v0.0.14

add :comment command for toggling comments

.gitignore

@@ -5,4 +5,6 @@
 /.mypy_cache
 /.pytest_cache
 /.tox
+/build
+/dist
 /venv*

.pre-commit-config.yaml

@@ -11,29 +11,33 @@ repos:
     -   id: name-tests-test
     -   id: requirements-txt-fixer
 -   repo: https://gitlab.com/pycqa/flake8
-    rev: 3.7.9
+    rev: 3.8.0
     hooks:
     -   id: flake8
         additional_dependencies: [flake8-typing-imports==1.7.0]
 -   repo: https://github.com/pre-commit/mirrors-autopep8
-    rev: v1.5
+    rev: v1.5.2
     hooks:
     -   id: autopep8
 -   repo: https://github.com/asottile/reorder_python_imports
-    rev: v1.9.0
+    rev: v2.3.0
     hooks:
     -   id: reorder-python-imports
         args: [--py3-plus]
 -   repo: https://github.com/asottile/add-trailing-comma
-    rev: v1.5.0
+    rev: v2.0.1
     hooks:
     -   id: add-trailing-comma
         args: [--py36-plus]
 -   repo: https://github.com/asottile/pyupgrade
-    rev: v2.1.0
+    rev: v2.4.1
     hooks:
     -   id: pyupgrade
         args: [--py36-plus]
+-   repo: https://github.com/asottile/setup-cfg-fmt
+    rev: v1.9.0
+    hooks:
+    -   id: setup-cfg-fmt
 -   repo: https://github.com/pre-commit/mirrors-mypy
     rev: v0.770
     hooks:

README.md

@@ -1,14 +1,20 @@
 [![Build Status](https://dev.azure.com/asottile/asottile/_apis/build/status/asottile.babi?branchName=master)](https://dev.azure.com/asottile/asottile/_build/latest?definitionId=29&branchName=master)
 [![Azure DevOps coverage](https://img.shields.io/azure-devops/coverage/asottile/asottile/29/master.svg)](https://dev.azure.com/asottile/asottile/_build/latest?definitionId=29&branchName=master)
 
+![babi logo](https://user-images.githubusercontent.com/1810591/89981369-9ed84e80-dc28-11ea-9708-5f4c49c09632.png)
+
 babi
 ====
 
 a text editor, eventually...
 
+### installation
+
+`pip install babi`
+
 ### why is it called babi?
 
-I usually use the text editor `nano`, frequently I typo this.  on a qwerty
+I used to use the text editor `nano`, frequently I typo this.  on a qwerty
 keyboard, when the right hand is shifted left by one, `nano` becomes `babi`.
 
 ### quitting babi
@@ -23,6 +29,7 @@ these are all of the current key bindings in babi
 - <kbd>^S</kbd>: save
 - <kbd>^O</kbd>: save as
 - <kbd>^X</kbd>: quit
+- <kbd>^P</kbd>: open file
 - arrow keys: movement
 - <kbd>^A</kbd> / <kbd>home</kbd>: move to beginning of line
 - <kbd>^E</kbd> / <kbd>end</kbd>: move to end of line
@@ -40,7 +47,7 @@ these are all of the current key bindings in babi
 - <kbd>tab</kbd> / <kbd>shift-tab</kbd>: indent or dedent current line (or
   selection)
 - <kbd>^K</kbd> / <kbd>^U</kbd>: cut and uncut the current line (or selection)
-- <kbd>M-u</kbd> / <kbd>M-U</kbd>: undo / redo
+- <kbd>M-u</kbd> / <kbd>M-U</kbd> or <kbd>M-e</kbd>: undo / redo
 - <kbd>^W</kbd>: search
 - <kbd>^\\</kbd>: search and replace
 - <kbd>^C</kbd>: show the current position in the file
@@ -62,10 +69,15 @@ in prompts (search, search replace, command):
 
 the syntax highlighting setup is a bit manual right now
 
-1. from a clone of babi, run `./bin/download-syntax` -- you will likely need
+1. find a visual studio code theme, convert it to json (if it is not already
-   to install some additional packages to download them (`pip install cson`)
+   json) and put it at `~/.config/babi/theme.json`.  a helper script is
-2. find a visual studio code theme, convert it to json (if it is not already
+   provided to make this easier: `./bin/download-theme NAME URL`
-   json) and put it at `~/.config/babi/theme.json`
+
+here's a modified vs dark plus theme that works:
+
+```bash
+./bin/download-theme vs-dark-asottile https://gist.github.com/asottile/b465856c82b1aaa4ba8c7c6314a72e13/raw/22d602fb355fb12b04f176a733941ba5713bc36c/vs_dark_asottile.json
+```
 
