Top-posting patch was incorporated into SVN r3087 and later. Just tested r3107 and noticed that while the patch works fine, signatures inserted at the top still use the '-- ' separator, which will cause problems for top-posters. See this post - http://www.roundcubeforum.net/api-based-plugins/5148-plugin-sig-above-reply-option.html#post21439
Ideally, the separator should be left alone for bottom-posters, but different or absent for top-posters.
Just checked out 3148, and this issue has still not been addressed.
I see only two choices - either the '-- ' separator needs to be changed/deleted for all replies/forwards, or better would be if the separator was left alone for bottom-posting but changed/deleted when top-posting is selected.
This definitely needs to be addressed prior to the next update to stable. Top-posting users will complain loudly if the quoted messages that they forward or reply to are stripped below their reply. This will negate all the good work on making top-posting possible.
I thought there was a ticket for this but I can't find it now, If there isn't a ticket for it you should start one.
I couldn't find a ticket for it either, so I started one, #1486330.
I've attached a diff file to the ticket on this:
#1486330 (Need signature separator change for top-posting replies) ? Roundcube Webmail (http://trac.roundcube.net/ticket/1486330)
The diff was created from r3177, and can be applied as a patch to that version. (patch -bp0 sig-separator.diff)
The goal was to retain the '-- ' separator if the signature is placed at the bottom (whether the user top-posts or not), but not add it if the prefs are set to signature-above-replies.
Comments and/or criticism are welcome.
I've attached an alternate patch to the ticket linked in my previous reply. Based on feedback from my internal testers, many people like having a separator of some kind. So the alternate diff/patch (sig-separator2.diff) uses "-- " for bottom posting, but "- " for top-posting. This diff/patch should be valid for svn 3177-3184 (the current svn as I write this).
Has anyone other than myself tested the patches I attached to the ticket on this yet? If so, any feedback?
Fixed in SVN 3315. Works well, and better thought out than I had hacked together. Good work!