Page 2 of 2

PostPosted: Sun Aug 22, 2004 12:46 am
by SysKoll
Besides, this say that you may, under special circumstance, use the double-quotes even when they are not necessary. It certainly doesn't authorize you to refuse legal strings. Our addresses (without quotes) are legal as per RFC2822, so GMX has to accept it or admit they break the RFC.

Hope it helps.

PostPosted: Sun Aug 22, 2004 2:19 am
by josh
The term that grabs my attention from all this is "locally interpreted" -- where does it say that a quoted value must be interpreted the same as it's unquoted equivalent? My interpretation was that "locally interpreted" means just that -- the destination host does what it wants with that part -- quotes could mean one thing and no-quotes could mean another. Is that wrong? We certainly accept both forms, and acceptance seems to be the subject of the RFCs.

Normally, I'd just go try and strip the quotes and be done with it -- the problem is that we'll have to re-think our header parsing because we use quotes to identify the "display name" and separate it from the account name -- eg, if the to: is:

"Bob Example" <bob@example.com>

we're zooming in on the quotes to get the Bob Example part for display. Rethinking our code in order to conform to an RFC is something we're happy to do (but, let's face it -- not required to do -- spamgourmet doesn't claim to be an SMTP compliant mail server) -- rethinking it because somebody else's supposedly SMTP compliant server doesn't conform is kind of a pain...

PostPosted: Sun Aug 22, 2004 10:57 am
by Skeeve
SysKoll wrote:Besides, this say that you may, under special circumstance, use the double-quotes even when they are not necessary. It certainly doesn't authorize you to refuse legal strings. Our addresses (without quotes) are legal as per RFC2822, so GMX has to accept it or admit they break the RFC.

Hope it helps.

100% agree. I will wait until monday and see what they will tell me. If I don't get a satisfying reply, I'll tell c't (IMHO best german computer magazin) about GMX breaking RFCs ;-)

PostPosted: Sun Aug 22, 2004 11:31 am
by Skeeve
josh wrote:The term that grabs my attention from all this is "locally interpreted" -- where does it say that a quoted value must be interpreted the same as it's unquoted equivalent?

I see that from this example of RFC 822:
Code: Select all
        Note:  In particular, quoting is NOT permitted  within  atoms.
               For  example  when  the local-part of an addr-spec must
               contain a special character, a quoted  string  must  be
               used.  Therefore, a specification such as:

                            Full\ Name@Domain

               is not legal and must be specified as:

                            "Full Name"@Domain


My interpretation was that "locally interpreted" means just that -- the destination host does what it wants with that part -- quotes could mean one thing and no-quotes could mean another. Is that wrong?

It's my interpretation of RFC's example, that your interpretation is wrong. BTW: I couldn't find the term "locally interpreted" in 822 or 821.

Normally, I'd just go try and strip the quotes and be done with it -- the problem is that we'll have to re-think our header parsing because we use quotes to identify the "display name" and separate it from the account name -- eg, if the to: is:

"Bob Example" <bob@example.com>

we're zooming in on the quotes to get the Bob Example part for display.

I think you'll have to rethink it already because ther is no need for double quotes!
Code: Select all
     mailbox     =  addr-spec                    ; simple address
                 /  phrase route-addr            ; name & addr-spec

     route-addr  =  "<" [route] addr-spec ">"

     phrase      =  1*word                       ; Sequence of words

     word        =  atom / quoted-string

So both of these are legal:
Code: Select all
    "Mikey-the-Mouse" <mikey@duck.village>

    Mikey-the-Mouse <mikey@duck.village>

You'll have to focus on the "<" ... ">" part!

Rethinking our code in order to conform to an RFC is something we're happy to do (but, let's face it -- not required to do -- spamgourmet doesn't claim to be an SMTP compliant mail server) -- rethinking it because somebody else's supposedly SMTP compliant server doesn't conform is kind of a pain...


I totally agree with you. I don't want to force you to do anything about it. My first intent was to verify that spamgourmet's addresses are legal and to show GMX that they are wrong. Then I found out about the double quotes, allowing me to work around GMX' bug. Unfortunately this led me to find out about spamgourmet's non-RFC compliancy (at least I think so). Do whatever you like with that information and let me tell you that I will be more than happy to help wherever I can. I'm a quite experienced perl programmer (You can find me on perlmonks.org too), so if you need my time or knowledge, then tell me.

PostPosted: Mon Aug 23, 2004 12:21 am
by josh
OK, never mind -- it's not that big of a deal to strip the quotes after all. Try it again.

PostPosted: Mon Aug 23, 2004 5:32 am
by Skeeve
josh wrote:OK, never mind -- it's not that big of a deal to strip the quotes after all. Try it again.

Thanks for implementing it but GMX is Oh-So-Clever! *sigh* They break RFC 821 by adding quotes and backslashes.

So instead of (e.g.) sending
RCPT TO: <"Mickey-the-Mouse"@my.do.main>

They try to be clever again and send to
RCPT TO: <""Mickey-the-Mouse""@my.do.main>

Oh how I hate it!

You should - if you can and your rules allow it - mail all of your GMX users to "bombard" GMX with support questions!

P.S. I already mailed c't this morning about GMX boycoting spamgourmet by ignoring RFCs 821 and 822

P.P.S. If you mail all of your GMX users, I will be more than happy to supply you with a german text for that.

P.P.P.S. They replied and they seem to want to do something about it.

PostPosted: Tue Feb 13, 2007 7:04 pm
by dcw601
Guess what - gmail doesn't like the leading "+" either. It works for a reply, or a hand-entered address, but it won't let me store it as a contact. Definitely a gmail bug.

PostPosted: Tue Feb 13, 2007 9:01 pm
by SysKoll
dcw601 wrote:Guess what - gmail doesn't like the leading "+" either. It works for a reply, or a hand-entered address, but it won't let me store it as a contact. Definitely a gmail bug.


Hum, interesting. I'll reread the specs and try to see if there is a part of the RFC I can pinpoint for the Google techs. (If you happen to know which part of the RFC specifies legal characters, feel free to post it!)