 ## demos
 
@@ -77,7 +89,7 @@ this opens the file, displays it, and can be edited and can save! unknown keys
 are displayed as errors in the status bar.  babi will scroll if the cursor
 goes off screen either from resize events or from movement.  babi can edit
 multiple files.  babi has a command mode (so you can quit it like vim
-<kbd>:q</kbd>!).  babi also support syntax highlighting
+<kbd>:q</kbd>!).  babi also supports syntax highlighting
 
 ![](https://i.fluffy.cc/5WFZBJ4mWs7wtThD9strQnGlJqw4Z9KS.png)
 

azure-pipelines.yml

@@ -10,11 +10,11 @@ resources:
       type: github
       endpoint: github
       name: asottile/azure-pipeline-templates
-      ref: refs/tags/v1.0.0
+      ref: refs/tags/v2.0.0
 
 jobs:
 - template: job--pre-commit.yml@asottile
 - template: job--python-tox.yml@asottile
   parameters:
-    toxenvs: [py36, py37, py38]
+    toxenvs: [pypy3, py36, py37, py38, py39]
     os: linux

babi/_types.py

@@ -2,7 +2,5 @@ from typing import TYPE_CHECKING
 
 if TYPE_CHECKING:
     from typing import Protocol  # python3.8+
-    from typing_extensions import TypedDict  # python3.8+
 else:
     Protocol = object
-    TypedDict = dict

