imap_thread() returns threads, but are confined to the current open mailbox you defined in imap_open(). This is not useful for, lets say, getting full threads ( from "Sent Messages" and "Inbox" [took me a day to figure this out]).
If you compare threads on Outlook vs gmail.com you will find that Outlook determines threads by subject title, not actual parent > child relationships.
Gmail however, seems to get threads right, but does not include mail you send using their web interface in {imap.google.com:993/imap/ssl}Sent Messages . What this means is that threads using php imap won't be perfect for gmail.
If you send mail using Outlook (or any mail client), gmail.com does put it in their "Sent Mail".
So all in all, threads for PHP imap are not perfect. But I blame the imap specifications (DEAR IMAP GUYS, please add better uids and parent ids. thx Chris) more than PHP
So I created the Outlook method for threading (comparing subjects) below:
<?php
$imap = imap_open('{imap.gmail.com:993/imap/ssl}INBOX', '[email protected]', 'yourpassword');
$subject = 'Item b';
$threads = array();
//remove re: and fwd:
$subject = trim(preg_replace("/Re\:|re\:|RE\:|Fwd\:|fwd\:|FWD\:/i", '', $subject));
//search for subject in current mailbox
$results = imap_search($imap, 'SUBJECT "'.$subject.'"', SE_UID);
//because results can be false
if(is_array($results)) {
//now get all the emails details that were found
$emails = imap_fetch_overview($imap, implode(',', $results), FT_UID);
//foreach email
foreach ($emails as $email) {
//add to threads
//we date date as the key because later we will sort it
$threads[strtotime($email->date)] = $email;
}
}
//now reopen sent messages
imap_reopen($imap, '{imap.gmail.com:993/imap/ssl}Sent Messages');
//and do the same thing
//search for subject in current mailbox
$results = imap_search($imap, 'SUBJECT "'.$subject.'"', SE_UID);
//because results can be false
if(is_array($results)) {
//now get all the emails details that were found
$emails = imap_fetch_overview($imap, implode(',', $results), FT_UID);
//foreach email
foreach ($emails as $email) {
//add to threads
//we date date as the key because later we will sort it
$threads[strtotime($email->date)] = $email;
}
}
//sort keys so we get threads in chronological order
ksort($threads);
echo '<pre>'.print_r($threads, true).'</pre>';
exit;
?>
so if you are going to use imap_thread() for something useful. This is probably the most optimal way I can think of:
<?php
$imap = imap_open('{imap.gmail.com:993/imap/ssl}INBOX', '[email protected]', 'password');
$threads = $rootValues = array();
$thread = imap_thread($imap);
$root = 0;
//first we find the root (or parent) value for each email in the thread
//we ignore emails that have no root value except those that are infact
//the root of a thread
//we want to gather the message IDs in a way where we can get the details of
//all emails on one call rather than individual calls ( for performance )
//foreach thread
foreach ($thread as $i => $messageId) {
//get sequence and type
list($sequence, $type) = explode('.', $i);
//if type is not num or messageId is 0 or (start of a new thread and no next) or is already set
if($type != 'num' || $messageId == 0
|| ($root == 0 && $thread[$sequence.'.next'] == 0)
|| isset($rootValues[$messageId])) {
//ignore it
continue;
}
//if this is the start of a new thread
if($root == 0) {
//set root
$root = $messageId;
}
//at this point this will be part of a thread
//let's remember the root for this email
$rootValues[$messageId] = $root;
//if there is no next
if($thread[$sequence.'.next'] == 0) {
//reset root
$root = 0;
}
}
//now get all the emails details in rootValues in one call
//because one call for 1000 rows to a server is better
//than calling the server 1000 times
$emails = imap_fetch_overview($imap, implode(',', array_keys($rootValues)));
//foreach email
foreach ($emails as $email) {
//get root
$root = $rootValues[$email->msgno];
//add to threads
$threads[$root][] = $email;
}
//there is no need to sort, the threads will automagically in chronological order
echo '<pre>'.print_r($threads, true).'</pre>';
imap_close($imap);
exit;
?>
PHP.mk документација
imap_thread
Почист и полокален преглед на PHP референцата, со задржана структура од PHP.net и подобра читливост за примери, секции и белешки.
Патека
function.imap-thread.php
Локална патека за оваа страница.
Извор
php.net/manual/en
Оригиналниот HTML се реупотребува и локално се стилизира.
Режим
Прокси + превод во позадина
Кодовите, табелите и белешките остануваат читливи во истиот тек.
Референца
function.imap-thread.php
imap_thread
Референца за `function.imap-thread.php` со подобрена типографија и навигација.
imap_thread
(PHP 4 >= 4.0.7, PHP 5, PHP 7, PHP 8)
imap_thread — Враќа дрво од нишки пораки
= NULL
Добива дрво од нишки пораки.
Вратени вредности
imap_thread() враќа асоцијативен низ кој содржи дрво од пораки нишки по REFERENCES, или false
при грешка.
Секоја порака во тековната поштенска кутија ќе биде претставена со три записи во резултирачкиот низ:
$thread["XX.num"] - тековен број на порака
$thread["XX.next"]
$thread["XX.branch"]
Дневник на промени
| Верзија | = NULL |
|---|---|
| 8.1.0 |
На imap параметарот очекува IMAP\Connection
инстанца сега; претходно, валидна imap resource се очекуваше.
|
Примери
Пример #1 imap_thread() Пример
<?php
// Here we're outputting the threads of a newsgroup, in HTML
$nntp = imap_open('{news.example.com:119/nntp}some.newsgroup', '', '');
$threads = imap_thread($nntp);
foreach ($threads as $key => $val) {
$tree = explode('.', $key);
if ($tree[1] == 'num') {
$header = imap_headerinfo($nntp, $val);
echo "<ul>\n\t<li>" . $header->fromaddress . "\n";
} elseif ($tree[1] == 'branch') {
echo "\t</li>\n</ul>\n";
}
}
imap_close($nntp);
?>Белешки од корисници 2 забелешки
cblanquera at gmail dot com ¶
пред 14 години
whamill at google mail ¶
пред 15 години
I figure other people may benefit from an explanation of the resulting array, My understanding is:
Key: Essentially a node ID
Num: The mail ID (where 0 indicates the start of a conversation)
Next: The node ID of the first child (where 0 indicates there are no children)
Branch: The node ID of the next sibling (where 0 indicates there is no sibling)