Decoding the subject with iconv_mime_decode()

[freescout-helpdesk]

Describe the bug

Currently Webklex/php-imap by default uses 'utf-8' decoder and \imap_utf8() function to decode subject:

php-imap/src/Header.php

Line 424 in abf080e

if ($decoder === 'utf-8' && extension_loaded('imap')) {

This approach does not always decode subjects properly.

For example Subject: =?ISO-2022-JP?B?GyRCIXlCaBsoQjEzMhskQjlmISEhViUsITwlRyVzGyhCJhskQiUoJS8lOSVGJWolIiFXQGxMZ0U5JE4kPyRhJE4jURsoQiYbJEIjQSU1JW0lcyEhIVo3bjQpJSglLyU5JUYlaiUiISYlbyE8JS8hWxsoQg==?= will be decoded into garbled string:

�$B!yBh�(B132�$B9f!!!V%,!<%G%s�(B&�$B%(%/%9%F%j%"!W@lLgE9$N$?$a$N#Q�(B&�$B#A%5%m%s!!!Z7n4)%(%/%9%F%j%"!&%o!<%/![�(B

So it does not make much sense to use by default the approach which is not 100% reliable.

Some time ago we've switched to decoding the subject simply with iconv_mime_decode():

$value = iconv_mime_decode($value, ICONV_MIME_DECODE_CONTINUE_ON_ERROR, "UTF-8");

It works always and properly - freescout-help-desk/freescout#2965 (comment).

So maybe it worth making iconv_mime_decode() the default and the only approach for decoding the subject like this: https://github.com/freescout-helpdesk/freescout/blob/master/overrides/webklex/php-imap/src/Header.php#L208

Function:
https://github.com/freescout-helpdesk/freescout/blob/dist/app/Misc/Mail.php#L909 (decodeSubject())

Related issue: Webklex/laravel-imap#295

[Webklex]

Hi @freescout-helpdesk ,
many thanks for your report and suggestion. I improved the decoder and added a new test case for the mentioned subject above.

Best regards & happy coding,

[nuernbergerA]

@Webklex Attachment::decodeName() has the same problem

[Webklex]

Hi @nuernbergerA ,
thanks for your report. I just pushed a new test including your mentioned scenario but can't reproduce the described issue.
Are you using the latest release? If not, please update and try again. There's a chance that the problem might have already been fixed in a newer release.

If the issue persists, please head over to https://github.com/Webklex/php-imap/issues/new?assignees=&labels=&projects=&template=bug_report.md and provide as much information as possible. Every little bit, even if it might seem unrelated could help to solve the mystery :)

Best regards & happy coding,

[nuernbergerA]

@Webklex i've opened a PR with a failing test for my issue ;)

[freescout-helpdesk]

@Webklex i've opened a PR with a failing test for my issue ;)

We've checked. This our function https://github.com/freescout-helpdesk/freescout/blob/dist/app/Misc/Mail.php#L883C8-L883C8 also successfully works for decoding headers and attachment names including your example:

=?iso-8859-1?Q?2021=5FM=E4ngelliste=5F0819306.xlsx?= >> 2021_Mängelliste_0819306.xlsx