>$p2)
            getpart($mbox,$mid,$p2,$partno.'.'.($partno0+1));  // 1.2, 1.2.1, etc.
    }
}
?>

david at hundsness dot com (2008-09-02 02:31:01)

Here is code to parse and decode all types of messages, including attachments. I've been using something like this for awhile now, so it's pretty robust.

<?
function getmsg($mbox,$mid) {
    // input $mbox = IMAP stream, $mid = message id
    // output all the following:
    global $charset,$htmlmsg,$plainmsg,$attachments;
    $htmlmsg = $plainmsg = $charset = '';
    $attachments = array();

    // HEADER
    $h = imap_header($mbox,$mid);
    // add code here to get date, from, to, cc, subject...

    // BODY
    $s = imap_fetchstructure($mbox,$mid);
    if (!$s->parts)  // simple
        getpart($mbox,$mid,$s,0);  // pass 0 as part-number
    else {  // multipart: cycle through each part
        foreach ($s->parts as $partno0=>$p)
            getpart($mbox,$mid,$p,$partno0+1);
    }
}

function getpart($mbox,$mid,$p,$partno) {
    // $partno = '1', '2', '2.1', '2.1.3', etc for multipart, 0 if simple
    global $htmlmsg,$plainmsg,$charset,$attachments;

    // DECODE DATA
    $data = ($partno)?
        imap_fetchbody($mbox,$mid,$partno):  // multipart
        imap_body($mbox,$mid);  // simple
    // Any part may be encoded, even plain text messages, so check everything.
    if ($p->encoding==4)
        $data = quoted_printable_decode($data);
    elseif ($p->encoding==3)
        $data = base64_decode($data);

    // PARAMETERS
    // get all parameters, like charset, filenames of attachments, etc.
    $params = array();
    if ($p->parameters)
        foreach ($p->parameters as $x)
            $params[strtolower($x->attribute)] = $x->value;
    if ($p->dparameters)
        foreach ($p->dparameters as $x)
            $params[strtolower($x->attribute)] = $x->value;

    // ATTACHMENT
    // Any part with a filename is an attachment,
    // so an attached text file (type 0) is not mistaken as the message.
    if ($params['filename'] || $params['name']) {
        // filename may be given as 'Filename' or 'Name' or both
        $filename = ($params['filename'])? $params['filename'] : $params['name'];
        // filename may be encoded, so see imap_mime_header_decode()
        $attachments[$filename] = $data;  // this is a problem if two files have same name
    }

    // TEXT
    if ($p->type==0 && $data) {
        // Messages may be split in different parts because of inline attachments,
        // so append parts together with blank row.
        if (strtolower($p->subtype)=='plain')
            $plainmsg. = trim($data) ."\n\n";
        else
            $htmlmsg. = $data ."<br><br>";
        $charset = $params['charset'];  // assume all parts are same charset
    }

    // EMBEDDED MESSAGE
    // Many bounce notifications embed the original message as type 2,
    // but AOL uses type 1 (multipart), which is not handled here.
    // There are no PHP functions to parse embedded messages,
    // so this just appends the raw source to the main message.
    elseif ($p->type==2 && $data) {
        $plainmsg. = $data."\n\n";
    }

    // SUBPART RECURSION
    if ($p->parts) {
        foreach ($p->parts as $partno0=>$p2)
            getpart($mbox,$mid,$p2,$partno.'.'.($partno0+1));  // 1.2, 1.2.1, etc.
    }
}
?>

misc at e2007 dot cynergi dot com (2007-10-21 07:53:47)

