Hi, the inflationary use of the word "virtual" connected with almost everything has been accompanying this list, exim, the Debian exim packages and all support media. The latest occurence of this is Debian bug #408467 (http://bugs.debian.org/408467), where I was faced with insults after denying to install a "virtual" router to the Debian default configurations. To make communication about this a little easier, I would like to have more "formal" definition of a "virtual foo" in exim context. I have created a page on my personal wiki, http://wiki.zugschlus.de/doku.php?id=definingvirtualforexim, and would like to solicit your comments and improvements. Once this document reaches maturity, I will try to find "new" words for the features we define to avoid misunderstandings with people asking for virtual stuff. I hope that the result of that discussion will be a list of definitions what is exactly meant when somebody talks about a "virtual foo", which could be more exactly supported by having boilerplate config snippets. Greetings Marc -- ----------------------------------------------------------------------------- Marc Haber | "I don''t trust Computers. They | Mailadresse im Header Mannheim, Germany | lose things." Winona Ryder | Fon: *49 621 72739834 Nordisch by Nature | How to make an American Quilt | Fax: *49 621 72739835
This wiki section describes exactly what I always used the word "virtual" in email context ... F?bio Rabelo Marc Haber escreveu:> Hi, > > the inflationary use of the word "virtual" connected with almost > everything has been accompanying this list, exim, the Debian exim > packages and all support media. The latest occurence of this is Debian > bug #408467 (http://bugs.debian.org/408467), where I was faced with > insults after denying to install a "virtual" router to the Debian > default configurations. > > To make communication about this a little easier, I would like to have > more "formal" definition of a "virtual foo" in exim context. I have > created a page on my personal wiki, > http://wiki.zugschlus.de/doku.php?id=definingvirtualforexim, and would > like to solicit your comments and improvements. > > Once this document reaches maturity, I will try to find "new" words > for the features we define to avoid misunderstandings with people > asking for virtual stuff. > > I hope that the result of that discussion will be a list of > definitions what is exactly meant when somebody talks about a "virtual > foo", which could be more exactly supported by having boilerplate > config snippets. > > Greetings > Marc > >
> This wiki section describes exactly what I always used the word > "virtual" in email context ...But without saying which of the multiple definitions on there describes your value of "exactly" it doesn''t help. As it happens, it describes mine too. Option a) under virtual domains. When I was upgrading from Exim 3 to Exim 4, the two definitions of a virtual router I found in the documentation didn''t do what the virtual router we had under Exim 3 actually did. In particular, neither supported the wildcard in the alias-like file. Fortunately, the migration tool helped me fix that problem, and thats what I call a virtual router. Of course, since then, we''ve removed most of the wild cards, thanks to the spammers. Regards Neil ------------------------------------------------------------------------------------ Scanned for viruses, spam and offensive content by CensorNet MailSafe Professional Web & E-mail Filtering from www.censornet.com
On 2 Feb 2007, at 12:03, Neil Briscoe wrote:>> This wiki section describes exactly what I always used the word >> "virtual" in email context ... > > But without saying which of the multiple definitions on there > describes > your value of "exactly" it doesn''t help.Virtually all of them? [/me goes and hides for a while] Nigel. -- [ Nigel Metheringham Nigel.Metheringham@InTechnology.co.uk ] [ - Comments in this message are my own and not ITO opinion/policy - ]
