Squashed commit of the following:

* Move things around a bit more
* Add project structure details to HACKING.md
* Create setup.py
* Split off A Boneless Datastructure Language

13 files changed, 190 insertions, 943 deletions

diff --git a/HACKING.md b/HACKING.md
new file mode 100644
index 0000000..ef461fa
--- /dev/null
+++ b/HACKING.md
@@ -0,0 +1,25 @@
+Hacking GAnarchy
+================
+
+Project Structure
+-----------------
+
+`requirements.txt` lists known-good, frozen dependencies. if needed or desired, install dependencies listed in setup.py directly.
+
+GAnarchy uses GAnarchy-based dependencies. they are identified by `gan$COMMIT` names. e.g. in requirements.txt:
+
+```
+-e git+https://soniex2.autistic.space/git-repos/abdl.git@1b26ad799217af7e187fdae78e862a6bf46e5591#egg=gan0f74bd87a23b515b45da7e6f5d9cc82380443dab
+```
+
+or in setup.py:
+
+```
+install_requires=[
+    "gan0f74bd87a23b515b45da7e6f5d9cc82380443dab",  # a boneless datastructure library
+]
+```
+
+(the comment is just a hint for humans to read)
+
+note however that not all forks are compatible with the project. requirements.txt provides known-good versions.
diff --git a/abdl/__init__.py b/abdl/__init__.py
deleted file mode 100644
index 8dde742..0000000
--- a/abdl/__init__.py
+++ /dev/null
@@ -1,528 +0,0 @@
-# A Boneless Datastructure Language
-# Copyright (C) 2019  Soni L.
-#
-# This program is free software: you can redistribute it and/or modify
-# it under the terms of the GNU Affero General Public License as published by
-# the Free Software Foundation, either version 3 of the License, or
-# (at your option) any later version.
-#
-# This program is distributed in the hope that it will be useful,
-# but WITHOUT ANY WARRANTY; without even the implied warranty of
-# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-# GNU Affero General Public License for more details.
-#
-# You should have received a copy of the GNU Affero General Public License
-# along with this program.  If not, see <https://www.gnu.org/licenses/>.
-
-"""A Boneless Datastructure Language, version 2.0.1.
-
-ABDL expressions are regex-like constructs for matching and validating object structures. They can be used
-with JSON and similar formats, and even self-referential data structures.
-
-Language Reference:
-
-    ABDL expressions have the ability to iterate, index, validate and filter data structures. This is
-    done with the use of the syntax elements listed below.
-
-    Syntax Elements:
-
-        An ABDL expression is a sequence of zero or more sequences starting with arrows followed by zero or
-        more subvalues.
-
-        An arrow is ``->`` and indicates indexing/iteration (Mappings, Sequences, Sets). It must be followed
-        by a variable, literal, parameter, regex or key match.
-
-        A variable is a string of alphanumeric characters, not starting with a digit. It may be followed by a
-        literal, parameter, regex, key match, or one or more type tests. A ``(key, value)`` tuple containing
-        the corresponding matched element will be identified by this name in the results dict.
-
-        A literal is a string delimited by single quotes (use ``%'`` to escape ``'`` and ``%%`` to escape ``%``).
-        A literal can be made "non-validating" by appending an ``?`` after it. It may be followed by one or more
-        type tests. It is exactly equivalent to indexing an object with a string key.
-
-        A parameter is the symbol ``$`` followed by a string of alphanumeric characters, not starting with
-        a digit. A parameter can be made "non-validating" by appending an ``?`` after it. It may be followed by
-        one or more type tests. It is exactly equivalent to indexing an object with an arbitrary object key.
-
-        A regex is an RE, as defined by the ``re`` module, delimited by forward slashes (use ``%/`` to escape
-        ``/`` and ``%%`` to escape ``%``). A regex can be made "non-validating" by appending an ``?`` after it.
-        It may be followed by one or more type tests. It attempts to match each key in the object.
-
-        A type test is ``:`` followed by a parameter. A type test can be made "non-validating" by appending
-        an ``?`` after the ``:``. It attempts to match the type of each matched value in the object.
-
-        A key match is an ABDL expression enclosed in ``[`` and ``]``, optionally prefixed with one or more type
-        tests. This matches keys (including the type tests).
-
-        A subvalue is an ABDL expression enclosed in ``(`` and ``)``. This allows matching multiple values on
-        the same object.
-
-        Some syntax elements can be validating or non-validating. Validating syntax elements will raise a
-        :py:exc:`abdl.ValidationError` whenever a non-matching element is encountered, whereas non-validating
-        ones will skip them. Note that it is possible for a validating syntax element to still yield results
-        before raising a :py:exc:`abdl.ValidationError`, so one needs to be careful when writing code where such
-        behaviour could result in a security vulnerability.
-
-    Examples:
-
-        >>> import abdl
-        >>> for m in abdl.match("->X:?$dict->Y", {"foo": 1, "bar": {"baz": 2}}, {'dict': dict}):
-        ...     print(m['X'][0], m['Y'][0], m['Y'][1])
-        bar baz 2
-
-        >>> pat = abdl.compile('''-> 'projects'?
-        ...                          -> commit /[0-9a-fA-F]{40}|[0-9a-fA-F]{64}/? :?$dict
-        ...                             -> url :?$dict
-        ...                                -> branch :?$dict''', {'dict': dict})
-        >>> data = {"projects": {
-        ...     "385e734a52e13949a7a5c71827f6de920dbfea43": {
-        ...         "https://soniex2.autistic.space/git-repos/ganarchy.git": {"HEAD": {"active": True}}
-        ...     }
-        ... }}
-        >>> for m in pat.match(data):
-        ...     print(m['commit'][0], m['url'][0], m['branch'][0], m['branch'][1])
-        385e734a52e13949a7a5c71827f6de920dbfea43 https://soniex2.autistic.space/git-repos/ganarchy.git HEAD {'active': True}
-
-    (If ``:?$dict`` wasn't present, a TypeError would be raised when trying to iterate the ``1`` from ``"foo": 1``.)
-"""
-
-import re
-
-from collections.abc import Mapping, Sequence, Iterator, Set
-
-class DeprecationError(Exception):
-    """Raised for deprecated features, if they are disabled.
-
-    This class controls warning/error behaviour of deprecated features."""
-    #enable_key_match_compat = False
-    #warn_key_match_compat = False
-
-    @classmethod
-    def warn_all(cls):
-        """Enables all deprecation warnings."""
-        pass
-
-class PatternError(Exception):
-    """Raised for invalid input or output expressions."""
-    # TODO implement formatting
-
-    def __init__(self, msg, pattern, defs, pos, toks):
-        self.msg = msg
-        self.pattern = pattern
-        self.defs = defs
-        self.pos = pos
-        self._toks = toks # internal use
-
-    def _normalize(self, pattern, defs):
-        if pattern is not None:
-            if self.pattern is not None:
-                raise ValueError("Attempt to normalize normalized pattern")
-            else:
-                self.pattern = pattern
-        if defs is not None:
-            if self.defs is not None:
-                raise ValueError("Attempt to normalize normalized defs")
-            else:
-                self.defs = defs
-
-    @classmethod
-    def _str_escape(cls, s, pos, toks):
-        raise cls("Error in string escape", None, None, pos, toks)
-
-    @classmethod
-    def _str_end(cls, s, pos, toks):
-        raise cls("Unfinished string", None, None, pos, toks)
-
-    @classmethod
-    def _re_escape(cls, s, pos, toks):
-        raise cls("Error in regex escape", None, None, pos, toks)
-
-    @classmethod
-    def _re_end(cls, s, pos, toks):
-        raise cls("Unfinished regex", None, None, pos, toks)
-
-    @classmethod
-    def _unexpected_tok(cls, s, pos, toks):
-        raise cls("Unexpected token", None, None, pos, toks)
-
-class ValidationError(Exception):
-    """Raised when the object tree doesn't validate against the given pattern."""
-    # FIXME TODO?
-
-class _PatternElement:
-    def on_not_in_key(self, frame, path, defs):
-        raise NotImplementedError
-
-    def on_in_key(self, frame, path, defs):
-        raise NotImplementedError
-
-    def collect_params(self, res: list):
-        pass
-
-class _Arrow(_PatternElement):
-    def on_not_in_key(self, frame, path, defs):
-        assert not path[-1].empty
-        path.append(_Holder(key=None, value=None, name=None, parent=path[-1].value, empty=True))
-        return False
-
-class _StringKey(_PatternElement):
-    def __init__(self, toks):
-        self.key = toks[0]
-        self.skippable = toks[1] == '?'
-
-    def on_in_key(self, frame, path, defs):
-        return self.on_not_in_key(frame, path, defs)
-
-    def on_not_in_key(self, frame, path, defs):
-        path[-1].iterator = self.extract(path[-1].parent)
-        path[-1].empty = False
-        return True
-
-    def extract(self, obj):
-        try:
-            yield (self.key, obj[self.key])
-        except (TypeError, IndexError, KeyError):
-            if not self.skippable:
-                raise ValidationError
-
-class _RegexKey(_PatternElement):
-    def __init__(self, toks):
-        self.key = toks[0]
-        self.compiled = re.compile(self.key)
-        self.skippable = toks[1] == '?'
-
-    def on_in_key(self, frame, path, defs):
-        return self.on_not_in_key(frame, path, defs)
-
-    def on_not_in_key(self, frame, path, defs):
-        filtered_iterator = self.filter(path[-1].iterator)
-        del path[-1].iterator
-        path[-1].iterator = filtered_iterator
-        del filtered_iterator
-        path[-1].empty = False
-        return True
-
-    def filter(self, it):
-        for el in it:
-            try:
-                if self.compiled.search(el[0]):
-                    yield el
-                elif not self.skippable:
-                    raise ValidationError
-            except TypeError:
-                if not self.skippable:
-                    raise ValidationError
-
-class _KeySubtree(_PatternElement):
-    def __init__(self, toks):
-        self.key = toks[0]
-        self.skippable = toks[1] == '?'
-
-    def on_not_in_key(self, frame, path, defs):
-        path[-1].subtree = True
-        filtered_iterator = self.filter(path[-1].iterator, defs)
-        del path[-1].iterator
-        path[-1].iterator = filtered_iterator
-        del filtered_iterator
-        path[-1].empty = False
-        return True
-
-    def filter(self, it, defs):
-        for x in it:
-            for y in _match_helper(self.key, defs, x[0]):
-                yield (y, x[1])
-
-    def collect_params(self, res: list):
-        for sub in self.key:
-            sub.collect_params(res)
-
-class _ValueSubtree(_PatternElement):
-    def __init__(self, toks):
-        self.key = toks[0]
-        self.skippable = toks[1] == '?'
-
-    def on_not_in_key(self, frame, path, defs):
-        assert not path[-1].empty
-        path.append(_Holder(key=None, value=None, name=None, parent=path[-1].value, empty=False, subtree=True))
-        path[-1].iterator = self.filter(path[-1].parent, defs)
-        return True
-
-    def filter(self, parent, defs):
-        for x in _match_helper(self.key, defs, parent):
-            yield (x, parent)
-
-    def collect_params(self, res: list):
-        for sub in self.key:
-            sub.collect_params(res)
-
-class _Ident(_PatternElement):
-    def __init__(self, toks):
-        self.key = toks[0]
-
-    def on_not_in_key(self, frame, path, defs):
-        path[-1].name = self.key
-        path[-1].empty = False
-        return True
-
-class _Param(_PatternElement):
-    def __init__(self, toks):
-        assert isinstance(toks[1], _Ident)
-        self.skippable = toks[0] == '?'
-        self.key = toks[1].key
-
-    def on_in_key(self, frame, path, defs):
-        return self.on_not_in_key(frame, path, defs)
-
-    def on_not_in_key(self, frame, path, defs):
-        path[-1].iterator = self.extract(path[-1].parent, defs[self.key])
-        path[-1].empty = False
-        return True
-
-    def extract(self, obj, key):
-        try:
-            yield (key, obj[key])
-        except (TypeError, IndexError, KeyError):
-            if not self.skippable:
-                raise ValidationError
-
-    def collect_params(self, res: list):
-        res.append(self.key)
-    
-    def get_value(self, defs):
-        return defs[self.key]
-
-class _Ty(_PatternElement):
-    def __init__(self, toks):
-        assert isinstance(toks[1], _Ident)
-        self.skippable = toks[0] == '?'
-        self.key = toks[1].key
-
-    def on_in_key(self, frame, path, defs):
-        filtered_iterator = self.filter(path[-1].iterator, defs[self.key])
-        del path[-1].iterator
-        path[-1].iterator = filtered_iterator
-        del filtered_iterator
-        path[-1].empty = False
-        return True
-
-    def on_not_in_key(self, frame, path, defs):
-        assert len(path) == 1
-        if isinstance(path[-1].value, defs[self.key]):
-            return False
-        elif not self.skippable:
-            raise ValidationError
-        path.clear()
-        return False
-
-    def filter(self, it, ty):
-        for el in it:
-            # this may TypeError if ty is not a type nor a tuple of types
-            # but that's actually the programmer's error
-            if isinstance(el[1], ty):
-                yield el
-            elif not self.skippable:
-                # and this one is for actual validation
-                raise ValidationError
-
-    def collect_params(self, res: list):
-        res.append(self.key)
-
-class _End(_PatternElement):
-    def on_in_key(self, frame, path, defs):
-        try:
-            path[-1].next()
-            return False
-        except StopIteration:
-            path.pop()
-            while frame.prev() and not isinstance(frame.current_op, _End):
-                pass
-            if not frame.prev():
-                # FIXME?
-                path.clear()
-        return True # FIXME?
-
-def _build_syntax():
-    from pyparsing import Suppress, Literal, Forward, CharsNotIn, StringEnd, Combine, Optional, Group, Word, srange, Empty
-    # original regex order: arrow, type/parameter/identifier, string, regex, failure
-    # better syntax: "arrow" changes from "value" to "key" and thus you need at least one key match before an arrow
-    subtree = Forward()
-    # where relevant, enforces match behaviour (skippable object tree branch vs required object tree branch)
-    skippable = Optional("?", default="")
-    #   r"|'(?:%'|%%|%(?P<EES>.|$)|[^%'])*?(?:'|(?P<ES>$))\??" # string literals
-    str_literal = (Combine(Suppress("'")
-            + (Suppress("%") + "'" | Suppress("%") + "%" | Literal("%") + (CharsNotIn("") | StringEnd()).setParseAction(PatternError._str_escape) | CharsNotIn("%'"))[...]
-            + (Suppress("'") | StringEnd().setParseAction(PatternError._str_end))) + skippable).setParseAction(lambda toks: [_StringKey(toks)])
-    #   r"|/(?:%/|%%|%(?P<EER>.|$)|[^%/])*?(?:/|(?P<ER>$))\??" # regex
-    re_literal = (Combine(Suppress("/")
-            + (Suppress("%") + "/" | Suppress("%") + "%" | Literal("%") + (CharsNotIn("") | StringEnd()).setParseAction(PatternError._re_escape) | CharsNotIn("%/"))[...]
-            + (Suppress("/") | StringEnd().setParseAction(PatternError._re_end))) + skippable).setParseAction(lambda toks: [_RegexKey(toks)])
-    arrow = Literal("->").setParseAction(lambda: [_Arrow()])
-    #   r"|(?::\??)?\$?[A-Za-z][A-Za-z0-9]*"                   # identifiers, parameters and type matches
-    identifier = Word(srange("[A-Za-z_]"), srange("[A-Za-z0-9_]")).setParseAction(lambda toks: [_Ident(toks)])
-    parameter = (Suppress("$") + skippable + identifier).setParseAction(lambda toks: [_Param(toks)])
-    ty = (Suppress(":") + skippable + Suppress("$") + identifier).setParseAction(lambda toks: [_Ty(toks)])
-    # support for objects-as-keys
-    keysubtree = (Suppress("[") + Group(ty[...] + subtree) + (Suppress("]") | CharsNotIn("").setParseAction(PatternError._unexpected_tok) | StringEnd().setParseAction(PatternError._unexpected_tok)) + Optional("?", default="")).setParseAction(lambda toks: [_KeySubtree(toks)])
-    # represents key matching - switches from "key" to "value"
-    tag = (identifier + Optional(parameter | str_literal | re_literal | keysubtree) | parameter | str_literal | re_literal | keysubtree) + ty[...] + Empty().setParseAction(lambda: [_End()])
-    # multiple value matching
-    valuesubtree = (Suppress("(") + Group(subtree) + (Suppress(")") | CharsNotIn("").setParseAction(PatternError._unexpected_tok) | StringEnd().setParseAction(PatternError._unexpected_tok)) + Optional("?", default="")).setParseAction(lambda toks: [_ValueSubtree(toks)])
-    # arrow and tag, value subtree
-    subtree <<= (arrow + tag)[...] + (valuesubtree + Empty().setParseAction(lambda: [_End()]))[...]
-    return ((subtree | CharsNotIn("").setParseAction(PatternError._unexpected_tok)) + StringEnd()).parseWithTabs()
-
-_built_syntax = _build_syntax()
-
-def _pairs(o):
-    if isinstance(o, Mapping):
-        return iter(o.items())
-    elif isinstance(o, Sequence):
-        return iter(enumerate(o, 0))
-    elif isinstance(o, Set):
-        return iter(((e, e) for e in o))
-    else:
-        # maybe there's more stuff I can implement later
-        raise TypeError
-
-class _Holder:
-    def __init__(self, key, value, name, parent=None, it=None, empty=False, subtree=False):
-        self.name = name
-        self.key = key
-        self.value = value
-        self.empty = empty
-        self._it = it
-        self.parent = parent
-        self.subtree = subtree
-
-    @property
-    def iterator(self):
-        if self._it is None:
-            self._it = _pairs(self.parent)
-        return self._it
-
-    @iterator.setter
-    def iterator(self, value):
-        assert self._it is None
-        self._it = value
-
-    @iterator.deleter
-    def iterator(self):
-        self._it = None
-
-    def next(self):
-        self.key, self.value = next(self.iterator)
-
-class _Frame:
-    def __init__(self, ops):
-        self.ops = ops
-        self.pc = -1
-
-    def next(self):
-        pc = self.pc + 1
-        if pc >= len(self.ops):
-            return False
-        self.pc = pc
-        return True
-
-    @property
-    def current_op(self):
-        return self.ops[self.pc]
-
-    def prev(self):
-        pc = self.pc - 1
-        if pc < 0:
-            return False
-        self.pc = pc
-        return True
-
-def _match_helper(ops, defs, tree):
-    frame = _Frame(ops)
-
-    path = [_Holder(key=None, value=tree, parent=None, it=iter(()), name=None)]
-    in_key = False
-    while path:
-        if not frame.next():
-            assert not path[-1].empty
-            res = {}
-            for h in path:
-                if h.subtree:
-                    for name, kv in h.key.items():
-                        res[name] = kv
-                elif h.name is not None:
-                    res[h.name] = (h.key, h.value)
-            yield res
-            assert len(path) == 1 or isinstance(frame.current_op, _End)
-            frame.prev()
-            in_key = True
-        else:
-            op = frame.current_op
-            if in_key:
-                in_key = op.on_in_key(frame, path, defs)
-            else:
-                in_key = op.on_not_in_key(frame, path, defs)
-
-class Pattern:
-    """A compiled pattern object.
