Trying to automatically detect and parse encoding or content type from e-mails

[imekul]

I have piping set up, where when I send an e-mail to a certain address, a script executes.

 

This script will parse the e-mail and will post the body of the e-mails (along with a picture, if attached) to a blog.

 

This all works okay, except the problem I keep running into is when I send from different devices (desktop, Android phone, BlackBerry), the e-mails have different character sets or encoding.  So one e-mail will show up just fine, and then the body of another will look something like this:

 

--90e6ba6e83a0846c8504926b9478

Content-Type: text/html; charset=ISO-8859-1

 

This is the content of the e-mail.

 

--90e6ba6e83a0846c8504926b9478--

 

So what I'm wondering is if there is any way to detect the charset/encoding and automatically decode it properly?  The only solution I have so far -- and I'm not very good with PHP, I'll admit -- is to just "hack" my way through, by using stristr() to see if the body contains certain characters (like "ISO-8859-1"), and then extracting the text that is in between, say, "ISO-8859-1" and "--".

 

Of course, this is very crude, and has many problems on its own.  So what I am looking for is if there is some sort of way to automatically get the body of an e-mail in plain text, regardless of whether the e-mail has an attachment, how it's encoded, or what mail client was used.

 

Thanks for your time!

[AbraCadaver]

This has already been done.  Look at phpclasses.org or PEAR for an email parsing or mime class.  There are bound to be many other good ones.

 

http://www.phpclasses.org/search.html?words=parse+email&x=0&y=0&go_search=1

http://pear.php.net/package/Mail_mimeDecode

[chintansshah]

Please write some code here.

[imekul]

Thanks, Shawn.  I'll take a look at that.

[imekul]

Are you aware of any samples or examples that show how to do something like this?

 

I found a sample script at http://www.nicholassolutions.com/tutorials/php/index.php#MimeDecode, but it doesn't seem to work fully.