mail2web strips out too much text when trying to strip email client signatures/quote lines

Hi Jose,

Can you tell me what mail application you replied with (gmail, yahoo, outlook, thunderbird, etc.)? Can you confirm your mailhandler setup is working? Finally, are you running cron and/or retrieving the messages manually on the admin/content/mailhandler (retrieve) page (if not running cron and/or mailhandler is not configured to retrieve messages on cron for that mailbox)?

thanks
Ian

Jose, can you check a few more things?

1) What do you see when you click "edit" on the comment in http://drupal.org/files/issues/empty%20reply.jpg that has no body?
2) What is your input format set at on the mailhandler mailbox? (go to admin/content/mailhandler/ and click on edit for that mailbox, and look at the last setting at the bottom - there is a "Input Format" radio selection)
3) What is your "Message body filter:" set at on /admin/messaging/settings ?

For the input format for mailhandler for that mailbox (not for messaging, you have that set correct) can you try switching to another input format?

Are you using Gmail to reply to the messages? This looks like a bug w/ overzealous/bad an preg_match when trying to strip out the remains of the signature/quote line that email clients stick in. I was able to get this to repeat while replying to messages via Gmail. Assigning to myself to fix. I've started to look how to rework this part of the code to a simpler solution, but it will be a few days or mid next week before it's rolled out into the dev version.

This is fixed in the latest commit to the dev version, Drupal 5 only. I'll do Drupal 6 soon. I've implemented an input format that will replace anything that mail2web thinks should be stripped with a collapsed fieldset at the end of the comment - much like "show quoted text" in gmail's client. This ensures no accidental data loss from over-aggressive or inaccurate regexing for all the corner cases and different mail cilents.

Please test on a dev site if you're able. I'm aiming for new betas w/in the next week or two.

Ian

There is now a new dev version to test out which should fix this issue. An input format is now used to simply hide the text instead of deleting it. Not all email signatures/junk will get stripped out because the regex is still focused on mainly two common signature patterns. Anyone who is able to test the dev version it would be great to get more feedback.

Thanks

I will keep this active until resolution is confirmed.

Ian,
the problem i have with the filter approach is one of workflow(?). what happens if the person goes to the site and edits the post and the same filter is not available to them? in my testing of D6 you get an error that complains about no input filter being set. this is just thinking out loud mostly. i could just be thinking about it too hard.

i ended up putting the regex stuff back in my d6 version to strip the text. but i am also not supporting html email on my sites, so that is not a good solution either.

thought about trying flexifilter or customfilter so there is no need to support this code in mail2web, but that adds complexity to admin.

r,
c

I am going to close this issue, please see my comment here http://drupal.org/node/349730#comment-1223209 and look for a release for both 5 and 6 by the end of this week. Please read the project page as well, which will discuss which versions of notifications/messaging you will need.

thanks
Ian

Switch to mailcomment project. Mail2web is now known as "mailcomment" due to a trademark issue with the name mail2web.