First of all, for a while now (not sure since which PHP version) a new "primary body type" constant exists: TYPEMODEL (int(7)) and TYPEOTHER became int(8). Furthermore the use of these constants can be misleading. The issue is that "other" MIME-types aren't just ==TYPEOTHER, but rather >=TYPEOTHER. I've added a PHP feature request to help solve this (http://bugs.php.net/43061).
To learn more, read this extract of a conversation I've had with Mark Crispin of IMAP's extension mailing list at UW (e-mail addresses deleted):
----
Mark Crispin:
I certainly understand the frustrations that end users experience with apparent mutual finger-pointing. The way to solve it is through communication.
For what it's worth, here are the current semantics of type codes:
0 TEXT
1 MULTIPART
2 MESSAGE
3 APPLICATION
4 AUDIO
5 IMAGE
6 VIDEO
7 MODEL
8 X-UNKNOWN (or expansion types filed up)
9 first expansion type
...
TYPEMAX last expansion type (currently 15)
On Wed, 17 Oct 2007, Cynergi wrote:
(...)
> -----Original Message-----
> Sent: ter?a-feira, 16 de Outubro de 2007 17:35
> To: Cynergi
> Subject: Re: [Imap-uw] Apparent bug in imap-2006k.DEV.SNAP-0710121414
> (and older?)
>
> Thank you for your report.
>
> This is not a bug. The c-client library is designed to behave this way.
>
> MIME types and encodings are, by definition in MIME, open-ended. From
> time to time, the IETF defines new MIME types. It is desirable that
> c-client be able to handle these without having to make a source code
> modification to c-client.
>
> To allow applications to support types and encodings that are unknown
> to c-client, c-client will automatically add a (limited) number of
> unknown types and encodings to its tables before resorting to
> TYPEOTHER and ENCOTHER. The names of these added types and encodings
> are available in the body_types[] and body_encodings[] arrays.
>
> Put another way, TYPEOTHER and ENCOTHER are only used if c-client is
> overflowed with unknown types and/or encodings.
>
> If the PHP developers had asked me about this, I would have explained
> this to them. Unfortunately, they have a habit of labelling
> unexpected behaviors as "bugs" rather than seeking answers. I have
> tried to post amended information on the PHP bugzilla in the past, but
> was rewarded with a "you are not authorized to do so" so I've given up.
>
> Hence, a more correct behavior for PHP is not to use type code values,
> but instead to use the body_types[type_code] string.
>
> I hope that this information is helpful.
>
> On Mon, 15 Oct 2007, Cynergi wrote:
>
>> Dear Sirs,
>>
>> I only want to make a bug report. I use PHP for development so I
>> really won't be able to keep up with messages from this list. Please
>> don't take offense if I unsubscribe in a few days after sending this
> message.
>>
>> While developing a Webmail for www.cynergi.com, we started coming
>> upon some
>> (spam) messages with bad MIME types. However PHP (via your c-client
>> library) reported a MIME type code of 9 instead of TYPEOTHER (8).
>>
>> Looking at your library I seem to have found the bug. From a comment
>> in our source code:
>>
>> Unknown MIME-types (such as "25-bit") are returned as int(9) due to
> a
>> bug in c-client's rfc822.c "body_types" array that defines TYPEOTHER
>> as "X-UNKNOWN"; then when imap4r1.c goes to match a MIME-type and
>> can't match any of body_types' strings (including TYPEOTHER's), it
>> returns the next available integer (9).
>> This same bug also seems to apply to ENCOTHER.
>>
>> I am unaware if the body_types array is also used to CREATE
>> MIME-types. If so, deleting "X-UNKNOWN" from there won't be the
>> solution (the solution will then have to be fixing imap4r1.c's code).
>>
>> Thank you for your time, and for such a great open-source library!!
>> :-)
>>
>> Pedro Freire
>> Cynergi

phpnet,a,emailaddress,cjb,net (2007-09-26 19:13:51)

Another comment to inform people about something that should really be in the function description:
imap_fetchstructure() downloads the entire email, attachments and all, rather than just the structure.
I guess it's an undocumented feature, not a bug.
I had assumed that the script would have only downloaded the amount of data that was returned, but my script downloaded a cumulative 2.5gig before i noticed. Hopefully no-one else will have this happen.

oersoep at gmail dot com (2006-01-29 23:58:57)

Please keep in mind that the "parameters" array is a stdClass object when there are no parameters, and NOT a zero-size array.
In PHP5 you'll get this error when iterating structure->parts[x]->parameters if there aren't any parameters for this part:
PHP Fatal error: Cannot use object of type stdClass as array in file.php on line 100

sirber at detritus dot qc dot ca (2006-01-04 11:48:58)

"Primary body type" of "unknown/unknown" will be int(9).

hans at lintoo dot dk (2005-09-22 07:07:46)

To fetch a single part (for example to investigate a charset or something like that)

<?php
class StructureEngine {
    private $structureObject;
    public function __construct($structureInfo) {
        $this->structureObject = $structureInfo;
        
    }

    // snip....

