The issue of the '?' characters has now also been solved. Not sure if
it's just our systems or everyone out there that has this problem but
we made the following change in imp/mailbox.php
changed:
$msg['subject'] = $imp_ui->getSubject($ob->subject);
to:
$msg['subject'] = utf8_encode($imp_ui->getSubject($ob->subject));
Doing this, also no longer requires the modification to
NLS::getCharset() I mentioned earlier since the subject effectively is
UTF8 now and the function works properly.
Cheers
I guess it could be done in one line as well
Quoting [email protected]:
For issue #2, I found the following by printing the Ascii codes for
each letter of the troubled words in the subject.
Taking for example the word: Académico
When it's being displayed properly, it's made up of the following codes:
65 - A
99 - c
97 - a
100 - d
195 - ?
169 - ?
109 - m
105 - i
99 - c
111 - o
When it's showing incorrectly, it's made up of:
65 - A
99 - c
97 - a
100 - d
233 - ?
109 - m
105 - i
99 - c
111 - o
I got these values by running the following code:
for ($x=0;$x<strlen($msg['subject']);$x++) {
print ord(substr($msg['subject'],$x,1))." -
".substr($msg['subject'],$x,1)."<br>";
}
In one example the é is made from two codes ( 195/169 ) and displays
properly while the second time around it's made from only one code (
233 ) and shows the question mark.
What gives ?
Quoting [email protected]:
I'm still unsure as to what issue #1 is and it still remains but I
did make headway with issue #2.
I tracked down the problem to Horde/Text.php in the function
htmlSpaces(). More precisely, this line:
$charset = NLS::getCharset();
NLS::getCharset always returned UTF-8 even though my subject was
encoded in ISO-8859-1. Further down, this line was effectively
emptying out my subject line whenever $charset was set to UTF-8:
$text = @htmlspecialchars($text, ENT_COMPAT, $charset);
For now, I changed $charset = NLS::getCharset(); with $charset =
NLS::getCharset(TRUE); which works because now the charset is
returned as ISO-8859-1
Is this a limitation of the htmlspecialchars() function ?
However, I still have issue #1 present where it shows '?' instead
of the actual character.
Quoting [email protected]:
I have two problems, not sure if they are a bug on my server or
maybe a bug within IMP.
I am using the latest Horde Webmail distribution:
Horde: 3.3.8
IMP: 4.3.7
c-client library: 2007e
libiconv: 1.13.1
Dovecot IMAP: 1.2.10
Issue #1.
IMP can't display characters from a different charset in the
subject and instead displays a ? instead of the character. In the
body of the message the subject displays just fine.
I enabled the rawlog option in Dovecot to see the actual IMAP
conversation between IMP and IMAP. When it just gets the message
index, for that particular message I see:
* 4 FETCH (INTERNALDATE "11-May-2010 20:22:31 -0400" RFC822.SIZE
2357 FLAGS (\Seen) ENVELOPE ("Tue, 11 May 2010 17:53:34 -0600" {86}
IT-F01 Solicitud de acción de Infraestructura Tecnológica. Otro
según detalle adjunto.
Which is perfectly fine. However, in Webmail I see the subject as being:
IT-F01 Solicitud de acci?n de Infraestructura Tecnol?gica. Otro
seg?n detalle adjunto.
Issue #2.
IMP shows no subject line at all for messages that contain
subjects defined encoded in ISO-8859-1. On the server the subject
starts like this:
=?iso-8859-1?Q?FW:_
What could this be ?
Thank you!
--
IMP mailing list - Join the hunt: http://horde.org/bounties/#imp
Frequently Asked Questions: http://horde.org/faq/
To unsubscribe, mail: [email protected]
--
IMP mailing list - Join the hunt: http://horde.org/bounties/#imp
Frequently Asked Questions: http://horde.org/faq/
To unsubscribe, mail: [email protected]
--
IMP mailing list - Join the hunt: http://horde.org/bounties/#imp
Frequently Asked Questions: http://horde.org/faq/
To unsubscribe, mail: [email protected]
--
IMP mailing list - Join the hunt: http://horde.org/bounties/#imp
Frequently Asked Questions: http://horde.org/faq/
To unsubscribe, mail: [email protected]