-
-    Warning:
-        Do not create instances of this class manually. Use :py:func:`abdl.compile`.
-
-    """
-
-    def __init__(self, pattern, defs):
-        try:
-            self._ops = _built_syntax.parseString(pattern)
-        except PatternError as e:
-            e._normalize(pattern, defs)
-            raise
-        else:
-            self._params = []
-            for op in self._ops:
-                op.collect_params(self._params)
-            self._defs = {param: defs[param] for param in self._params}
-
-    def match(self, obj):
-        """Matches this compiled pattern against the given object.
-
-        Args:
-            obj: The object to match against.
-
-        Returns:
-            An iterator. This iterator yields ``(key, value)`` pairs
-            wrapped in a dict for each variable in the pattern.
-
-        """
-        return _match_helper(self._ops, self._defs, obj)
-
-def compile(pattern, defs={}):
-    """Compiles the pattern and returns a compiled :py:class:`abdl.Pattern` object.
-
-    Args:
-        pattern (str): The pattern. Refer to module-level documentation for
-            pattern syntax.
-        defs (dict): The parameter list. Used by parameters in the pattern.
-
-    Returns:
-        Pattern: A compiled pattern object.
-
-    """
-    # TODO caching
-    return Pattern(pattern, defs)
-
-def match(pattern, obj, defs={}):
-    """Matches the pattern against the given obj.
-
-    This method is equivalent to ``abdl.compile(pattern, defs).match(obj)``.
-
-    Args:
-        pattern (str): The pattern. Refer to module-level documentation for
-            pattern syntax.
-        obj: The object to match against.
-        defs (dict): The parameter list. Used by parameters in the pattern.
-
-    Returns:
-        An iterator. This iterator yields ``(key, value)`` pairs
-        wrapped in a dict for each variable in the pattern.
-
-    """
-    return compile(pattern, defs).match(obj)
diff --git a/ganarchy/__init__.py b/ganarchy/__init__.py
index 0f13b44..6d24994 100644
--- a/ganarchy/__init__.py
+++ b/ganarchy/__init__.py
@@ -32,7 +32,7 @@ import requests
 from collections import defaultdict
 from urllib.parse import urlparse
 
-import ganarchy.config
+import ganarchy.config as m_ganarchy_config
 
 MIGRATIONS = {
         "toml-config": (
@@ -210,10 +210,8 @@ def get_env():
     env.filters['tomle'] = env.filters['tomlescape']
     return env
 
-
-@click.group()
-def ganarchy():
-    pass
+# FIXME
+from ganarchy.cli import main as ganarchy
 
 @ganarchy.command()
 def initdb():
@@ -457,12 +455,11 @@ class Config:
         # TODO re.compile("(^" + "|^".join(map(re.escape, domains)) + "|" + "|".join(map(re.escape, suffixes) + ")$")
         if base:
             # FIXME is remove=remove the right thing to do?
-            self._update_projects(base.projects, remove=remove, sanitize=False) # already sanitized
-        projects = config_data.get('projects', {})
-        self._update_projects(projects, remove=remove)
+            self._update_projects({'projects': base.projects}, remove=remove, sanitize=False) # already sanitized
+        self._update_projects(config_data, remove=remove)
 
     def _update_projects(self, projects, remove, sanitize=True):
-        m = (ganarchy.config.CONFIG_PATTERN_SANITIZE if sanitize else ganarchy.config.CONFIG_PATTERN).match(projects)
+        m = (m_ganarchy_config.CONFIG_REPOS_SANITIZE if sanitize else m_ganarchy_config.CONFIG_REPOS).match(projects)
         for v in m:
             commit, repo_url, branchname, options = v['commit'][0], v['url'][0], v['branch'][0], v['branch'][1]
             try:
diff --git a/ganarchy/__main__.py b/ganarchy/__main__.py
index b270251..ca17cb7 100644
--- a/ganarchy/__main__.py
+++ b/ganarchy/__main__.py
@@ -15,9 +15,12 @@
 # along with this program.  If not, see <https://www.gnu.org/licenses/>.
 
 # The base CLI
+import ganarchy.cli
+
+# FIXME this shouldn't be here
 import ganarchy
 
 # Additional CLI commands
-import ganarchy.debug
+import ganarchy.cli.debug
 
-ganarchy.ganarchy(prog_name='ganarchy')
+ganarchy.cli.main(prog_name='ganarchy')
diff --git a/ganarchy/debug.py b/ganarchy/cli/__init__.py
index 1310549..9effabb 100644
--- a/ganarchy/debug.py
+++ b/ganarchy/cli/__init__.py
@@ -16,21 +16,6 @@
 
 import click
 
-import ganarchy
-import ganarchy.config
-
-@ganarchy.ganarchy.group()
-def debug():
+@click.group()
+def main():
     pass
-
-@debug.command()
-def paths():
-    click.echo('Config home: {}'.format(ganarchy.config_home))
-    click.echo('Additional config search path: {}'.format(ganarchy.config_dirs))
-    click.echo('Cache home: {}'.format(ganarchy.cache_home))
-    click.echo('Data home: {}'.format(ganarchy.data_home))
-
-@debug.command()
-def configs():
-    pass
-
diff --git a/ganarchy/cli/debug.py b/ganarchy/cli/debug.py
new file mode 100644
index 0000000..10e91e5
--- /dev/null
+++ b/ganarchy/cli/debug.py
@@ -0,0 +1,70 @@
+# This file is part of GAnarchy - decentralized project hub
+# Copyright (C) 2019  Soni L.