    public function getPart($partNum) {
        $path = split("[.]",$partNum);
        $currentPart = $this->structureObject;
        foreach ($path as $key => $num) {
            $currentPart = $currentPart->parts[$num-1];
        }
        return $currentPart;
    }
}
?>

hans at lintoo dot dk (2005-09-22 06:14:33)

A class for searching through a stucture object. 
You can search matching 1-3 parameters or use the methods included. 
Feel free to use it, correct it or extend it if you please. Enjoy!

http://lintoo.dk/public/structureengine.class.phps

The code was too wide for this manual, so I hope the link above will do. Here a quick overview instead:

<?php
class StructureEngine {
    private $structureObject;
    public function __construct($structureInfo) {}
    public function locatePlain() {}
    public function locateHTML() {}
    public function locateAttachments() {}
    private function checkParam($part, $type, $value) {}
    private function findParts($partsArray, $prefix, $param1Type = 'type', $param1Value = 0, $param2Type = null, ...) {}
}
?>

john at vetsurgeon dot org dot uk (2005-07-04 09:07:37)

A little script I threw together to break a message down and process it into a usable array

<?
//script will fetch an email identified by $msgid, and parse the its parts into an
//array $partsarray
//structure of array:
//$partsarray[<name of part>][<attachment/text>]
//if attachment- subarray is [filename][binary data]
//if text- subarray is [type of text(HTML/PLAIN)][text string]

//i.e.
//$partsarray[3.1][attachment][filename]=filename of attachment in part 3.1
//$partsarray[3.1][attachment][binary]=binary data of attachment in part 3.1
//$partsarray[2][text][type]=type of text in part 2
//$partsarray[2][text][string]=decoded text string in part 2
//$partsarray[not multipart][text][string]=decoded text string in message that isn't multipart

function parsepart($p,$i){
    global $link,$msgid,$partsarray;
    //where to write file attachments to:
    $filestore = '[full/path/to/attachment/store/(chmod777)]';

    //fetch part
    $part=imap_fetchbody($link,$msgid,$i);
    //if type is not text
    if ($p->type!=0){
        //DECODE PART        
        //decode if base64
        if ($p->encoding==3)$part=base64_decode($part);
        //decode if quoted printable
        if ($p->encoding==4)$part=quoted_printable_decode($part);
        //no need to decode binary or 8bit!
        
        //get filename of attachment if present
        $filename='';
        // if there are any dparameters present in this part
        if (count($p->dparameters)>0){
            foreach ($p->dparameters as $dparam){
                if ((strtoupper($dparam->attribute)=='NAME') ||(strtoupper($dparam->attribute)=='FILENAME')) $filename=$dparam->value;
                }
            }
        //if no filename found
        if ($filename==''){
            // if there are any parameters present in this part
            if (count($p->parameters)>0){
                foreach ($p->parameters as $param){
                    if ((strtoupper($param->attribute)=='NAME') ||(strtoupper($param->attribute)=='FILENAME')) $filename=$param->value;
                    }
                }
            }
        //write to disk and set partsarray variable
        if ($filename!=''){
            $partsarray[$i][attachment] = array('filename'=>$filename,'binary'=>$part);
            $fp=fopen($filestore.$filename,"w+");
            fwrite($fp,$part);
            fclose($fp);
            }
    //end if type!=0        
    }
    
    //if part is text
    else if($p->type==0){
        //decode text
        //if QUOTED-PRINTABLE
        if ($p->encoding==4) $part=quoted_printable_decode($part);
        //if base 64
        if ($p->encoding==3) $part=base64_decode($part);
        
        //OPTIONAL PROCESSING e.g. nl2br for plain text
        //if plain text

        if (strtoupper($p->subtype)=='PLAIN')1;
        //if HTML
        else if (strtoupper($p->subtype)=='HTML')1;
        $partsarray[$i][text] = array('type'=>$p->subtype,'string'=>$part);
    }
    
    //if subparts... recurse into function and parse them too!
    if (count($p->parts)>0){
        foreach ($p->parts as $pno=>$parr){
            parsepart($parr,($i.'.'.($pno+1)));            
            }
        }
return;
}

//open resource
$link=imap_open("{localhost:110/pop3}INBOX",'[YOUR USERNAME]','[YOUR PASSWORD]');

//fetch structure of message
$s=imap_fetchstructure($link,$msgid);

//see if there are any parts
if (count($s->parts)>0){
foreach ($s->parts as $partno=>$partarr){
    //parse parts of email
    parsepart($partarr,$partno+1);
    }
}

