INN config file parsing infrastructure

[Russ Allbery]

Fabien Tassin <fta at oleane.net> writes:

> I really would like to discourage usage of non-semantic comments...  I
> agree that such comments are very useful for people that manage their
> servers by hand but if you want to use a GUI (or Web based admin tools,
> a conf managers, etc.), comments will not survive between changes.
> That's the purpose of the 'description' and 'skip' fields in
> incoming.conf.  Unfortunatly, I've never had time to finish the project
> I've had in mind :(

> IHMO, such a lib should be able to do some manipulations on a file such as :
> - check/read/write a whole file (and perhaps its dependancies)
> - add/remove/desactivate/activate a section or a group
> - launch optionnal actions for each event (previously quoted) using hooks.

The checking we definitely want regardless of how the file is managed.
For the rest, I think that most users are going to fall into two
categories:  people who want to use a GUI, tool, or some other automated
way to manage the file; and people who just edit them with a text editor.
(I'm likely to stay in the latter category for a long time, for example.)

Given that, I don't expect people to be doing a lot of hand-editing files
that are maintained by a library or vice versa, so I'd consider it
acceptable if, when a hand-edited file were read in and written out using
a library behind a GUI app, information not used by INN (like non-semantic
comments) were lost and insigificant (to INN) things like the order of
keys inside a group or the exact encoding of strings with special
characters in them were changed.

Maybe I'm wrong on this, though....

-- 
Russ Allbery (rra at stanford.edu)             <http://www.eyrie.org/~eagle/>