Email not displaying/being parsed properly #6713
Comments
MrBMT
changed the title
|
I have a similar problem, the first line of the message is missing. I'm also using the current git master (4d1b3e2). The message is something like: If we decode the base64, the message is something like: I found that |
|
The @MrBMT's sample contains |
|
It's a $html = "<html>Hello, This is a test.<br />Does it work this time?</html>";
$html = "Hello, This is a test.<br />Does it work this time?";it still thinks it is in // This is sort of a last-ditch attempt to correct for cases where no head/body
// elements are provided.
if ($this->insertMode <= static::IM_BEFORE_HEAD && 'head' !== $name && 'html' !== $name) {
$this->insertMode = static::IM_IN_BODY;
}so, |
|
I agree that we need a better fix. I created a ticket for Mastermind Masterminds/html5-php#166. |
|
Implemented workaround. Fixed. |
|
@alecpl I updated to the latest git master and it still wasn't being processed correctly. I ran what was being passed through to the fix_html5() function through htmlspecialchars() then var_dumped the result, which turned out to be: So for some reason, roundcube is adding In addition, the In order to fix things, I had to amend the function as follows: /**
* Cleanup and workarounds on input to Masterminds/HTML5
*/
protected function fix_html5($html)
{
$html = str_replace('<head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8" /></head>', "", $html);
// HTML5 requires <head> or <body> (#6713)
// https://github.com/Masterminds/html5-php/issues/166
if (!preg_match('/<(head|body)/i', $html)) {
$pos = stripos($html, '<html');
if ($pos === false) {
$html = '<html><body>' . $html;
}
else {
$pos = strpos($html, '>', $pos);
$html = substr_replace($html, '<body>', $pos+1, 0);
}
}
return $html;
}I haven't had a chance this evening to look in to where the Thanks for looking in to this! |
|
I forgot we add that meta tag in another place. Should be fixed now. |
|
@alecpl unfortunately, this fix doesn't solve my issue. My problematic message skips the added Não, não, era mesmo só para lhe dizer que foi um erro meu. O problema do overfitting mantém-se.
<div class="quote" style="line-height: 1.5"><br>
<br>-------- Mensagem original --------
<br>Assunto: Re: Re.: Gráficos
<br>De: (...)
<br>Para: (...)<br><br>
<blockquote (...)>
<html>
<head><meta charset="UTF-8" /></head>
<body dir="auto">Mas acha que deveríamos repensar e (...)For my case, the following PHPUnit test case must work: $html = 'First line<br /><html><body>Second line';
$washed = $washer->wash($html);
$this->assertContains('First line', $washed);While it's not fixed, i must patch the diff --git a/program/lib/Roundcube/rcube_washtml.php b/program/lib/Roundcube/rcube_washtml.php
index 8e4edde..459d9cf 100644
--- a/program/lib/Roundcube/rcube_washtml.php
+++ b/program/lib/Roundcube/rcube_washtml.php
@@ -783,6 +783,14 @@ class rcube_washtml
*/
protected function fix_html5($html)
{
+ // work only with the first part of the message, before any HTML tag
+ if ($pos = strpos($html, '<')) {
+ $html_after = substr($html, $pos);
+ $html = substr($html, 0, $pos);
+ } else {
+ $html_after = '';
+ }
+
// HTML5 requires <head> or <body> (#6713)
// https://github.com/Masterminds/html5-php/issues/166
if (!preg_match('/<(head|body)/i', $html)) {
@@ -797,6 +805,8 @@ class rcube_washtml
}
}
+ $html .= $html_after;
+
return $html;
} |
|
Problem with this sample is that it is not really a valid HTML. DOMDocument-based code didn't indeed skip that text, so maybe we should support that case too. What a mess. |
|
Fixed. |
Note: This is with the current git master (4d1b3e2) being used.
I sent an email to myself today from another system, however the text displayed via Roundcube is incorrect - the first line is completely missing.
The message sent was:
However it displays via Roundcube:
As shown in this screenshot:
Viewing the HTML being produced, it is completely ignoring the first line of the email:
<div id="messagebody"><div class="message-htmlpart" id="message-htmlpart1"><div class="rcmBody"><!-- html ignored --><!-- head ignored --><!-- meta ignored --><!-- html ignored --><br />Does it work this time?</div></div>For your reference, an anonymised version of the raw message source is below:
The text was updated successfully, but these errors were encountered: