Plain text version ?eats out“ ?

@dainius,

Having done a little reading about this it seems to be because š is not available in a plain text character set so it cannot be replaced.

Lithuanian (and some others) alphabet has such ?abnormal“ characters: ???????ū?. Every character present in utf8. And only ? letter causes ?problems“ in some WYSIWYG editors – it is converted to HTML entity.
I fixed this problem at you plugin code by adding line at class-s2-core.php:252:
$message = str_replace( array('š','Š'), array('?','?'), $message);
But still not tested.
I know, I will loose this on next plugin update. So it would be kind if you add this fix in your future update (maybe another – more ?correct“ place in code).
As I mentioned, this problem appears to many users, because this character is ?popular“ not only in one language.

Also you delete a lot of potential characters by
$message = preg_replace('|&[^a][^m][^p].{0,3};|', '', $message);
Why not to decode them back using html_entity_decode() ?
This could be safe, because here we talk about converting from HTML to plain-text mail.

I think instead of doing:

$message = preg_replace('|&[^a][^m][^p].{0,3};|', '', $message);
$message = preg_replace('|&|', '&', $message);

it would be better to do:

$message = html_entity_decode( $message, ENT_NOQUOTES, 'UTF-8' );

BTW this also will fix Š problems ??

@dainius,

Those preg_replace() calls are pretty old code now and I think your approach is simpler and better. I’ll get your change rolled into the next version.