Rework tool importation along with cli module

Specifically, now config.<tool> (like config.cli) is handled by the import
machinery (it is like a shorter alias for config.import.<tool>.<tool>.exe
that we already had). And the cli module now uses that instead of custom
logic.

This also adds support for uniform tool metadata extraction that is handled by
the import machinery. As a result, a tool that follows the "build2 way" can be
imported with metadata by the buildfile and/or corresponding module without
any tool-specific code or brittleness associated with parsing --version or
similar outputs. See the cli tool/module for details.

Finally, two new flavors of the import directive are now supported: import!
triggers immediate importation skipping any rule-specific logic while import?
is optional import (analogous to using?). Note that optional import is always
immediate. There is also the import-specific metadata attribute which can be
specified for these two import flavors in order to trigger metadata
importation. For example:

import? [metadata] cli = cli%exe{cli}

if ($cli != [null])
  info "cli version $($cli:cli.version)"

1 files changed, 14 insertions, 0 deletions

diff --git a/libbuild2/config/utility.hxx b/libbuild2/config/utility.hxx
index 614791f..fb3023b 100644
--- a/libbuild2/config/utility.hxx
+++ b/libbuild2/config/utility.hxx
@@ -311,6 +311,20 @@ namespace build2
     // be used to "remember" that the module is left unconfigured in order to
     // avoid re-running the tests, etc.
     //
+    // @@ This functionality is WIP/unused and still has a number of issues:
+    //
+    // - This seems to be a subset of a bigger problem of caching discovered
+    //   configuration results. In fact, what we do in the configured case,
+    //   for example in the cc module (multiple path extraction runs, etc), is
+    //   a lot more expensive.
+    //
+    // - The current semantics does not work well for the case where, say, the
+    //   missing tool has appeared in PATH and can now be used via the default
+    //   configuration. In fact, even reconfiguring will not help without a
+    //   "nudge" (e.g., config.<tool>=<tool>). So maybe this value should be
+    //   ignored during configuration? See the "Tool importation: unconfigured
+    //   state" page for more notes.
+    //
     LIBBUILD2_SYMEXPORT bool
     unconfigured (scope& rs, const string& var);