hellkvist.org

[cristian]

Hi all

I install peffisaur on my server and everything works fine in email-mode except for one mobile operator that send a strange mms via email.

The message (msgtxt) is into an html email !!!

Ho can I modify the recmail.pl script to read and save this message from the html ??

Any idea ??

Ciao and thank you !

[Peffis]

I see. The changes require some perl knowledge I'm afraid. You should have a look at the getTextForMid function in recmail.pl. First change the SQL query so that it says

[cristian]

Thank you for your reply but the problem is that after the unpack process into the msgstore->email folder i don't have any msg txt file!!

I notice that the message is only in to the html that we can see into the email client.

So I think that the only way is to save the message from the while cicle into the recmail.pl. (where we assing $to, $from, $subject ...)

Ciao

[Peffis]

Hmmm, so munpack doesn't manage to extract the html-part? I thought that was sent just as any other attachment, but it was a long time ago since I had a look at email at that level. Perhaps there is some exception here that munpack doesn't handle, or perhaps there is some switch that you can apply to munpack to make it work. Could you send me a sample message on email if you have one? I could have a look perhaps.

Extracting and storing the message in the beginning is a bit tricky really. The process is reading from standard in and after processing the headers (such as to, from and so on) stdin is piped on to munpack. So if you read through the message and storing it to file there would be nothing left for munpack to read. But you could of course then pipe the just stored file to munpack instead and you are all done.

I'm thinking that there must probably be hundreds of perl modules out there that does exactly this - parsing email and extracting parts like munpack does it. Perhaps one could consider rewriting the script completely later on. I also think there is email-parsing support in PHP even. Perhaps the script should be in PHP instead, thus it could use the same config file and have the whole thing in the same language. As it is now, the email support is just a quick hack as you can see.

[cristian]

If you want I can send you an MMS - email with this mobile operator and you can show wath happen?

What do you think ??

Ciao

Cristian

[Peffis]

Yes, do that! (I suppose you have my email address)

[cristian]

Yes I have your email address and just now I send you the MMS.

Ciao

[Peffis]

Thanks, for that particular message you will really need to do some parsing to get the message out because it's deep down there within the nested tables. I'm afraid I don't have the time to do that specific parsing for you right now so I would stick with the idea of writing it all to a file first, detect and parse the HTML and then pass the rest to munpack.

I still haven't tested to unpack it though. I'll do that and see if munpack does unpack the html-content or not.

[cristian]

Thank you for your help!

Into your last post I found the key!!!!

munpack option -t save the html into the email folder and so I need only a perl script to extract the message from html.

Thank you,

Cristian

[Peffis]

That's great! Thanks for sharing that info.

If it was only removal of tags you were after there are tons of perl lines out there on the net. Unfortunately in your example there is lots of other junk as well, such as operator messages, so it will take some parsing unfortunately. But perhaps it can be made simple and hard-coded, like taking the 5th (I haven't counted) "<td>" from the start or something

[cristian]

I think that I found the solution.

Whit a perl script I found the text between a specific and distinct html tag (table cell).

The script works well but I have a strange character at the end of the text. Now I try to remove with chomp() function.

What do you think ??

Executing the script terminal side I don't have this character. But with the recmail.pl found it into my Mysql field.

Ciao and thank you.

[Peffis]

Hmm, I don't know really where that character can come from. But chomping it off of cutting off the last char should do the trick if it's there all the time I guess.