//for not multipart messages
else{
    //get body of message
    $text=imap_body($link,$msgid);
    //decode if quoted-printable
    if ($s->encoding==4) $text=quoted_printable_decode($text);
    //OPTIONAL PROCESSING
    if (strtoupper($s->subtype)=='PLAIN') $text=$text;
    if (strtoupper($s->subtype)=='HTML') $text=$text;
    
    $partsarray['not multipart'][text]=array('type'=>$s->subtype,'string'=>$text);
}

print_r($partsarray);
?>

masterbassist (2005-04-18 09:52:54)

I think the following line (when building attachment information)
>>> "filename" => $parts[$i]->parameters[0]->value
needs to be
>>> "filename" => $parts[$i]->dparameters[0]->value
The first version generated a PHP warning under PHP 5.0.3. The second version actually gets the filename.

spam at emielmols dot info (2005-04-01 03:47:00)

I've created a function which simply extracts basic message content as a string ($content) and attachments as an array ($attachments). I've tested it with messages from Outlook, Outlook Express and Hotmail. Perhaps it's useful to anyone.

<?php
        $struct = imap_fetchstructure($mbox, $mid);
        
        $parts = $struct->parts;
        $i = 0;

        if (!$parts) { /* Simple message, only 1 piece */
          $attachment = array(); /* No attachments */
          $content = imap_body($mbox, $mid);
        } else { /* Complicated message, multiple parts */
        
          $endwhile = false;
        
          $stack = array(); /* Stack while parsing message */
          $content = "";    /* Content of message */
          $attachment = array(); /* Attachments */
        
          while (!$endwhile) {
            if (!$parts[$i]) {
              if (count($stack) > 0) {
                $parts = $stack[count($stack)-1]["p"];
                $i     = $stack[count($stack)-1]["i"] + 1;
                array_pop($stack);
              } else {
                $endwhile = true;
              }
            }
          
            if (!$endwhile) {
              /* Create message part first (example '1.2.3') */
              $partstring = "";
              foreach ($stack as $s) {
                $partstring .= ($s["i"]+1) . ".";
              }
              $partstring .= ($i+1);
            
              if (strtoupper($parts[$i]->disposition) == "ATTACHMENT") { /* Attachment */
                $attachment[] = array("filename" => $parts[$i]->parameters[0]->value,
                                      "filedata" => imap_fetchbody($mbox, $mid, $partstring));
              } elseif (strtoupper($parts[$i]->subtype) == "PLAIN") { /* Message */
                $content .= imap_fetchbody($mbox, $mid, $partstring);
              }
            }

            if ($parts[$i]->parts) {
              $stack[] = array("p" => $parts, "i" => $i);
              $parts = $parts[$i]->parts;
              $i = 0;
            } else {
              $i++;
            }
          } /* while */
        } /* complicated message */

        echo "Analyzed message $mid, result: <br />";
        echo "Content: $content<br /><br />";
        echo "Attachments:"; print_r ($attachment);
?>

y dot daradkeh at gmail dot com (2004-12-20 07:40:16)

Hey people, get the message attachments with this code:

<?php

$mbox = imap_open("$imap_server".$f,$name,$pass);

// delibertely choose a message with an attachment

$info = imap_fetchstructure($mbox, $mno);

// find out how may parts the object has
$numparts = count($info->parts);

$i=0;
// find if multipart message
if ($numparts >1)
{
   echo "<b>More than one part</b><br><br>";

   foreach ($info->parts as $part)
   {
      if ($part->disposition == "INLINE")
         printf("Inline message has %s lines<BR>", $part->lines);
     elseif ($part->disposition == "attachment")
     {
       $i++;
       echo $i." Attachment/s found!<br>";
       echo "Filename: ".$part->dparameters[0]->value."<br><br>";
     }
   }
}
else
   echo "Only one part";

imap_close($mbox);
?>

aperez at informatica dot 24ruedas dot com (2004-06-16 12:10:15)

About above comment and source code I wrote: certainly, plain/html text files are attachments, but I've written my code thinking about web IMAP clients, where text parts are not for downloading, simply they are shown. That's the reason.
Thanks anyway. :)

passeniermaxime at hotmail dot com (2004-06-15 08:15:56)

