Re: [bug-mailutils] sieve: malformed address when moving to maildir: address@hidden: Re: THETAWORLD ORDER NO. 17459185107]

[Jean Louis]

Hello Sergey,

Those were mailboxes possibly created by Thunderbird, and later
converted by evolution to mboxes. One of those was making it that
way with lines terminating with CRLF, but I don't know which one. Some
mailboxes were standard, some not.

The evolution on its first run, starts converting mailboxes, so I
cannot recommend that software to anybody, for security reasons. I
have just tried running it, and now I have thousands of emails to sort
back.

By listening to your advice, I have now automatically converted
mailboxes with cat mbox | tr -d "\r" > new-mbox and processed
thousands of emails with the sieve.

The processing so far was almost perfect, as few of mailboxes still
remained unprocessed, and messages remained empty, with the original
date. Those are probably a problem on my side.

mutt was handling and reading those messages correctly.

I have spared a lot of time, I need to find out ways to donate for
this software.

Jean Louis

On Mon, Dec 26, 2016 at 12:06:03AM -2200, Sergey Poznyakoff wrote:
> Hi Jean,
> 
> [Cc-ing to the list]
> 
> Thanks for the samples. I can confirm that I reliably get 'Malformed
> email address' for two out of three messages in mailbox Safecharge.com.
> The addresses are malformed indeed, as well as the messages themselves.
> The problem with these messages is that they contain lines terminated
> with CRLF (ASCII 13 10) characters insead of single LF (ASCII
> 10).  Mailutils won't recognize such messages.
> 
> The same holds true for the second mailbox (PaySimple.com).  In both
> mailboxes in question, only the first message has properly terminated
> lines and only this message is processed (I wonder what program could
> have created such mailboxes).
> 
> There is nothing that can be done in that case, except making sure input
> mailboxes are properly formatted (e.g. by filtering them through 
> tr -d '\r').
> 
> Another point: although I was able to reproduce the mailformed address
> error, I wasn't able to reproduce the reported input mailbox
> corruption.  In both testcases, the malformed messages were retained in
> the corresponding input mailboxes.
> 
> Regards,
> Sergey