babi/buf.py

@@ -0,0 +1,307 @@
+import bisect
+import contextlib
+from typing import Callable
+from typing import Generator
+from typing import Iterator
+from typing import List
+from typing import NamedTuple
+from typing import Optional
+from typing import Tuple
+
+from babi._types import Protocol
+from babi.horizontal_scrolling import line_x
+from babi.horizontal_scrolling import scrolled_line
+from babi.horizontal_scrolling import wcwidth
+from babi.margin import Margin
+
+SetCallback = Callable[['Buf', int, str], None]
+DelCallback = Callable[['Buf', int, str], None]
+InsCallback = Callable[['Buf', int], None]
+
+
+def _offsets(s: str, tab_size: int) -> Tuple[int, ...]:
+    ret = [0]
+    for c in s:
+        if c == '\t':
+            ret.append(ret[-1] + (tab_size - ret[-1] % tab_size))
+        else:
+            ret.append(ret[-1] + wcwidth(c))
+    return tuple(ret)
+
+
+class Modification(Protocol):
+    def __call__(self, buf: 'Buf') -> None: ...
+
+
+class SetModification(NamedTuple):
+    idx: int
+    s: str
+
+    def __call__(self, buf: 'Buf') -> None:
+        buf[self.idx] = self.s
+
+
+class InsModification(NamedTuple):
+    idx: int
+    s: str
+
+    def __call__(self, buf: 'Buf') -> None:
+        buf.insert(self.idx, self.s)
+
+
+class DelModification(NamedTuple):
+    idx: int
+
+    def __call__(self, buf: 'Buf') -> None:
+        del buf[self.idx]
+
+
+class Buf:
+    def __init__(self, lines: List[str], tab_size: int = 4) -> None:
+        self._lines = lines
+        self.tab_size = tab_size
+        self.file_y = self.y = self._x = self._x_hint = 0
+
+        self._set_callbacks: List[SetCallback] = [self._set_cb]
+        self._del_callbacks: List[DelCallback] = [self._del_cb]
+        self._ins_callbacks: List[InsCallback] = [self._ins_cb]
+
+        self._positions: List[Optional[Tuple[int, ...]]] = []
+
+    # read only interface
+
+    def __repr__(self) -> str:
+        return (
+            f'{type(self).__name__}('
+            f'{self._lines!r}, x={self.x}, y={self.y}, file_y={self.file_y}'
+            f')'
+        )
+
+    def __bool__(self) -> bool:
+        return bool(self._lines)
+
+    def __getitem__(self, idx: int) -> str:
+        return self._lines[idx]
+
+    def __iter__(self) -> Iterator[str]:
+        yield from self._lines
+
+    def __len__(self) -> int:
+        return len(self._lines)
+
+    # mutators
+
+    def __setitem__(self, idx: int, val: str) -> None:
+        if idx < 0:
+            idx %= len(self)
+        victim = self._lines[idx]
+
+        self._lines[idx] = val
+
+        for set_callback in self._set_callbacks:
+            set_callback(self, idx, victim)
+
+    def __delitem__(self, idx: int) -> None:
+        if idx < 0:
+            idx %= len(self)
+        victim = self._lines[idx]
+
+        del self._lines[idx]
+
+        for del_callback in self._del_callbacks:
+            del_callback(self, idx, victim)
+
+    def insert(self, idx: int, val: str) -> None:
+        if idx < 0:
+            idx %= len(self)
+
+        self._lines.insert(idx, val)
+
+        for ins_callback in self._ins_callbacks:
+            ins_callback(self, idx)
+
+    # also mutators, but implemented using above functions
+
+    def append(self, val: str) -> None:
+        self.insert(len(self), val)
+
+    def pop(self, idx: int = -1) -> str:
+        victim = self[idx]
+        del self[idx]
+        return victim
+
+    def restore_eof_invariant(self) -> None:
+        """the file lines will always contain a blank empty string at the end'
+        to simplify rendering.  call this whenever the last line may change
+        """
+        if self[-1] != '':
+            self.append('')
+
+    def set_tab_size(self, tab_size: int) -> None:
+        self.tab_size = tab_size
+        self._positions = [None]
+
+    # event handling
+
+    def add_set_callback(self, cb: SetCallback) -> None:
+        self._set_callbacks.append(cb)
+
+    def remove_set_callback(self, cb: SetCallback) -> None:
+        self._set_callbacks.remove(cb)
+
+    def add_del_callback(self, cb: DelCallback) -> None:
+        self._del_callbacks.append(cb)
+
+    def remove_del_callback(self, cb: DelCallback) -> None:
+        self._del_callbacks.remove(cb)
+
+    def add_ins_callback(self, cb: InsCallback) -> None:
+        self._ins_callbacks.append(cb)
+
+    def remove_ins_callback(self, cb: InsCallback) -> None:
+        self._ins_callbacks.remove(cb)
+
+    @contextlib.contextmanager
+    def record(self) -> Generator[List[Modification], None, None]:
+        modifications: List[Modification] = []
+
+        def set_cb(buf: 'Buf', idx: int, victim: str) -> None:
+            modifications.append(SetModification(idx, victim))
+
+        def del_cb(buf: 'Buf', idx: int, victim: str) -> None:
+            modifications.append(InsModification(idx, victim))
+
+        def ins_cb(buf: 'Buf', idx: int) -> None:
+            modifications.append(DelModification(idx))
+
+        self.add_set_callback(set_cb)
+        self.add_del_callback(del_cb)
+        self.add_ins_callback(ins_cb)
+        try:
+            yield modifications
+        finally:
+            self.remove_ins_callback(ins_cb)
+            self.remove_del_callback(del_cb)
+            self.remove_set_callback(set_cb)
+
+    def apply(self, modifications: List[Modification]) -> List[Modification]:
+        with self.record() as ret_modifications:
+            for modification in reversed(modifications):
+                modification(self)
+        return ret_modifications
+
+    # position properties
+
+    @property
+    def displayable_count(self) -> int:
+        return len(self._lines) - self.file_y
+
+    @property
+    def x(self) -> int:
+        return self._x
+
+    @x.setter
+    def x(self, x: int) -> None:
+        self._x = x
+        self._x_hint = self._cursor_x
+
+    def _extend_positions(self, idx: int) -> None:
+        self._positions.extend([None] * (1 + idx - len(self._positions)))
+
+    def _set_cb(self, buf: 'Buf', idx: int, victim: str) -> None:
+        self._extend_positions(idx)
+        self._positions[idx] = None
+
+    def _del_cb(self, buf: 'Buf', idx: int, victim: str) -> None:
+        self._extend_positions(idx)
+        del self._positions[idx]
+
+    def _ins_cb(self, buf: 'Buf', idx: int) -> None:
+        self._extend_positions(idx)
+        self._positions.insert(idx, None)
+
+    def line_positions(self, idx: int) -> Tuple[int, ...]:
+        self._extend_positions(idx)
+        value = self._positions[idx]
+        if value is None:
+            value = _offsets(self._lines[idx], self.tab_size)
+            self._positions[idx] = value
+        return value
+
+    def line_x(self, margin: Margin) -> int:
+        return line_x(self._cursor_x, margin.cols)
+
+    @property
+    def _cursor_x(self) -> int:
+        return self.line_positions(self.y)[self.x]
+
+    def cursor_position(self, margin: Margin) -> Tuple[int, int]:
+        y = self.y - self.file_y + margin.header
+        x = self._cursor_x - self.line_x(margin)
+        return y, x
+
+    # rendered lines
+
+    def rendered_line(self, idx: int, margin: Margin) -> str:
+        x = self._cursor_x if idx == self.y else 0
+        expanded = self._lines[idx].expandtabs(self.tab_size)
+        return scrolled_line(expanded, x, margin.cols)
+
+    # movement
+
+    def scroll_screen_if_needed(self, margin: Margin) -> None:
+        # if the `y` is not on screen, make it so
+        if not (self.file_y <= self.y < self.file_y + margin.body_lines):
+            self.file_y = max(self.y - margin.body_lines // 2, 0)
+
+    def _set_x_after_vertical_movement(self) -> None:
+        positions = self.line_positions(self.y)
+        x = bisect.bisect_left(positions, self._x_hint)
+        x = min(len(self._lines[self.y]), x)
+        if positions[x] > self._x_hint:
+            x -= 1
+        self._x = x
+
+    def up(self, margin: Margin) -> None:
+        if self.y > 0:
+            self.y -= 1
+            if self.y < self.file_y:
+                self.file_y = max(self.file_y - margin.scroll_amount, 0)
+            self._set_x_after_vertical_movement()
+
+    def down(self, margin: Margin) -> None:
+        if self.y < len(self._lines) - 1:
+            self.y += 1
+            if self.y >= self.file_y + margin.body_lines:
+                self.file_y += margin.scroll_amount
+            self._set_x_after_vertical_movement()
+
+    def right(self, margin: Margin) -> None:
+        if self.x >= len(self._lines[self.y]):
+            if self.y < len(self._lines) - 1:
+                self.down(margin)
+                self.x = 0
+        else:
+            self.x += 1
+
+    def left(self, margin: Margin) -> None:
+        if self.x == 0:
+            if self.y > 0:
+                self.up(margin)
+                self.x = len(self._lines[self.y])
+        else:
+            self.x -= 1
+
+    # screen movement
+
+    def file_up(self, margin: Margin) -> None:
+        if self.file_y > 0:
+            self.file_y -= 1
+            if self.y > self.file_y + margin.body_lines - 1:
+                self.up(margin)
+
+    def file_down(self, margin: Margin) -> None:
+        if self.file_y < len(self._lines) - 1:
+            self.file_y += 1
+            if self.y < self.file_y:
+                self.down(margin)

babi/color.py

@@ -1,5 +1,9 @@
 from typing import NamedTuple
 
+# TODO: find a standard which defines these
+# limited number of "named" colors
+NAMED_COLORS = {'white': '#ffffff', 'black': '#000000'}
+
 
 class Color(NamedTuple):
     r: int
@@ -8,4 +12,9 @@ class Color(NamedTuple):
 
     @classmethod
     def parse(cls, s: str) -> 'Color':
-        return cls(r=int(s[1:3], 16), g=int(s[3:5], 16), b=int(s[5:7], 16))
+        if s.startswith('#') and len(s) >= 7:
+            return cls(r=int(s[1:3], 16), g=int(s[3:5], 16), b=int(s[5:7], 16))
+        elif s.startswith('#'):
+            return cls.parse(f'#{s[1] * 2}{s[2] * 2}{s[3] * 2}')
+        else:
+            return cls.parse(NAMED_COLORS[s])

babi/color_manager.py

@@ -1,4 +1,3 @@
-import contextlib
 import curses
 from typing import Dict
 from typing import NamedTuple
@@ -19,26 +18,25 @@ class ColorManager(NamedTuple):
     raw_pairs: Dict[Tuple[int, int], int]
 
     def init_color(self, color: Color) -> None:
-        if curses.COLORS < 256:
+        if curses.can_change_color():
-            return
-        elif curses.can_change_color():
             n = min(self.colors.values(), default=256) - 1
             self.colors[color] = n
             curses.init_color(n, *_color_to_curses(color))
-        else:
+        elif curses.COLORS >= 256:
             self.colors[color] = color_kd.nearest(color, color_kd.make_256())
+        else:
+            self.colors[color] = -1
 
     def color_pair(self, fg: Optional[Color], bg: Optional[Color]) -> int:
-        if curses.COLORS < 256:
-            return 0
-
         fg_i = self.colors[fg] if fg is not None else -1
         bg_i = self.colors[bg] if bg is not None else -1
         return self.raw_color_pair(fg_i, bg_i)
 
     def raw_color_pair(self, fg: int, bg: int) -> int:
-        with contextlib.suppress(KeyError):
+        try:
             return self.raw_pairs[(fg, bg)]
+        except KeyError:
+            pass
 
         n = self.raw_pairs[(fg, bg)] = len(self.raw_pairs) + 1
         curses.init_pair(n, fg, bg)

babi/fdict.py

@@ -3,8 +3,10 @@ from typing import Iterable
 from typing import Mapping
 from typing import TypeVar
 
-TKey = TypeVar('TKey')
+from babi._types import Protocol
-TValue = TypeVar('TValue')
+
+TKey = TypeVar('TKey', contravariant=True)
+TValue = TypeVar('TValue', covariant=True)
 
 
 class FDict(Generic[TKey, TValue]):
@@ -22,3 +24,21 @@ class FDict(Generic[TKey, TValue]):
 
     def values(self) -> Iterable[TValue]:
         return self._dct.values()
+
+
+class Indexable(Generic[TKey, TValue], Protocol):
+    def __getitem__(self, key: TKey) -> TValue: ...
+
+
+class FChainMap(Generic[TKey, TValue]):
+    def __init__(self, *mappings: Indexable[TKey, TValue]) -> None:
+        self._mappings = mappings
+
+    def __getitem__(self, key: TKey) -> TValue:
+        for mapping in reversed(self._mappings):
+            try:
+                return mapping[key]
+            except KeyError:
+                pass
+        else:
+            raise KeyError(key)

babi/highlight.py

@@ -1,4 +1,3 @@
-import contextlib
 import functools
 import json
 import os.path
@@ -10,20 +9,31 @@ from typing import Match
 from typing import NamedTuple
 from typing import Optional
 from typing import Tuple
+from typing import TypeVar
+
+from identify.identify import tags_from_filename
 
 from babi._types import Protocol
-from babi.fdict import FDict
+from babi.fdict import FChainMap
 from babi.reg import _Reg
 from babi.reg import _RegSet
 from babi.reg import ERR_REG
+from babi.reg import expand_escaped
 from babi.reg import make_reg
 from babi.reg import make_regset
 
+T = TypeVar('T')
 Scope = Tuple[str, ...]
 Regions = Tuple['Region', ...]
 Captures = Tuple[Tuple[int, '_Rule'], ...]
 
 
+def uniquely_constructed(t: T) -> T:
+    """avoid tuple.__hash__ for "singleton" constructed objects"""
+    t.__hash__ = object.__hash__  # type: ignore
+    return t
+
+
 def _split_name(s: Optional[str]) -> Tuple[str, ...]:
     if s is None:
         return ()
@@ -57,8 +67,11 @@ class _Rule(Protocol):
     def include(self) -> Optional[str]: ...
     @property
     def patterns(self) -> 'Tuple[_Rule, ...]': ...
+    @property
+    def repository(self) -> 'FChainMap[str, _Rule]': ...
 