I have written a code that check's for attachments.
There is a slight problem with the above code;
when you send a mail with a .txt document as attachment
the type will be 0 and so there won't be a attachment found.
I check for the disposition of the body part witch can be inline or attachment.
If this is an attachment he will display it.
You an also mix up both codes by checking on the type and the disposition.
//////////////////////////////////
/// $mbox = connection
/// $a_mails_sort[$i] = message id
$structuur = imap_fetchstructure($mbox,$a_mails_sort[$i]);
$bericht_delen=$structuur->parts;
for($i_delen = 0; $i_delen<=count($bericht_delen);$i_delen++)
{
if($bericht_delen[$i_delen]->disposition == "attachment")
{
$attachment = "OK";
}// end if
}//end for

aperez at informatica dot 24ruedas dot com (2004-05-25 12:12:43)

This is my function to check attachments:

<?php
  $est=imap_fetchstructure($con,$msg);

  if(checkAttachments($est))
  {
    echo("You have attachment(s)");
  }
  else
  {
    echo("You don't have attachment(s)");
  }
  
  function checkAttachments($estMsg)
  {
    if(($estMsg->type!=0) && ($estMsg->type!=1))
    {
      // Text and multipart parts will not be shown as attachments
      return(true);
    }
    else
    {
      // If there's no attachments, parts inside will be checked
      if($estMsg->parts)
      {
        $partMsg=$estMsg->parts;
        $i=0;
        // Parts will be checked while no attachments found or not all of them checked
        while(!(checkAttachments($partMsg[$i])) && ($i<sizeof($estMsg->parts)))
        {
          $i++;
        }

        // If any 'checkAttachment' calls returned 'true', 'i' should be
        // equal to number of parts(after increased in while). So, no
        // attachments found
        if($i==sizeof($estMsg->parts))
        {
          return(false);
        }
        else
        {
          return(true);
        }
      }
      else
      {
        // If no parts and text or multipart type, no attachments
        return(false);  
      }
    }
  }
  imap_close($con);
?>

I hope this helps somebody else(if bugs, thanks for your fixes).

richy at smilingsouls dot net (2004-01-28 03:08:07)

After many long hours of pouring over the object returned by imap_fetchstructure() I came up with the solution posted at the following url:
http://p2p.wrox.com/topic.asp?TOPIC_ID=9063
The url above provides an example of how to access variables in the object returned by imap_fetchstructure(). This class creates an object with member variables containing the correct part numbers which are compatible with the imap_fetchbody() function.
The structure returned looks something like this:
0 Raw Headers
$this->pid[$mid][0] = 1 text/plain top level message
$this->pid[$mid][1] = 2 message/rfc822 entire unparsed message
multipart/mixed - part is skipped!
2.0 Raw Headers
$this->pid[$mid][2] = 2.1 multipart/alternative contains unparsed contents of next subpart (text/plain && text/html)
2.1.0 Raw Headers
$this->pid[$mid][3] = 2.1.1 text/plain 2nd level inline embedded message
$this->pid[$mid][4] = 2.1.2 text/html
$this->pid[$mid][5] = 2.2 message/rfc822 entire unparsed message
multipart/mixed - part is skipped!
2.2.0 Raw Headers
$this->pid[$mid][6] = 2.2.1 multipart/alternative
$this->pid[$mid][7] = 2.2.1.1 text/plain 3rd level inline embedded message
$this->pid[$mid][8] = 2.2.1.2 text/html
$this->pid[$mid][9] = 2.2.2 image/jpeg
$this->pid[$mid][10] = 2.2.3 image/jpeg
$this->pid[$mid][11] = 2.2.4 image/jpeg
$this->pid[$mid][12] = 2.2.5 image/gif
Its important to note why multipart/related, multipart/mixed and certain multipart/alternative (if parent part is message/rfc822 && not multipart/mixed) parts are skipped, imap_fetchbody **does not** retreive a message part for these parts. Which isn't a bug, as I have read in other places, these parts exist for the purpose of designing the viewer portion of a mail application. Also, headers aren't assigned part numbers here because they are easy to pick out.
: )
Rich

chrislhill at "O_o" hotmail dot com (2003-10-23 11:36:49)

For people just beging this may help alot as I spent a couple hours just trying to relize how exact the array was stored. (at the time I did not know the print_r function :P)
$struct = imap_fetchstructure($mbox, $msgnumber);
print_r($struct);
Will give you a better example for your messages but they are called as $struct using the varible method above.
$struct->type; //would return the type
$struct->encoding //would return the encoding
and etc..
This can be done many different ways but that is the basics on pulling the info out of the structure of fetchstructure itself, I hope it helps someone starting because it wouldve helped me :/.

