For a variety of reasons, Mercury is quite strict about enforcing the syntax of the MAIL FROM: and RCTP TO: SMTP commands. The main reason is that syntax errors in these commands are a common indicator of spam (no properly-written SMTP client should ever be getting the protocol this badly wrong). The secondary reason is that a number of open relay testers use deviant or invalid forms like this as part of their test, and evaluate the server based on how it handles them.
As Thomas has pointed out, the syntax for these commands in RFC2821 is quite specific - you can't just throw any address format in there. If your mail clients are getting this so badly wrong, you need to get back to their authors and suggest that they re-read RFC2821, then correct their code.
Installed the beta version, duplicates not resolved, haven't had a chance to see whether the slowness has been fixed. Meantime I did find out that my duplicates issue seems to be Android related; see , though the thread suggests that its a bad sever implementation. But the number of different servers mentioned with the problem imho indicate a client issue nontheless. Still not entirely sure as I had the issue with AgendusMail on the Palm device as well...
It turns out that this was nothing to do with Mercury or Pegasus. We use file permissions to control who has access toeach mail box and the user accoutnin question had not had permission for the administrator account which runs Mercury to access it properly. This was a mistake as it happens. I corrected the directory permissions and the file hieracy appeared correctly.
... and I can make the modifications necessary to make it work with any browser. Which is what the vast majority of webmailers aim to do. Fancy stuff, like drag and drop from inbox to another folder is possible too (see GMX), but that doesn't mean that it has to work that way.
Actually, I don't see why David has to spend the little time he has developing yet another webmailer. That wheel has already been invented. Horde ( http://www.horde.org/webmail/ ) works fine with any IMAP server.
using server 2003 and 4.51: i only have the problem, when i use mercury with the help of a service-wrapper and connect to the console. then i can´t use the help. otherwise no problem.
Thank you, Han. I am going to move this topic to the Pegasus Mail support forum (which is where it should have been in the first place - sorry!).
I'm looking for a solution that doesn't require user interaction. There's two reasons for this; one: I've trained the users to be suspicious of email that wants to be activated in such a fashion (that training has paid off significantly in the past - we got hundreds of email zip viruses on day zero) and two: my users, my management, and I myself want this sort of thing to be handled in a completely user-transparent fashion. These people do not want to mess about tuning their email system, they want to do their real jobs.
Most of Pegasus's configuration is extremely well suited to this sort of thing; flipping switches from N to Y or editing strings is pretty easy to do on a massive scale, and using INI files means that configurations can be stored and backed up using very OS-agnostic methods. There's no putzing about with registry imports and exports that may or may not be written to disk or held in memory at any given time, there's no relative identifiers encoded into anything. The only exception is the bit-mapped and byte-mapped variables, and even those can generally be decoded by simply turning stuff on and off and diffing the resultant INI files.
For example: If I wanted to set the "organization string" for 10,000 users without having to co-ordinate the efforts of 10,000 human beings, I could do it all by myself with four or five lines of code. Pegasus's simple, clean ASCII configuration is what allows me to do so. Registry-driven software is inherently more difficult and cost-ineffective to manage.
mercury32 4.51 crashes if submitted IMAP folder name does not match character set selected in IMAP module configuration.
[/quote]
This problem is fixed in the v4.52 patch, which will be out either today or tomorrow. Note that the "fix" only prevents the crash - if you try to create a folder with a name that cannot be represented using the default character set configured in Mercury, the foldername almost certainly won't be the way you expect it to be, but the program won't crash. In the examples in this thread, where attempts were made to create a folder with a single illegal character as a name, you'll just get a "NO" failure response from Mercury.
The aim is to add full unicode support for folder names in an upcoming release, but for now, this workaround will have to do.
I was also experiencing this with my Palm Treo 700wx and a few other windows mobile devices. I have applied the beta patch and all has been working without issue for a few days now. Thanks so much David!
content control to worry about, 1313 messages of about 10kb to active addresses on
our main mailing list passed through the core process in just under 8
minutes. The list is still running as a VERP list and this is with ClamWall also scanning all outgoing mail. This is running on a 2.8Ghz P4 with 1Gb RAM, running on Server
Yes indeed we have done this and it does work. It was a few weeks back but feel free to mail me chrisc@wwl.co.uk if you have any questions you think we may be able to answer.
OK sorry - think I found the answer in an earlier post ( was stressed and failed to spot it at the time!)
The 'Accept mail for invalid local addresses' tickbox seems to do the trickin MercuryS - but does that apply on older version ( Novell version 1.48) - can't see the line in the ini file?
[/quote]
The line in MERCURY.INI is Allow_Illegals : 0 and is in the [MercuryS] section for v1.48. If this is turned off (off =0, on=1) then MercuryS does the rejection and the bounce goes to the connecting system and not the MAIL FROM: address. If this is a spammer it gets dropped on the floor, if this is a good server the server sends the bounce message back to there sender.
Rules and policies are executed in the core module in Mercury, so once a message has been handed over to MercuryC or MercuryE for delivery it's out of their reach. It might be possible to create a function like the one you describe using the new daemon interface, but as the specifications haven't been released yet I can't say for sure.