 
+@uniquely_constructed
 class Rule(NamedTuple):
     name: Tuple[str, ...]
     match: Optional[str]
@@ -72,9 +85,24 @@ class Rule(NamedTuple):
     while_captures: Captures
     include: Optional[str]
     patterns: Tuple[_Rule, ...]
+    repository: FChainMap[str, _Rule]
 
     @classmethod
-    def from_dct(cls, dct: Dict[str, Any]) -> _Rule:
+    def make(
+            cls,
+            dct: Dict[str, Any],
+            parent_repository: FChainMap[str, _Rule],
+    ) -> _Rule:
+        if 'repository' in dct:
+            # this looks odd, but it's so we can have a self-referential
+            # immutable-after-construction chain map
+            repository_dct: Dict[str, _Rule] = {}
+            repository = FChainMap(parent_repository, repository_dct)
+            for k, sub_dct in dct['repository'].items():
+                repository_dct[k] = Rule.make(sub_dct, repository)
+        else:
+            repository = parent_repository
+
         name = _split_name(dct.get('name'))
         match = dct.get('match')
         begin = dct.get('begin')
@@ -84,7 +112,7 @@ class Rule(NamedTuple):
 
         if 'captures' in dct:
             captures = tuple(
-                (int(k), Rule.from_dct(v))
+                (int(k), Rule.make(v, repository))
                 for k, v in dct['captures'].items()
             )
         else:
@@ -92,7 +120,7 @@ class Rule(NamedTuple):
 
         if 'beginCaptures' in dct:
             begin_captures = tuple(
-                (int(k), Rule.from_dct(v))
+                (int(k), Rule.make(v, repository))
                 for k, v in dct['beginCaptures'].items()
             )
         else:
@@ -100,7 +128,7 @@ class Rule(NamedTuple):
 
         if 'endCaptures' in dct:
             end_captures = tuple(
-                (int(k), Rule.from_dct(v))
+                (int(k), Rule.make(v, repository))
                 for k, v in dct['endCaptures'].items()
             )
         else:
@@ -108,12 +136,16 @@ class Rule(NamedTuple):
 
         if 'whileCaptures' in dct:
             while_captures = tuple(
-                (int(k), Rule.from_dct(v))
+                (int(k), Rule.make(v, repository))
                 for k, v in dct['whileCaptures'].items()
             )
         else:
             while_captures = ()
 
+        # some grammars (at least xml) have begin rules with no end
+        if begin is not None and end is None and while_ is None:
+            end = '$impossible^'
+
         # Using the captures key for a begin/end/while rule is short-hand for
         # giving both beginCaptures and endCaptures with same values
         if begin and end and captures:
@@ -126,7 +158,7 @@ class Rule(NamedTuple):
         include = dct.get('include')
 
         if 'patterns' in dct:
-            patterns = tuple(Rule.from_dct(d) for d in dct['patterns'])
+            patterns = tuple(Rule.make(d, repository) for d in dct['patterns'])
         else:
             patterns = ()
 
@@ -143,69 +175,34 @@ class Rule(NamedTuple):
             while_captures=while_captures,
             include=include,
             patterns=patterns,
-        )
-
-
-class Grammar(NamedTuple):
-    scope_name: str
-    first_line_match: Optional[_Reg]
-    file_types: FrozenSet[str]
-    patterns: Tuple[_Rule, ...]
-    repository: FDict[str, _Rule]
-
-    @classmethod
-    def from_data(cls, data: Dict[str, Any]) -> 'Grammar':
-        scope_name = data['scopeName']
-        if 'firstLineMatch' in data:
-            first_line_match: Optional[_Reg] = make_reg(data['firstLineMatch'])
-        else:
-            first_line_match = None
-        if 'fileTypes' in data:
-            file_types = frozenset(data['fileTypes'])
-        else:
-            file_types = frozenset()
-        patterns = tuple(Rule.from_dct(dct) for dct in data['patterns'])
-        if 'repository' in data:
-            repository = FDict({
-                k: Rule.from_dct(dct) for k, dct in data['repository'].items()
-            })
-        else:
-            repository = FDict({})
-        return cls(
-            scope_name=scope_name,
-            first_line_match=first_line_match,
-            file_types=file_types,
-            patterns=patterns,
             repository=repository,
         )
 
-    @classmethod
+
-    def parse(cls, filename: str) -> 'Grammar':
+@uniquely_constructed
-        with open(filename) as f:
+class Grammar(NamedTuple):
-            return cls.from_data(json.load(f))
+    scope_name: str
+    repository: FChainMap[str, _Rule]
+    patterns: Tuple[_Rule, ...]
 