es96 at hotmail dot com (2003-08-13 02:38:58)

If you are getting CHARSET as US-ASCII even if the header has a Content-Type: field, make sure the header also has a MIME-Version: field.
For example, the following header will correcty report charset as KOI8-R
MIME-Version: 1.0
Content-Type: text/plain; charset="koi8-r"
Without the MIME-Version it will be reported as US-ASII

as1 at powersite dot com dot ar (2002-12-30 13:57:43)

Sorry, my before code is not complet:
$struct = imap_fetchstructure( $conn, $curMsg );
$i = 1;
while( imap_fetchbody( $conn, $curMsg, $i ) ){
$struct->parts[$i-1] = imap_bodystruct( $conn, $curMsg, $i );
$j = 1;
while( imap_fetchbody( $conn, $curMsg, "$i.$j" ) ){
$struct->parts[$i-1]->parts[$j-1] = imap_bodystruct( $conn, $curMsg, "$i.$j" );
$j++;
}
$i++;
}

as1 at powersite dot com dot ar (2002-12-30 13:01:36)

If your imap_fetchstructure not retrive all parts, write it:
$struct = imap_fetchstructure( $conn, $curMsg );
$i = 1;
while( imap_fetchbody( $conn, $curMsg, $i ) ){
$struct->parts[$i-1] = imap_bodystruct( $conn, $curMsg, $i );
$i++;
}

paradise at ml75 dot com (2002-12-04 07:40:47)

I'm using freeBSD and had lots of problems with all functions/mail applications i could find on the net...
So i'm writing my own now..
Most of the stuff works fine now.. but i'm still working on the chooser to select the best part to show. (plain or HTML).
The #1 problem is that most functions give the wrong ID numbers beck to fetch the body.
Here's a sollution to that problem.. :P
// Debug function.. always handy until it's 100% completed..
$attachments = array(); $dispos=0;
$debug=true;
function doDebug($msg){
global $debug, $debug_depth, $dispos;
if($debug){
for($y = 0; $y < $debug_depth; $y++) echo("&nbsp;&nbsp;&nbsp;&nbsp;");
if($dispos == 0) echo($msg."<br/>"); else echo("<i>".$msg."</i><br/>");
}
}
function Check_Part($part,$id){
global $attachments, $debug, $debug_depth, $dispos; $ans=sizeof($attachments);
doDebug("+Checking part $id");
$debug_depth++;
doDebug("Parttype=".$part->type ."/". mimetype($part->type) ."/". $part->subtype);
if($part->ifdisposition){ $dispos++; doDebug("<b>Disposition</b>"); }
if($part->ifdescription){ doDebug("<b>Description=".$part->description."</b>"); }
switch($part->type){
case 1: if((strtolower($part->subtype) == "mixed")
or (strtolower($part->subtype) == "alternative")
or (strtolower($part->subtype) == "related")) break;
default: $an = sizeof($attachments);
if($part->ifdparameters){
$dpara = $part->dparameters;
for ($v=0;$v<sizeof($dpara);$v++){
if (eregi("filename", $dpara[$v]->attribute))
$fname = $dpara[$v]->value;}
}
if($part->ifparameters){
if(empty($fname)){
$para = $part->parameters;
for ($v=0;$v<sizeof($para);$v++){
if (eregi("name", $para[$v]->attribute)) $fname = $para[$v]->value;
}
}
}
if(empty($fname))$fname = "Onbekend";
$attachments[$an]->id = ($an+1);
$attachments[$an]->part = $id;
$attachments[$an]->filename = $fname;
$attachments[$an]->type=$part->type;
$attachments[$an]->subtype=$part->subtype;
$attachments[$an]->dispos=$dispos;
$attachments[$an]->size=$part->bytes;
doDebug("Filename=".$attachments[$an]->filename); break;
}
for($x = 0; $x < count($part->parts); $x++){
Check_Part($part->parts[$x], $id.".".($x+1));
}
doDebug("Dit deel had iets nuttigs.");
if($part->ifdisposition) $dispos--;
$debug_depth--;
doDebug("-End of part $id");
}
function decode_mail($mbox, $msg){
global $attachments, $debug, $debug_depth;
doDebug("+Decoding message $msg"); $debug_depth++;
$obj = imap_fetchstructure($mbox, $msg, FT_UID); $moet_werken=true; $id=1;
while($moet_werken){
Check_Part($obj, $id);
$moet_werken=false;
}
$debug_depth--; doDebug("-End of message $msg");
}
Now to get this to work..
Decode_Mail([Mailbox connection], [Message Number]);
this is the outpu attachment array..
1--1.1.1--Onbekend--0--PLAIN--0--31
2--1.1.2--Onbekend--0--HTML--0--29
3--1.2--Beschrijving Nieuw E.doc--3--MSWORD--1--613718
--------------------------------------
PS.. "Onbekend" is dutch for "Unknown".
Good luck.
Paradise (www.ml75.nl)