+#
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU Affero General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU Affero General Public License for more details.
+#
+# You should have received a copy of the GNU Affero General Public License
+# along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+import click
+import qtoml
+
+import ganarchy
+import ganarchy.cli
+import ganarchy.config
+
+@ganarchy.cli.main.group()
+def debug():
+    pass
+
+@debug.command()
+def paths():
+    click.echo('Config home: {}'.format(ganarchy.config_home))
+    click.echo('Additional config search path: {}'.format(ganarchy.config_dirs))
+    click.echo('Cache home: {}'.format(ganarchy.cache_home))
+    click.echo('Data home: {}'.format(ganarchy.data_home))
+
+@debug.command()
+def configs():
+    def print_conf(conf):
+        click.echo("\tRepos:")
+        for i, repo in enumerate(ganarchy.config.CONFIG_REPOS.match({'projects': conf.projects})):
+            click.echo("\t\t{}.".format(i))
+            click.echo("\t\t\tProject: {}".format(repo['commit'][0]))
+            click.echo("\t\t\tURI: {}".format(repo['url'][0]))
+            click.echo("\t\t\tBranch: {}".format(repo['branch'][0]))
+            click.echo("\t\t\tActive: {}".format(repo['branch'][1] == {'active': True}))
+
+    click.echo("Breaking down the configs.")
+    conf = None
+    # reverse order is intentional
+    for d in reversed(ganarchy.config_dirs):
+        click.echo("Config: {}/config.toml".format(d))
+        try:
+            f = open(d + "/config.toml", 'r', encoding='utf-8', newline='')
+            conf = ganarchy.Config(f, conf)
+            click.echo("Updated entries:")
+            print_conf(conf)
+            f.close()
+        except (OSError, UnicodeDecodeError, qtoml.decoder.TOMLDecodeError) as e:
+            click.echo("\tError: {}".format(e))
+    try:
+        click.echo("Config: {}/config.toml".format(ganarchy.config_home))
+        f = open(ganarchy.config_home + "/config.toml", 'r', encoding='utf-8', newline='')
+        conf = ganarchy.Config(f, conf)
+        click.echo("Updated entries:")
+        print_conf(conf)
+        click.echo("-----")
+        click.echo("\tTitle: {}".format(conf.base_url))
+        click.echo("\tBase URI: {}".format(conf.base_url))
+        f.close()
+    except (OSError, UnicodeDecodeError, qtoml.decoder.TOMLDecodeError) as e:
+        click.echo("\tError: {}".format(e))
diff --git a/ganarchy/config.py b/ganarchy/config.py
index 154447b..ae1615d 100644
--- a/ganarchy/config.py
+++ b/ganarchy/config.py
@@ -155,3 +155,55 @@ class RemoteConfigSource(ConfigSource):
         for r in CONFIG_PATTERN_SANITIZE.match(self.tomlobj):
             yield (r['commit'][0], r['url'][0], r['branch'][0], r['branch'][1])
 
+class ConfigManager:
+    def __init__(self):
+        # FIXME ???
+        self.sources = []
+
+# class Config:
+#     def __init__(self, toml_file, base=None, remove=True):
+#         self.projects = defaultdict(lambda: defaultdict(lambda: defaultdict(lambda: defaultdict(dict))))
+#         config_data = qtoml.load(toml_file)
+#         self.remote_configs = config_data.get('config_srcs', [])
+#         self.title = config_data.get('title', '')
+#         self.base_url = config_data.get('base_url', '')
+#         # TODO blocked domains (but only read them from config_data if remove is True)
+#         self.blocked_domains = []
+#         self.blocked_domain_suffixes = []
+#         self.blocked_domains.sort()
+#         self.blocked_domain_suffixes.sort(key=lambda x: x[::-1])
+#         # FIXME remove duplicates and process invalid entries
+#         self.blocked_domains = tuple(self.blocked_domains)
+#         self.blocked_domain_suffixes = tuple(self.blocked_domain_suffixes) # MUST be tuple
+#         # TODO re.compile("(^" + "|^".join(map(re.escape, domains)) + "|" + "|".join(map(re.escape, suffixes) + ")$")
+#         if base:
+#             # FIXME is remove=remove the right thing to do?
+#             self._update_projects(base.projects, remove=remove, sanitize=False) # already sanitized
+#         projects = config_data.get('projects', {})
+#         self._update_projects(projects, remove=remove)
+# 
+#     def _update_projects(self, projects, remove, sanitize=True):
+#         m = (m_ganarchy_config.CONFIG_PATTERN_SANITIZE if sanitize else m_ganarchy_config.CONFIG_PATTERN).match(projects)
+#         for v in m:
+#             commit, repo_url, branchname, options = v['commit'][0], v['url'][0], v['branch'][0], v['branch'][1]
+#             try:
+#                 u = urlparse(repo_url)
+#                 if not u:
+#                     raise ValueError
+#                 # also raises for invalid ports, see https://docs.python.org/3/library/urllib.parse.html#urllib.parse.urlparse
+#                 # "Reading the port attribute will raise a ValueError if an invalid port is specified in the URL. [...]"
