Mailto: anomaly

[ScottHeath]

Webmail usually works very well, but an obscure bug seems to have bit me.

My daughter's school uses a listserv-like program to forward messages from each teacher to all the parents, If you want to reply to a teacher, you can't just "Reply" to the mailing, you have to click on the "From:" field for the teacher. But the teachers' addresses are of the form "Name[at]dpsk12.org" and the "12.org" gets cut off when a new message composition window opens.

I checked the source of the school's message, and the teacher's whole address is there. But every time I click on it, a new "compose message" window opens with the trimmed address :angry: .

Is this a bug in webmail? I haven't found a mention of it in this forum...

[Wazoo]

First time I have heard of this one. The problem here is how to try to work the issue. One would have to see the e-mail involved for starters, and not sure if you'd want to post that directly. The general suggestion would be to submit one of these as a spam, capture the Tracking URL, cancel the parse/report, and provide the Tracking URL here. That would allow some analysis from the folks 'here' to see if there's something obvious. Something silly like a bit of word-wrapping going on such that your teacher's address is actually broken off at the end of a line (though that really does seem odd) .... but without seeing something of the structure involved, there's not really a good way to try to guess at things like this.

Failing that, then you'd have to hit JT directly, with your account details to identify your account .. making sure that there's at least one of these messages still sitting on a SpamCop E-Mail server ... and wait until he has the time to catch your e-mail and do the research ... He's a busy guy, so don't expect an immediate response (though of course, he could surprise us both <g>)

[DavidT]

Scott,

I just ran a quick test of the phenomen, sending myself a forged message with a From like this:

From: "David T." <test[at]dpsk12.org>

I opened the message using SpamCop webmail, clicked on the "From" link to compose a reply, and the system addressed it properly. Therefore, I strongly suspect that there's something about the formatting of the "From" field in the headers of those messages that is causing the problem.

Rather than posting one here, please send the headers, or just the actual "From" field from one of those messages to me privately (click on my user ID for options) and I'll take a quick look.

DT

[StevenUnderwood]

I have never used that function and none of the messages I have available (2 messages and 13 spam at the moment) have the problem you decsribe. Some of these had multiple domain levels as well, and most much longer that the one you are having problems with.

In IE6SP1, when I mouse over the address, the status line shows "Compose Message (address for recipient). In your account, does it show the correct address in that area (if you have a comperable feature)?

Also, does the message source show the correct information (assuming it does)?

Just trying to determine if it is actually a problem with webmail of the format of the message as it was received. I am assuming this is repeatable with other messages from the same source.

[Jeff G.]

Are you mousing over or clicking while reading a message or while reading the index of messages in a mailbox/Folder?

Also, what is the status of "Options / Other Options / Display Options / The From: column of the message should be linked:", of the following?

• Clicking on the address will compose a new message to the sender
  
• Clicking on the address will open the message to be read
  
• Do not generate a link in the From: column
  

[turetzsr]

Webmail usually works very well, but an obscure bug seems to have bit me.

My daughter's school uses a listserv-like program to forward messages from each teacher to all the parents, If you want to reply to a teacher, you can't just "Reply" to the mailing, you have to click on the "From:" field for the teacher.  But the teachers' addresses are of the form "Name[at]dpsk12.org" and the "12.org" gets cut off when a new message composition window opens.

<snip>

22716[/snapback]

...Sounds like what happens when there's a space in the "mailto" link ....

[ScottHeath]

Oops! I forgot that, on this board, you are not subscribed to a thread just because you started it. Lots of wonderful people did respond, and I missed it.

Anyway, the bug is more complicated than I thought. It works like this: the "digest" I receive is properly formed and arrives in straight text and also HTML form. The digest might contain several e-mails, and each e-mail might have a "mailto:" reference. If I click on the mailto: link in the plain text version, the webmail composition window opens with the wrong address (joe1287[at]aol.com becomes joe). But if I call up the html version and click on the same link, my browser opens my own mail program with the complete address in the 'to' field.