On Fri, Feb 02, 2007 at 11:47:43AM +0100, Marc Haber wrote:> To make communication about this a little easier, I would like to have > more "formal" definition of a "virtual foo" in exim context. I have > created a page on my personal wiki, > http://wiki.zugschlus.de/doku.php?id=definingvirtualforexim, and would > like to solicit your comments and improvements. > > Once this document reaches maturity, I will try to find "new" words > for the features we define to avoid misunderstandings with people > asking for virtual stuff.This has now been done, taking Tony''s advice to shun the term "virtual" altogether. Please review http://wiki.zugschlus.de/doku.php?id=definingvirtualforexim and tell me what you think. Greetings Marc -- ----------------------------------------------------------------------------- Marc Haber | "I don''t trust Computers. They | Mailadresse im Header Mannheim, Germany | lose things." Winona Ryder | Fon: *49 621 72739834 Nordisch by Nature | How to make an American Quilt | Fax: *49 621 72739835
Hi Tony, nice to hear from you. On Wed, Feb 14, 2007 at 04:56:08PM +0000, Tony Finch wrote:> I have a > suggestion for a more axiomatic/orthogonal approach. > > * Does the domain use the traditional system aliases file, or its own > aliases file, or a shared (fully-qualified) aliases file, or none? > > * Is email to this domain delivered locally? If so, are unix user names > valid local parts at this domain? Or is the message store independent > of unix accounts? > > * If multiple domains point to the same message store, are local parts at > the various domains the same across all the domains or different? Does > the same local part at different domains refer to different mailboxes? > > * Is the domain supposed to be entirely equivalent to some other domain?I really really like that idea, but it comes with the price that the configuration recipies that might be delivered with this approach depart rather radically from what the community is used to. Having corresponding experiences with the configuration itself, I''d rather stay more along the lines of the common approaches. Would it be acceptable to (a) give "new" definition for the common "virtual foo" stuff as we already have and (b) additionally give a second section of explanations and examples following your new scheme? Or do you think this would add to general confusion? Greetings Marc -- ----------------------------------------------------------------------------- Marc Haber | "I don''t trust Computers. They | Mailadresse im Header Mannheim, Germany | lose things." Winona Ryder | Fon: *49 621 72739834 Nordisch by Nature | How to make an American Quilt | Fax: *49 621 72739835
Marc Haber <mh+pkg-exim4-users@zugschlus.de> wrote:> >http://wiki.zugschlus.de/doku.php?id=definingvirtualforexim and tellIt seems pretty good to me. I''ve made a few minor edits, but I have a suggestion for a more axiomatic/orthogonal approach. * Does the domain use the traditional system aliases file, or its own aliases file, or a shared (fully-qualified) aliases file, or none? * Is email to this domain delivered locally? If so, are unix user names valid local parts at this domain? Or is the message store independent of unix accounts? * If multiple domains point to the same message store, are local parts at the various domains the same across all the domains or different? Does the same local part at different domains refer to different mailboxes? * Is the domain supposed to be entirely equivalent to some other domain? Tony. -- f.a.n.finch <dot@dotat.at> http://dotat.at/ PLYMOUTH: NORTHWEST BACKING SOUTH 4 OR 5, INCREASING 6 TO GALE 8. SLIGHT OR MODERATE INCREASING VERY ROUGH. FAIR. GOOD.
Marc Haber <mh+pkg-exim4-users@zugschlus.de> wrote:> >I really really like that idea, but it comes with the price that the >configuration recipies that might be delivered with this approach >depart rather radically from what the community is used to.I was thinking more of using this as a way of classifying the config recipes and for guiding conversations with users, more than anything else. Tony. -- f.a.n.finch <dot@dotat.at> http://dotat.at/ SHANNON ROCKALL: SOUTHERLY GALE 8 TO STORM 10. VERY HIGH. OCCASIONAL RAIN. MODERATE OR POOR.