+#                 if u.port == 0:
+#                     raise ValueError
+#                 if u.scheme not in ('http', 'https'):
+#                     raise ValueError
+#                 if (u.hostname in self.blocked_domains) or (u.hostname.endswith(self.blocked_domain_suffixes)):
+#                     raise ValueError
+#             except ValueError:
+#                 continue
+#             if branchname == "HEAD":
+#                 branchname = None
+#             active = options.get('active', None)
+#             if active not in (True, False):
+#                 continue
+#             branch = self.projects[commit][repo_url][branchname]
+#             branch['active'] = active or (branch.get('active', False) and not remove)
diff --git a/requirements.txt b/requirements.txt
index 2f964c0..0fd42f6 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,5 +1,11 @@
+-e git+https://soniex2.autistic.space/git-repos/abdl.git@1b26ad799217af7e187fdae78e862a6bf46e5591#egg=gan0f74bd87a23b515b45da7e6f5d9cc82380443dab
+certifi==2019.11.28
+chardet==3.0.4
 Click==7.0
+idna==2.8
 Jinja2==2.10.1
-qtoml==0.2.4
+MarkupSafe==1.1.1
 pyparsing==2.4.2
+qtoml==0.2.4
 requests==2.22.0
+urllib3==1.25.7
diff --git a/requirements_test.txt b/requirements_test.txt
index 737d0df..ebe51db 100644
--- a/requirements_test.txt
+++ b/requirements_test.txt
@@ -1,3 +1,21 @@
+astroid==2.3.3
+atomicwrites==1.3.0
+attrs==19.3.0
+decompyle3==3.3.2
 hypothesis==4.42.7
+isort==4.3.21
+lazy-object-proxy==1.4.3
+mccabe==0.6.1
+more-itertools==7.2.0
+packaging==19.2
+pluggy==0.13.1
+py==1.8.0
+pylint==2.4.4
 pytest==5.2.2
 pytest-sphinx==0.2.2
+six==1.13.0
+spark-parser==1.8.9
+uncompyle6==3.6.2
+wcwidth==0.1.7
+wrapt==1.11.2
+xdis==4.2.2
diff --git a/setup.py b/setup.py
new file mode 100644
index 0000000..a14854e
--- /dev/null
+++ b/setup.py
@@ -0,0 +1,5 @@
+import setuptools
+
+setuptools.setup(packages=["gan385e734a52e13949a7a5c71827f6de920dbfea43"], install_requires=[
+    "gan0f74bd87a23b515b45da7e6f5d9cc82380443dab",  # a boneless datastructure library
+    "Click", "Jinja2", "qtoml", "requests"])
diff --git a/testing/test_abdl.py b/testing/test_abdl.py
deleted file mode 100644
index 4f5241f..0000000
--- a/testing/test_abdl.py
+++ /dev/null
@@ -1,199 +0,0 @@
-# Tests abdl.py
-
-import abdl
-
-import hypothesis
-import hypothesis.strategies as st
-
-import collections.abc
-
-import re
-
-import traceback
-
-
-abdl.DeprecationError.enable_key_match_compat = False
-
-# use abdl's _pairs for consistency.
-pairs = abdl._pairs
-
-# do not put integers, floats, etc here
-# do not put bytes, they iterate as integers
-hashables = st.deferred(lambda: st.text() | st.frozensets(hashables) | st.lists(hashables).map(tuple))
-values = st.deferred(lambda: hashables | objtree)
-objtree = st.deferred(lambda: st.text() | st.dictionaries(hashables, values) | st.lists(values) | st.sets(hashables) | st.lists(hashables).map(tuple))
-
-# note: use all() so as to not eat all the RAM :p
-
-class LogAndCompare:
-    def __init__(self, left, right):
-        self._itl = left
-        self._itr = right
-        self.left = []
-        self.right = []
-    def __iter__(self):
-        return self
-    def __next__(self):
-        try:
-            left = next(self._itl)
-        except abdl.ValidationError as e:
-            e.tb = traceback.format_exc()
-            left = e
-        try:
-            right = next(self._itr)
-        except abdl.ValidationError as e:
-            e.tb = traceback.format_exc()
-            right = e
-        self.left.append(left)
-        self.right.append(right)
-        return left == right or (type(left), type(right)) == (abdl.ValidationError,)*2
-    def __repr__(self):
-        return "LogAndCompare(left=" + repr(self.left) + ", right=" + repr(self.right) + ")"
-
-
-@hypothesis.given(objtree, st.just(abdl.compile("->X")))
-def test_basic_iterator(foo, pat):
-    assert all(LogAndCompare(pat.match(foo), map(lambda x: {"X": x}, pairs(foo))))
-
-@hypothesis.given(objtree, st.just(abdl.compile("->X->Y")))
-def test_two_depths(foo, pat):
-    def deep(foo):
-        for x in pairs(foo):
-            for y in pairs(x[1]):
-                yield {"X": x, "Y": y}
-    assert all(LogAndCompare(pat.match(foo), deep(foo)))
-
-@hypothesis.given(objtree, st.just(abdl.compile("->X->Y->Z->W")))
-def test_four_depths(foo, pat):
-    def deep(foo):
-        for x in pairs(foo):
-            for y in pairs(x[1]):
-                for z in pairs(y[1]):
-                    for w in pairs(z[1]):
-                        yield {"X": x, "Y": y, "Z": z, "W": w}
-    assert all(LogAndCompare(pat.match(foo), deep(foo)))
-
-@hypothesis.given(st.dictionaries(st.text(), st.text()) | st.sets(st.text()), st.just(abdl.compile("->/.../")))
-def test_regex(foo, pat):
-    # no bindings on this one :<
-    def deep(foo):
-        for x in pairs(foo):
-            if re.search("...", x[0]):
-                    yield {}
-            else:
-                raise abdl.ValidationError
-    assert all(LogAndCompare(pat.match(foo), deep(foo)))
-
-@hypothesis.given(objtree, st.just(abdl.compile("->/.../?")))