     @classmethod
-    def blank(cls) -> 'Grammar':
+    def make(cls, data: Dict[str, Any]) -> 'Grammar':
-        return cls(
+        scope_name = data['scopeName']
-            scope_name='source.unknown',
+        if 'repository' in data:
-            first_line_match=None,
+            # this looks odd, but it's so we can have a self-referential
-            file_types=frozenset(),
+            # immutable-after-construction chain map
-            patterns=(),
+            repository_dct: Dict[str, _Rule] = {}
-            repository=FDict({}),
+            repository = FChainMap(repository_dct)
-        )
+            for k, dct in data['repository'].items():
-
+                repository_dct[k] = Rule.make(dct, repository)
-    def matches_file(self, filename: str, first_line: str) -> bool:
-        _, ext = os.path.splitext(filename)
-        if ext.lstrip('.') in self.file_types:
-            return True
-        elif self.first_line_match is not None:
-            return bool(
-                self.first_line_match.match(
-                    first_line, 0, first_line=True, boundary=True,
-                ),
-            )
         else:
-            return False
+            repository = FChainMap()
+        patterns = tuple(Rule.make(d, repository) for d in data['patterns'])
+        return cls(
+            scope_name=scope_name,
+            repository=repository,
+            patterns=patterns,
+        )
 
 
 class Region(NamedTuple):
@@ -276,6 +273,7 @@ class CompiledRegsetRule(CompiledRule, Protocol):
 class Entry(NamedTuple):
     scope: Tuple[str, ...]
     rule: CompiledRule
+    start: Tuple[str, int]
     reg: _Reg = ERR_REG
     boundary: bool = False
 
@@ -287,7 +285,7 @@ def _inner_capture_parse(
         scope: Scope,
         rule: CompiledRule,
 ) -> Regions:
-    state = State.root(Entry(scope + rule.name, rule))
+    state = State.root(Entry(scope + rule.name, rule, (s, 0)))
     _, regions = highlight_line(compiler, state, s, first_line=False)
     return tuple(
         r._replace(start=r.start + start, end=r.end + start) for r in regions
@@ -369,6 +367,7 @@ def _do_regset(
     return state, match.end(), boundary, tuple(ret)
 
 
+@uniquely_constructed
 class PatternRule(NamedTuple):
     name: Tuple[str, ...]
     regset: _RegSet
@@ -395,6 +394,7 @@ class PatternRule(NamedTuple):
         return _do_regset(idx, match, self, compiler, state, pos)
 
 
+@uniquely_constructed
 class MatchRule(NamedTuple):
     name: Tuple[str, ...]
     captures: Captures
@@ -420,6 +420,7 @@ class MatchRule(NamedTuple):
         raise AssertionError(f'unreachable {self}')
 
 
+@uniquely_constructed
 class EndRule(NamedTuple):
     name: Tuple[str, ...]
     content_name: Tuple[str, ...]
@@ -439,8 +440,9 @@ class EndRule(NamedTuple):
         next_scope = scope + self.content_name
 
         boundary = match.end() == len(match.string)
-        reg = make_reg(match.expand(self.end))
+        reg = make_reg(expand_escaped(match, self.end))
-        state = state.push(Entry(next_scope, self, reg, boundary))
+        start = (match.string, match.start())
+        state = state.push(Entry(next_scope, self, start, reg, boundary))
         regions = _captures(compiler, scope, match, self.begin_captures)
         return state, True, regions
 
@@ -455,7 +457,16 @@ class EndRule(NamedTuple):
         if m.start() > pos:
             ret.append(Region(pos, m.start(), state.cur.scope))
         ret.extend(_captures(compiler, state.cur.scope, m, self.end_captures))
-        return state.pop(), m.end(), False, tuple(ret)
+        # this is probably a bug in the grammar, but it pushed and popped at
+        # the same position.
+        # we'll advance the highlighter by one position to get past the loop
+        # this appears to be what vs code does as well
+        if state.entries[-1].start == (m.string, m.end()):
+            ret.append(Region(m.end(), m.end() + 1, state.cur.scope))
+            end = m.end() + 1
+        else:
+            end = m.end()
+        return state.pop(), end, False, tuple(ret)
 
     def search(
             self,
@@ -480,6 +491,7 @@ class EndRule(NamedTuple):
                 return _do_regset(idx, match, self, compiler, state, pos)
 