Just a suggestion: you define Non-System User An e-mail address that is not associated with a local system account but which uses the local part as input for the routing decision. Incoming mail for such addresses can either be delivered to a non-system mail spool, a ?real? system user or forwarded elsewhere. This is frequently called a ?Virtual User? My suggestion: Non-System User has the potential for just as much confusion as Virtual User. The System User/Non-System User distinction could be seen as uid<1000 vs. uid>1000. Perhaps another name should be found, although I don''t have a good alternative. Non-local user would be one way to say it, although in an MTA context this is even more confusing. How about Non-UNIX user? Non-Local Account user? Cheers, Alex -- Alex King Linuxworks Phone: +64 3 473 1611 Mobile: +64 21 410 420
Hi Alex> How about Non-UNIX user?I think I prefer that definition. I had cause to modify a router I''d written for that database based system you''d helped me with. The need for the change was precisely because I couldn''t use checks like check_local_user, since as far as the system, and thus Exim, is concerned, none of them are. Regards Neil ------------------------------------------------------------------------------------ Scanned for viruses, spam and offensive content by CensorNet MailSafe Professional Web & E-mail Filtering from www.censornet.com
On Wed, Feb 14, 2007 at 05:43:04PM +0000, Tony Finch wrote:> Marc Haber <mh+pkg-exim4-users@zugschlus.de> wrote: > >I really really like that idea, but it comes with the price that the > >configuration recipies that might be delivered with this approach > >depart rather radically from what the community is used to. > > I was thinking more of using this as a way of classifying the config > recipes and for guiding conversations with users, more than anything else.I understand. Thanks. Greetings Marc -- ----------------------------------------------------------------------------- Marc Haber | "I don''t trust Computers. They | Mailadresse im Header Mannheim, Germany | lose things." Winona Ryder | Fon: *49 621 72739834 Nordisch by Nature | How to make an American Quilt | Fax: *49 621 72739835
On Wed, Feb 14, 2007 at 04:56:08PM +0000, Tony Finch wrote:> * Is email to this domain delivered locally? If so, are unix user names > valid local parts at this domain? Or is the message store independent > of unix accounts? > > * If multiple domains point to the same message store, are local parts at > the various domains the same across all the domains or different? Does > the same local part at different domains refer to different mailboxes?These two introduce the new expression "message store" that was up to now not in use in the definition document. Is it possible to rewrite the two points to only refer to expressions that are already in use, or (preferred) give a definition of "message store"? Other than that, I have put the definitions nearly verbatim on the wiki page (http://wiki.zugschlus.de/doku.php?id=definingvirtualforexim) to avoid losing them. Greetings Marc -- ----------------------------------------------------------------------------- Marc Haber | "I don''t trust Computers. They | Mailadresse im Header Mannheim, Germany | lose things." Winona Ryder | Fon: *49 621 72739834 Nordisch by Nature | How to make an American Quilt | Fax: *49 621 72739835
On Sat, Feb 24, 2007 at 10:32:26PM +0000, Tony Finch wrote:> Marc Haber <mh+pkg-exim4-users@zugschlus.de> wrote: > >These two introduce the new expression "message store" that was up to > >now not in use in the definition document. Is it possible to rewrite > >the two points to only refer to expressions that are already in use, > >or (preferred) give a definition of "message store"? > > Your other categories used the term "mail spool" which I avoid because > it can be confused with the MTA queue.Agreed. So we can get consensus when I ditch my expression "mail spool", replacing it with "mail store"? Greetings Marc -- ----------------------------------------------------------------------------- Marc Haber | "I don''t trust Computers. They | Mailadresse im Header Mannheim, Germany | lose things." Winona Ryder | Fon: *49 621 72739834 Nordisch by Nature | How to make an American Quilt | Fax: *49 621 72739835
On Sat, Feb 24, 2007 at 11:16:05PM +0100, Marc Haber wrote:> These two introduce the new expression "message store" that was up to > now not in use in the definition document. Is it possible to rewrite > the two points to only refer to expressions that are already in use, > or (preferred) give a definition of "message store"?I have given the definition a try in the wiki, and have rewritten the other definitions to use message store instead of mail store and/or mail spool. Comments? Greetings Marc -- ----------------------------------------------------------------------------- Marc Haber | "I don''t trust Computers. They | Mailadresse im Header Mannheim, Germany | lose things." Winona Ryder | Fon: *49 621 72739834 Nordisch by Nature | How to make an American Quilt | Fax: *49 621 72739835
Marc Haber <mh+pkg-exim4-users@zugschlus.de> wrote:> >These two introduce the new expression "message store" that was up to >now not in use in the definition document. Is it possible to rewrite >the two points to only refer to expressions that are already in use, >or (preferred) give a definition of "message store"?Your other categories used the term "mail spool" which I avoid because it can be confused with the MTA queue. Tony. -- f.a.n.finch <dot@dotat.at> http://dotat.at/ BISCAY FITZROY: SOUTHWESTERLY 5 TO 7, OCCASIONALLY GALE 8. ROUGH OR VERY ROUGH. THUNDERY SHOWERS. GOOD.