Re: sysinst split project - The Configuration File

[David Holland <dholland-current%netbsd.org@localhost>]

On Tue, Dec 15, 2009 at 01:03:54PM -0200, Silas Silva wrote:
 > Well, I was in doubt after reading the Turing-completeness problem of
 > configuration vs. scripting and whether or not sysinst.conf should be
 > Turing complete. But I mulled over it and I think using Lua on the
 > project would be a good thing.

It seems to me that there are at least three different questions here.

One is: whether canned answers to canned questions (like "which
password cipher?" or "what IP address do I use?") should be expressed
as static data or as script hooks.

The second is: whether it should be possible to direct the overall
progress of the install process using a user-supplied script, which
e.g. plugs together components of sysinst and executes them.

The third is: whether pieces of sysinst itself should be written as
scripts, and if so, whether part of the available configuration
process should involve editing those scripts.

ISTM that canned answers should be expressed as canned data. (Or
possibly via some non-Turing-complete indirect map, e.g. based on the
hostname or IP address. But even then I think it'd probably be better
to map from hostnames to complete install profiles rather than retype
the map for every install question.)

It also seems to me though that directing the overall progress of the
install is useful and perhaps important, and that should be done in a
script language. Striking the right balance between this and requiring
the user to edit complex scripts that ought to be internal to sysinst
will take some thought.

 > Couldn't sysinst import/export from/to both formats, human-readable (Lua
 > script or another) and XML? That would solve the controversy, but it may
 > be difficult to maintain support for two different standards.

Trying to transform scripts into XML is a bad idea.

-- 
David A. Holland
dholland%netbsd.org@localhost