-def test_regex_skippable_vs_objtree(foo, pat):
-    assert all(LogAndCompare(pat.match(foo), ({} for x in pairs(foo) if isinstance(x[0], str) and re.search("...", x[0]))))
-
-@hypothesis.given(st.dictionaries(st.text(), st.text()) | st.sets(st.text()), st.just(abdl.compile("->/.../->Y")))
-def test_regex_and_bind(foo, pat):
-    def deep(foo):
-        for x in pairs(foo):
-            if re.search("...", x[0]):
-                for y in pairs(x[1]):
-                    yield {"Y": y}
-            else:
-                raise abdl.ValidationError
-    assert all(LogAndCompare(pat.match(foo), deep(foo)))
-
-@hypothesis.given(objtree, st.just(abdl.compile("->/.../?->Y")))
-def test_regex_skippable_and_bind_vs_objtree(foo, pat):
-    def deep(foo):
-        for x in pairs(foo):
-            if isinstance(x[0], str) and re.search("...", x[0]):
-                for y in pairs(x[1]):
-                    yield {"Y": y}
-    assert all(LogAndCompare(pat.match(foo), deep(foo)))
-
-@hypothesis.given(objtree, st.just(abdl.compile("->/^...$/?->Y")))
-def test_regex_anchored_skippable_and_bind_vs_objtree(foo, pat):
-    def deep(foo):
-        for x in pairs(foo):
-            if isinstance(x[0], str) and re.search("^...$", x[0]):
-                for y in pairs(x[1]):
-                    yield {"Y": y}
-    assert all(LogAndCompare(pat.match(foo), deep(foo)))
-
-@hypothesis.given(objtree, st.just(abdl.compile("->''?->Y")))
-def test_empty_literal_vs_objtree(foo, pat):
-    def deep(foo):
-        for x in pairs(foo):
-            if x[0] == '':
-                for y in pairs(x[1]):
-                    yield {"Y": y}
-    assert all(LogAndCompare(pat.match(foo), deep(foo)))
-
-defs = {'a': (dict, list, set)}
-@hypothesis.given(objtree, st.just(abdl.compile("->X:?$a->Y", defs=defs)))
-def test_type(foo, pat):
-    def deep(foo):
-        for x in pairs(foo):
-            if isinstance(x[1], (dict, list, set)):
-                for y in pairs(x[1]):
-                    yield {"X": x, "Y": y}
-    assert all(LogAndCompare(pat.match(foo), deep(foo)))
-
-defs = {'a': (dict, list, set), 'b': (dict, set), 'c': dict}
-@hypothesis.given(objtree, st.just(abdl.compile("->X:?$a:?$b:?$c->Y", defs=defs)))
-def test_multi_type(foo, pat):
-    def deep(foo):
-        for x in pairs(foo):
-            if isinstance(x[1], dict):
-                for y in pairs(x[1]):
-                    yield {"X": x, "Y": y}
-    assert all(LogAndCompare(pat.match(foo), deep(foo)))
-
-defs = {'a': (dict, list, set), 'b': (dict, set), 'c': dict}
-@hypothesis.given(objtree, st.just(abdl.compile("->X:$a:$b:$c->Y", defs=defs)))
-@hypothesis.settings(suppress_health_check=[hypothesis.HealthCheck.too_slow])
-def test_multi_type_with_validation_errors(foo, pat):
-    def deep(foo):
-        for x in pairs(foo):
-            if isinstance(x[1], dict):
-                for y in pairs(x[1]):
-                    yield {"X": x, "Y": y}
-            else:
-                raise abdl.ValidationError
-    assert all(LogAndCompare(pat.match(foo), deep(foo)))
-
-@hypothesis.given(st.dictionaries(st.frozensets(st.text()), st.text()), st.just(abdl.compile("->[:?$sets->A]->D", {'sets': collections.abc.Set})))
-def test_subtree_partial(foo, pat):
-    def deep(foo):
-        for x in pairs(foo):
-            if isinstance(x[0], collections.abc.Set):
-                for a in pairs(x[0]):
-                    for d in pairs(x[1]):
-                        yield {"A": a, "D": d}
-    assert all(LogAndCompare(pat.match(foo), deep(foo)))
-
-@hypothesis.given(objtree, st.just(abdl.compile("->X->$a->Z", {'a': '0'})))
-def test_param(foo, pat):
-    def deep(foo):
-        for x in pairs(foo):
-            try:
-                y = x['0']
-            except (TypeError, IndexError, KeyError):
-                raise abdl.ValidationError
-            else:
-                for z in pairs(y):
-                    yield {"X": x, "Z": z}
-    assert all(LogAndCompare(pat.match(foo), deep(foo)))
-
-def test_basic_value_subtree():
-    matcher = abdl.match("(->foo'foo')(->bar'bar')", {'foo': 1, 'bar': 2})
-    assert list(matcher) == [{'foo': ('foo', 1), 'bar': ('bar', 2)}]
-
-# FIXME
-#@hypothesis.given(objtree, st.text())
-#def test_exhaustive(foo, pat):
-#    hypothesis.assume(not re.match("^%s+$", pat))
-#    hypothesis.assume(pat)
-#    try:
-#        compiled = abdl.compile(pat)
-#        print(pat)
-#    except abdl.PatternError:
-#        hypothesis.assume(False)
-#    compiled.match(foo)
diff --git a/testing/test_examples.py b/testing/test_examples.py
deleted file mode 100644
index 02532cc..0000000
--- a/testing/test_examples.py
+++ /dev/null
@@ -1,26 +0,0 @@
-import abdl
-
-def test_basic_example():
-    m = next(abdl.match("->X:?$dict->Y", {"foo": 1, "bar": {"baz": 2}}, {'dict': dict}))
-    assert m['X'][0] == 'bar'
-    assert m['Y'][0] == 'baz'
-    assert m['Y'][1] == 2
-
-def test_basic_2():
-    m = next(abdl.match("->'projects':?$d->P/[0-9a-fA-F]{40}|[0-9a-fA-F]{64}/?:?$d->U:?$d->B", {"projects": {"385e734a52e13949a7a5c71827f6de920dbfea43": {"https://soniex2.autistic.space/git-repos/ganarchy.git": {"HEAD": {"active": True}}}}}, {'d': dict}))
-    assert m['P'][0] == "385e734a52e13949a7a5c71827f6de920dbfea43"
-    assert m['U'][0] == "https://soniex2.autistic.space/git-repos/ganarchy.git"
-    assert m['B'][0] == "HEAD"
-    assert m['B'][1] == {"active": True}
-
-def test_spaces():
-    pat = abdl.compile("""-> 'projects'?
