rework prerequisite data naming conventions & add some validation

We now give an exhaustive specification of which IDEN1 are valid, rather than
simply reserving some patterns.  Also introduce terminology of "prerequisite
data identifiers".

Drop the idea that IDEN1 might be nil.  Additionally reserve IDEN1 beginning
with three hyphens.  State that third party extensions shouldn't use _CONTEXT.
Disallow forms not listed.

Thanks to David Bremner for discussion.

Signed-off-by: Sean Whitton <spwhitton@spwhitton.name>

2 files changed, 31 insertions, 25 deletions

diff --git a/doc/data.rst b/doc/data.rst
index a32acb1..ae657c6 100644
--- a/doc/data.rst
+++ b/doc/data.rst
@@ -4,32 +4,19 @@ Prerequisite data
 Naming
 ------
 
-A piece of prerequisite data is identified by two strings.  Typically the
-first of these specifies the context in which the data is relevant.  For an
-ssh host key, for example, this context would be a hostname.  If it's ``nil``
-then the data is valid in any context.  The second of these identifies the
-data within its context.  This is often just the filename in which the
-prerequisite data will eventually be stored.  It might also be a
-human-readable string describing the purpose of the data.
-
-Reserved names
-~~~~~~~~~~~~~~
-
-These are exclusive semantics for certain possible pairs of strings
-identifying prerequisite data -- to avoid confusion and potential clashes, do
-not use prerequisite data identified by strings matching these conditions for
-other purposes.
+An item of prerequisite data is identified by two strings, called ``IDEN1``
+and ``IDEN2``; together these are the *prerequisite data identifiers* for an
+item of prerequisite data.  Typically ``IDEN1`` specifies the context in which
+the data is relevant, and ``IDEN2`` identifies the data within its context.
+``IDEN2`` is very often the filename in which the prerequisite data will
+eventually be stored.  It might also be a human-readable string describing the
+purpose of the data.  The following are the valid forms of ``IDEN1``, and
+their meanings.
 
 - ``(HOSTNAME . PATH)`` means the data that should be uploaded to ``PATH`` on
   ``HOSTNAME`` (and usually nowhere else, except in the case of, e.g., a
   public key).  ``PATH`` must be absolute, not relative.
 
-- ``(_CONTEXT . ITEM)`` is an arbitrary prerequisite data context named
-  ``CONTEXT``; typically ``CONTEXT`` will be a network or grouping name,
-  rather than referring to a single host.  ``ITEM`` might be a path or some
-  other identifier.  Reserved for consfigs; will not be used by property
-  definitions included with Consfigurator.
-
 - ``("--lisp-system" . SYSTEM)`` means the data is Lisp code which, when
   loaded, defines the packages and symbols contained in the ASDF system
   ``SYSTEM``
@@ -49,9 +36,25 @@ other purposes.
 - ``("--luks-passphrase" . VOLUME-LABEL)`` means a LUKS passphrase for volume
   with label ``VOLUME-LABEL``.
 
-(Proposed convention: Except for the first two items above, these reserved
-names should start with ``--`` and use ``--`` to separate parameter values
-within the string.  Hostnames cannot start with a hyphen.)
+- Any other ``IDEN1`` beginning with exactly two hyphens is reserved for
+  future use.
+
+- ``(_CONTEXT . ITEM)`` is an arbitrary prerequisite data context named
+  ``CONTEXT``; typically ``CONTEXT`` will be a network or grouping name,
+  rather than referring to a single host.  ``ITEM`` might be a path or some
+  other identifier.  Reserved for consfigs; will not be used by property
+  definitions included with Consfigurator, and should not be used by third
+  party extensions.
+
+- ``(---CONTEXT . ITEM)`` is, similarly, an arbitrary prerequisite data
+  context named ``CONTEXT``.  This form is intended for contexts similar to
+  the reserved names beginning with two hyphens: types of information rather
+  than site-local network or grouping names.  This form will not be used by
+  property definitions included with Consfigurator, but may be used by both
+  consfigs and third party extensions.
+
+Any other forms are invalid.  In particular, an ``IDEN1`` that is not a valid
+hostname and does not begin with a hyphen or an underscore must not be used.
 
 Mechanics
 ---------
diff --git a/src/data.lisp b/src/data.lisp
index 307d0c1..ac5c6e9 100644
--- a/src/data.lisp
+++ b/src/data.lisp
@@ -1,6 +1,6 @@
 ;;; Consfigurator -- Lisp declarative configuration management system
 
-;;; Copyright (C) 2021  Sean Whitton <spwhitton@spwhitton.name>
+;;; Copyright (C) 2021-2022  Sean Whitton <spwhitton@spwhitton.name>
 
 ;;; This file is free software; you can redistribute it and/or modify
 ;;; it under the terms of the GNU General Public License as published by
@@ -173,6 +173,9 @@ This function is called by property :APPLY and :UNAPPLY subroutines."
                      (missing-iden1 condition) (missing-iden2 condition)))))
 
 (defun %get-data (iden1 iden2)
+  (alet (first-char iden1)
+    (unless (or (char= #\_ it) (char= #\- it) (valid-hostname-p iden1))
+      (simple-program-error "Invalid IDEN1: ~S" iden1)))
   (let* ((idenpair (cons iden1 iden2))
          (from-source (query-data-sources iden1 iden2))
          (from-source-version (and from-source (car from-source)))