theUNDERSCOREVoodooChileADGMXdotNET (2002-10-25 05:21:04)

you can easily see through everything with these functions:
function printarray($array){
while(list($key,$value) = each($array)){
if(is_array($value)){
echo $key."(array):<blockquote>";
printarray($value);//recursief!!
echo "</blockquote>";
}elseif(is_object($value)){
echo $key."(object):<blockquote>";
printobject($value);
echo "</blockquote>";
}else{
echo $key."==>".$value."<br />";
}
}
}
function printobject($obj){
$array = get_object_vars($obj);
printarray($array);
}
end then do this:
$struct = imap_fetchstructure($box,$mailnr);
//$struct is een object!!
drukobjectaf($struct);
you get an extended tree-view into the hierarchie of objects and arrays
Comparing different mails(changing thenumber) may bring more clarity into this matter
(these functions may not work properly as they are translated from my dutch original, but they ought to be OK)

jcastro at elnuevodia dot com (2002-09-20 13:16:28)

I think the above note about attachments is wrong. I tested sending files with and without attachments and I get the following<br>
with attachment: type=3 bytes=343648
no attachment: type=0 bytes=2
so checking for $struct->bytes == " " means nothing. At least in my test
running windows 2000, php4.2.1 using outlook and exchange. It seems that cheking the type will be more reliable

(2002-09-19 15:20:06)

This function sets MIME charset
parameter value to US-ASCII when
either Content-Type header is missing
or Content-Type header doesn't have
charset parameter. This may be compliant
to RFC 822, but this poses
a problem for a calling function that
has to deal with untagged messages
in MIME charset other than US-ASCII.
To work around this, US-ASCII
has to be treated as the absence of MIME charset.

kmakaveev(at)yahoo(dot)com (2002-06-25 07:29:21)

There's a little mistake in the code above!
The second echo must return "The message has no attachment!".
And that code must have some some additional thing to work properly! Maybe I must use ->parts to determine how many parts has a particular message!

kmakaveev(at)yahoo(dot)com (2002-06-24 11:54:57)

There's a simple way to check weither a particular mail message has an attachment or not. I'm not sure that it is 100% true but here it is a sample code:
<?
$msg_structure = imap_fetchstructure($mail_inbox, $msg_id);
if($msg_structure->bytes == "")
{
echo "The message has attachment!";
}
else
{
echo "The message has attachment!";
}
?>
Please, test this code and let me know is it working properly or not! Thanks in advance!

elcarlos at skaainet dot be (2002-06-22 15:11:36)

Somebody must include some good examples of how to decode a message with and without an attachement.
I'm writing a web based email, and it wasn't verry easy for me to understand that part.
After a week of searching I still don't get it very well, but now (I think and hope) I'm able to write it...
Yet I'm not that good with it, so I'm not able to put an example in it myself...
Regards,
El_Carlos

trionon at mail dot ru (2002-02-22 08:42:24)

It's better to use special constants instead of digits for primary content-type:
TYPETEXT unformatted text
TYPEMULTIPART multiple part
TYPEMESSAGE encapsulated message
TYPEAPPLICATION application data
TYPEAUDIO audio
TYPEIMAGE static image (GIF, JPEG, etc.)
TYPEVIDEO video
TYPEOTHER unknown
These constants are documented in the source code of IMAP extension but they are still undocumented here :(

php AT firstnetimpressions.com (2002-02-18 13:15:02)

Point of clarification:
The seventh primary body type is not "Other" as documented, but actually "Model". This encompasses IGES, VRML, MESH, DWF, etc.
http://www.isi.edu/in-notes/iana/assignments/media-types/media-types
"Other" is the eigth primary body type.