I checked the source, and it looks fine, BTW.

Oh, and I did turn on tracking, so I'll see your responses this time.

[DavidT]

Scott was trusting enough to send me a copy of one of the problematic messages, and I've just finished running some tests based upon what I saw. There does indeed seem to be a bug in the current SC implementation of IMP/Horde webmail (there's a new version in a seemingly-stalled Beta testing mode). The bug invoves any "mailto" link in the text portion of an email, which the webmail software tries to make a clickable link, but which break, if the email address includes the numeral 1. The webmail software stops reading the email address at the numeral 1 (but this doesn't happen with pure alpha addresses or addresses containing numerals other than 1).

So, for example, if someone sends you this:

<mailto:abcdxyz[at]this.couldnt.possibly.be.a.domain.name>

then the SC webmail makes it clickable, meaning you can click on it and start composing a message to that address, but if they were to send you this:

<mailto:abcdxyz1[at]this.couldnt.possibly.be.a.domain.name>

then the address in both the mouseover display and also the resulting compose window is truncated at the numeral one, leaving only: abcdxyz

There's a slight difference in the behavior of FireFox vs. IE when mousing over the resulting links and also when clicking on them, in that Firefox grabs only the address, but IE sees the last "greater than" symbol and appends it to the email address.

If the mailto is formatted like this (using quotes):

<"mailto:abcdxyz[at]this.couldnt.possibly.be.a.domain.name">

then that takes care of the IE issue and it will produce similar results in both browsers, but the links still break if they contain a "1." This isn't really related to the problem at hand, however, and can be ignored.

I also opened my test messages in the Beta version of the proposed upgrade to SC webmail, and the problem doesn't happen, so this is an odd bug in the current SC webmail implementation that will go away when the Beta becomes the production version.

Addendum: I did a Google search on "horde bug mailto" (without the quotes) and the very first hit is to an archived March 2003 IMP support list message from SpamCop's very own JT:

http://lists.horde.org/archives/imp/Week-o...310/031166.html

So it's been a problem for at least the last two years.

DT

[ScottHeath]

Scott was trusting enough to send me a copy of one of the problematic messages, and I've just finished running some tests based upon what I saw. There does indeed seem to be a bug in the current SC implementation of IMP/Horde webmail (there's a new version in a seemingly-stalled Beta testing mode). The bug invoves any "mailto" link in the text portion of an email, which the webmail software tries to make a clickable link, but which break, if the email address includes the numeral 1. The webmail software stops reading the email address at the numeral 1 (but this doesn't happen with pure alpha addresses or addresses containing numerals other than 1).

Thanks! That's way more than I needed to know , but that's cool . Nice detective work, even though this particular bug seems to have been documented two years ago by another SC person! At least now I know I encountered a real bug, and I wasn't imagining it. The e-mails that I had tried to send, BTW, did not bounce or tell me their address was not valid. They just went to that big bit-bucket in the sky.

Also, even though I clicked on the 'track this thread' link and got the acknowledgement, I received nothing fom the invision software when David posted his report, but I did get David's off-list message. Maybe the tracking message is in the works...

[StevenUnderwood]

Also, even though I clicked on the  'track this thread' link and got the acknowledgement, I received nothing fom the invision software when David posted his report, but I did get David's off-list message.  Maybe the tracking message is in the works...

26466[/snapback]

I believe the email is only sent if you are not logged into the forum (including the 15 minute lag time) on the basis that you can check the forum if you are logged in.

[ScottHeath]

I believe the email is only sent if you are not logged into the forum (including the 15 minute lag time) on the basis that you can check the forum if you are logged in.

26468[/snapback]

I got a notice from your post, but not his. Or maybe the notice that sent me here this time was for his message, but it came thru in time for me to see yours. It doesn't matter, really, but thanks for the info.