-                             -> commit /[0-9a-fA-F]{40}|[0-9a-fA-F]{64}/? :?$dict
-                                -> url :?$dict
-                                   -> branch :?$dict""", {'dict': dict})
-    data = {"projects": {"385e734a52e13949a7a5c71827f6de920dbfea43": {"https://soniex2.autistic.space/git-repos/ganarchy.git": {"HEAD": {"active": True}}}}}
-    m = next(pat.match(data))
-    assert m['commit'][0] == "385e734a52e13949a7a5c71827f6de920dbfea43"
-    assert m['url'][0] == "https://soniex2.autistic.space/git-repos/ganarchy.git"
-    assert m['branch'][0] == "HEAD"
-    assert m['branch'][1] == {"active": True}
diff --git a/testing/test_ops_abdl.py b/testing/test_ops_abdl.py
deleted file mode 100644
index fe50369..0000000
--- a/testing/test_ops_abdl.py
+++ /dev/null
@@ -1,161 +0,0 @@
-# Tests abdl.py internals
-
-import abdl
-
-import re
-
-class OpHelper:
-    def __init__(self, pat, ops=None):
-        self.pat = pat
-        if not ops:
-            self.ops = pat._ops
-        else:
-            self.ops = ops
-        self.pos = -1
-
-    def done(self):
-        assert self.pos + 1 == len(self.ops)
-
-    def __enter__(self):
-        self.pos += 1
-        first = self.pos
-        assert not isinstance(self.ops[first], abdl._End)
-        while not isinstance(self.ops[self.pos], abdl._End):
-            self.pos += 1
-        assert isinstance(self.ops[self.pos], abdl._End)
-        return self.ops[first:self.pos]
-
-    def __exit__(self, exc_type, exc_value, traceback):
-        pass
-
-def expect_types(seq, *tys):
-    assert len(seq) == len(tys)
-    assert(all(map(lambda x: isinstance(*x), zip(seq, tys))))
-
-def expect_idents(oph, *idents):
-    for ident in idents:
-        with oph as ops:
-            expect_types(ops, abdl._Arrow, abdl._Ident)
-            assert ops[1].key == ident
-
-def test_empty_iterator_pattern():
-    oph = OpHelper(abdl.compile(""))
-    oph.done()
-
-def test_four_depths_pattern():
-    oph = OpHelper(abdl.compile("->X->Y->Z->W"))
-    expect_idents(oph, "X", "Y", "Z", "W")
-    oph.done()
-
-def test_regex_pattern():
-    oph = OpHelper(abdl.compile("->/.../"))
-    with oph as ops:
-        expect_types(ops, abdl._Arrow, abdl._RegexKey)
-        assert ops[1].key == '...'
-        assert ops[1].compiled == re.compile('...')
-        assert ops[1].skippable == False
-    oph.done()
-
-def test_regex_skippable_pattern():
-    oph = OpHelper(abdl.compile("->/.../?"))
-    with oph as ops:
-        expect_types(ops, abdl._Arrow, abdl._RegexKey)
-        assert ops[1].key == '...'
-        assert ops[1].compiled == re.compile('...')
-        assert ops[1].skippable == True
-    oph.done()
-
-def test_regex_and_bind_pattern():
-    oph = OpHelper(abdl.compile("->/.../->Y"))
-    with oph as ops:
-        expect_types(ops, abdl._Arrow, abdl._RegexKey)
-        assert ops[1].key == '...'
-        assert ops[1].compiled == re.compile('...')
-        assert ops[1].skippable == False
-    expect_idents(oph, "Y")
-    oph.done()
-
-def test_empty_literal_skippable_and_bind_pattern():
-    oph = OpHelper(abdl.compile("->''?->Y"))
-    with oph as ops:
-        expect_types(ops, abdl._Arrow, abdl._StringKey)
-        assert ops[1].key == ''
-        assert ops[1].skippable == True
-    expect_idents(oph, "Y")
-    oph.done()
-
-def test_type_pattern():
-    oph = OpHelper(abdl.compile("->X:?$a->Y", defs={'a': (dict, list, set)}))
-    assert oph.pat._defs['a'] == (dict, list, set)
-    with oph as ops:
-        expect_types(ops, abdl._Arrow, abdl._Ident, abdl._Ty)
-        assert ops[1].key == 'X'
-        assert ops[2].key == 'a'
-        assert ops[2].skippable == True
-    expect_idents(oph, "Y")
-    oph.done()
-
-def test_multi_type_pattern():
-    oph = OpHelper(abdl.compile("->X:$a:?$b:?$c->Y", defs={'a': (dict, list, set), 'b': (dict, set), 'c': dict}))
-    assert oph.pat._defs['a'] == (dict, list, set)
-    assert oph.pat._defs['b'] == (dict, set)
-    assert oph.pat._defs['c'] == dict
-    with oph as ops:
-        expect_types(ops, abdl._Arrow, abdl._Ident, abdl._Ty, abdl._Ty, abdl._Ty)
-        assert ops[1].key == 'X'
-        assert ops[2].key == 'a'
-        assert ops[2].skippable == False
-        assert ops[3].key == 'b'
-        assert ops[3].skippable == True
-        assert ops[4].key == 'c'
-        assert ops[4].skippable == True
-    expect_idents(oph, "Y")
-    oph.done()
-
-def test_key_subtree_pattern():
-    oph = OpHelper(abdl.compile("->[:?$set->A]->D", defs={'set': set}))
-    assert oph.pat._defs['set'] == set
-    with oph as ops:
-        expect_types(ops, abdl._Arrow, abdl._KeySubtree)
-        oph2 = OpHelper(None, ops=ops[1].key)
-        with oph2 as ops2:
-            expect_types(ops2, abdl._Ty, abdl._Arrow, abdl._Ident)
-            assert ops2[0].key == 'set'
-            assert ops2[0].skippable == True
-            assert ops2[2].key == 'A'
-        oph2.done()
-    expect_idents(oph, "D")
-    oph.done()
-
-def test_param_pattern():
-    oph = OpHelper(abdl.compile("->X->$a->Z", defs={'a': '0'}))
-    assert oph.pat._defs['a'] == '0'
-    expect_idents(oph, "X")
-    with oph as ops:
-        expect_types(ops, abdl._Arrow, abdl._Param)
-        assert ops[1].key == 'a'
-        assert ops[1].skippable == False
-    expect_idents(oph, "Z")
-    oph.done()
-
-def test_value_subtree_pattern():
-    oph = OpHelper(abdl.compile("(->foo'foo')(->bar'bar')"))
-    with oph as ops:
-        expect_types(ops, abdl._ValueSubtree)
-        oph2 = OpHelper(None, ops=ops[0].key)
-        with oph2 as ops2:
-            expect_types(ops2, abdl._Arrow, abdl._Ident, abdl._StringKey)
-            assert ops2[1].key == 'foo'
-            assert ops2[2].key == 'foo'
-            assert ops2[2].skippable == False
-        oph2.done()
-    with oph as ops:
-        expect_types(ops, abdl._ValueSubtree)
-        oph2 = OpHelper(None, ops=ops[0].key)
-        with oph2 as ops2:
-            expect_types(ops2, abdl._Arrow, abdl._Ident, abdl._StringKey)
-            assert ops2[1].key == 'bar'
-            assert ops2[2].key == 'bar'
-            assert ops2[2].skippable == False
-        oph2.done()
-    oph.done()