mkknapp at quadrapod dot com (2001-04-25 18:48:45)

Assuming $struct = imap_fetchstructure($x,$y);
It is important to note that if a message has NO attachements, $struct->parts is an empty array, and $struct->bytes has a value. If a message as ANY attachements, $struct->bytes ALWAYS = 0. To get the size of the primary body, you have to call structure->part[0]->bytes. To get the size of the whole message, either strlen(imap_body) or add up the ->bytes of all the parts.
Another interesting note:
When there is body text and no attachements:
count($struct->parts) = 0
When there is body text and 1 attachement:
count($struct->parts) = 2
These imap functions could really use better documentation. Like giving the meathods for the dparameter and parameter classes...

jbrandt at europa dot com (2001-04-16 15:45:04)

Parsing the recursive object returned from imap_fetchstructure is easy. However trying to figure out the name of the section to pass to imap_fetchbody() isn't quite so straight forward because the naming conventions for the MIME sections are odd. If you copy and paste the following function you can pass an UID of an IMAP message and it will print out a hierarchy of the message and label each section.

The only required parameters are $mbox and $em_uid. The other parameters are used when the function needs to call itself.

Two functions mime_encoding() and mime_type() are not included. All they do is return string for the integer represention of a mime type/encoding.

<?php
function imap_dbody($mbox, $em_uid, $mimeobj, $depth = 0, $section = 0)
{
        // If this is the first run of the function then grab a object of the MIME structure of the message

        if($section == 0) $mimeobj = imap_fetchstructure($mbox, $em_uid, FT_UID);

        for($y = 0; $y < $depth; $y++) echo("???????");

        echo(mime_type($mimeobj->type) . "/{$mimeobj->subtype}, ");
        echo(mime_encoding($mimeobj->encoding) . " (<B>$section</B>)<BR>");

        for($x = 0; $x < count($mimeobj->parts); $x++)
        {
                // If we are in the root of the object increment by whole integers

                if($section == 0) $nsection = $x + 1;

                // If we are in the object and the first sub-object of our object isn't multipart
                // then increment the postfix by ".1" otherwise we are multipart or a message
                // and leave the section id alone to be handled by the next code block

                else if(($pos = strrpos($section, ".")) && $mimeobj->parts[0]->type != TYPEMULTIPART)
                        $nsection = substr($section, 0, $pos) . "." . ($x + 1);
                else
                        $nsection = $section;

                // If there are more parts to the part about to be processed reference it as a header with ".0"
                // but only if the child of this child isn't MULTIPART

                if(count($mimeobj->parts[$x]->parts))
                {
                        // Funny really, if a mime section is a inline message that has a multipart body you reference the message
                        // mime section with "2" the inline message header with "2.0" and the subsections with "2.x"
                        // However if the mime section is a inline message with only 1 part then you reference the
                        // mime section in the message with 2.0 and the inline message body with 2.1

                        if(!($mimeobj->parts[$x]->type == TYPEMESSAGE && $mimeobj->parts[$x]->parts[0]->type == TYPEMULTIPART))
                                $nsection .= ".0";
                        else
                                $nsection .= "";
                }

                imap_dbody($mbox, $em_uid, $mimeobj->parts[$x], $depth + 1, $nsection);

        }

        // If after processing the entire MIME object the $x variable is still zero then we didn't
        // process a multipart mime message, it's just normal email so say so here.

        if($x == 0 && $section == 0)
        {
                echo(mime_type($mimeobj->type) . "/{$mimeobj->subtype}, ");
                echo(mime_encoding($mimeobj->encoding) . " (<B>1</B>) (<B>NOT MIME MULTIPART</B>)<BR>");
        }
}
?>

vksgeneric at hotmail dot com (2000-02-02 17:22:45)

It looks like some clients (e.g. outlook) can attach a file that has two mime-types associated with it, like:
Content-Type: model/vrml,x-world/x-vrml
This bombs the parsing procedure somewhere in c-client, I guess, and as a result ifparameters is set to false for the respective part, and you can not retrieve filename, mimetype, or any other parameters.

hholzgra at media-engineering dot de (1999-12-15 13:59:00)

the parts objects are identical in
structure to those returned by imap_fetchstructure, describing one subfolder each
parameters and dparameters are MIME-specific, theese contain the
extra parameters provided in the Content-Type and Content-Disposion Header lines such as Charset or Filename