>> >> 2) it's not intuitive to me that the query I have given to
>> >> postifx >> >> is conditionally executed. >> > >> > And THAT is precisely what the documentation says. >> >> And THAT is merely what I was asking to have confirmation of >> (not sure why that is so difficult)... > > It is difficult only because you prefer to believe fanciful > interpretations a plainly stated behaviour. NO. If you read my inquiry, I did read the documentation and understand it correctly. I merely had some lingering doubts and was asking for simple confirmation. Everyone's steadfast refusal to confirm (nothing more than "yes" would have been necessary), instead responding with hostile responses like yours is alienating, especially to a polite inquiry. Is this how you treat people face to face as well? Everyone has their own interpretations of what they read, based on varying degrees of expertise. If you refuse to hear the impression others get from reading the documentation, then maybe it will never improve. > It is intuititive if you understand that Postfix makes multiple > lookups for a single key, some of which contain only the user or > the domain part. The low-level table driver (LDAP, MySQL, ...) > does not know which part of the address it is handed, or even > whether the lookup key is an address at all. > > Thus, when a lookup key does not resembe an full email address, it > is not possible to determine whether the key has a domain part or > not. Instead of making meaningless queries that may produce incorrect > results, Postfix suppresses the query. If Postfix knows what's going on, then it may be able to communicate the situation through to the backend. I don't follow you entirely, since my response to this would be that if the query needs to know what part of the address the key is, it should be able to learn this by examining %u or %d to be empty/NULL when the key doesn't have @ in it. Of course, the query would only need to do this if it had advanced requirements. But I don't have the level of intimate knowledge of the workings and only speak as a user, so I guess you can dismiss my input with a Monty Python wisecrack. >> and also because this fact is IMO buried too deep in the documentation. > > It is documented in the paragraph that describes the "%d" syntax. > Don't use "%d" in queries without reading what it does. There is > surely no better place to describe "%d" semantics than in the > associated paragraph of the query syntax definitions. I am not suggesting that text be removed, only that it is surprising to learn that the query is ignored in some cases and this might be a good thing to note elsewhere in addition. But it doesn't really sound like you want feedback anyway. <shrug>