From: Noel Jones (njonesmegan.vbhcs.org)
Date: Thu Oct 11 2007 - 14:56:45 CDT

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

At 02:13 PM 10/11/2007, Terry Carmen wrote:
>>Yes, those commands demonstrate how to test or query a cidr table.
>>It doesn't say you need to compile the cidr table. It's not
>>practical for the documentation to list everything you should NOT do.
>Although it's not practical to list everything that the user should
>not do with postmap, it would make things clearer if the docs simply
>listed the various file types and mentioned which get compiled and which don't.

It wouldn't be unreasonable to add something to the documentation if
it can be stated in a concise way that is likely to reduce
confusion. Documentation patches are always welcome, but not always adopted.

Wording indicating "don't postmap these tables" would be needed for
the {regexp, pcre, cidr}_table man pages, and maybe for the
header_checks, access, canonical, virtual, and transport pages too
since they mention using postmap to create an index file. Original
files are in the proto directory of the tarball.

>It's not a huge deal, but would have saved me (and I suspect other
>people) a bunch of time.

Unnecessary postmap of a table only costs you the few seconds to type
the command and doesn't affect postfix operation at all.

If I remember correctly, your original problems were using CIDR style
notation in a hash style access table and using check_sender_access
where check_client_access was needed. Again, specific suggestions or
patches for improving the documentation are welcome.

--
Noel Jones

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]