 
+@uniquely_constructed
 class WhileRule(NamedTuple):
     name: Tuple[str, ...]
     content_name: Tuple[str, ...]
@@ -499,8 +511,10 @@ class WhileRule(NamedTuple):
         next_scope = scope + self.content_name
 
         boundary = match.end() == len(match.string)
-        reg = make_reg(match.expand(self.while_))
+        reg = make_reg(expand_escaped(match, self.while_))
-        state = state.push_while(self, Entry(next_scope, self, reg, boundary))
+        start = (match.string, match.start())
+        entry = Entry(next_scope, self, start, reg, boundary)
+        state = state.push_while(self, entry)
         regions = _captures(compiler, scope, match, self.begin_captures)
         return state, True, regions
 
@@ -534,13 +548,13 @@ class WhileRule(NamedTuple):
 
 
 class Compiler:
-    def __init__(self, grammar: Grammar, grammars: Dict[str, Grammar]) -> None:
+    def __init__(self, grammar: Grammar, grammars: 'Grammars') -> None:
         self._root_scope = grammar.scope_name
         self._grammars = grammars
         self._rule_to_grammar: Dict[_Rule, Grammar] = {}
         self._c_rules: Dict[_Rule, CompiledRule] = {}
         root = self._compile_root(grammar)
-        self.root_state = State.root(Entry(root.name, root))
+        self.root_state = State.root(Entry(root.name, root, ('', 0)))
 
     def _visit_rule(self, grammar: Grammar, rule: _Rule) -> _Rule:
         self._rule_to_grammar[rule] = grammar
@@ -550,19 +564,23 @@ class Compiler:
     def _include(
             self,
             grammar: Grammar,
+            repository: FChainMap[str, _Rule],
             s: str,
     ) -> Tuple[List[str], Tuple[_Rule, ...]]:
         if s == '$self':
             return self._patterns(grammar, grammar.patterns)
         elif s == '$base':
-            return self._include(self._grammars[self._root_scope], '$self')
+            grammar = self._grammars.grammar_for_scope(self._root_scope)
+            return self._include(grammar, grammar.repository, '$self')
         elif s.startswith('#'):
-            return self._patterns(grammar, (grammar.repository[s[1:]],))
+            return self._patterns(grammar, (repository[s[1:]],))
         elif '#' not in s:
-            return self._include(self._grammars[s], '$self')
+            grammar = self._grammars.grammar_for_scope(s)
+            return self._include(grammar, grammar.repository, '$self')
         else:
             scope, _, s = s.partition('#')
-            return self._include(self._grammars[scope], f'#{s}')
+            grammar = self._grammars.grammar_for_scope(scope)
+            return self._include(grammar, grammar.repository, f'#{s}')
 
     @functools.lru_cache(maxsize=None)
     def _patterns(
@@ -574,7 +592,9 @@ class Compiler:
         ret_rules: List[_Rule] = []
         for rule in rules:
             if rule.include is not None:
-                tmp_regs, tmp_rules = self._include(grammar, rule.include)
+                tmp_regs, tmp_rules = self._include(
+                    grammar, rule.repository, rule.include,
+                )
                 ret_regs.extend(tmp_regs)
                 ret_rules.extend(tmp_rules)
             elif rule.match is None and rule.begin is None and rule.patterns:
@@ -634,8 +654,10 @@ class Compiler:
             return PatternRule(rule.name, make_regset(*regs), rules)
 
     def compile_rule(self, rule: _Rule) -> CompiledRule:
-        with contextlib.suppress(KeyError):
+        try:
             return self._c_rules[rule]
+        except KeyError:
+            pass
 
         grammar = self._rule_to_grammar[rule]
         ret = self._c_rules[rule] = self._compile_rule(grammar, rule)
@@ -643,49 +665,87 @@ class Compiler:
 
 
 class Grammars:
-    def __init__(self, grammars: List[Grammar]) -> None:
+    def __init__(self, *directories: str) -> None:
-        self.grammars = {grammar.scope_name: grammar for grammar in grammars}
+        self._scope_to_files = {
-        self._compilers: Dict[Grammar, Compiler] = {}
+            os.path.splitext(filename)[0]: os.path.join(directory, filename)
+            for directory in directories
+            if os.path.exists(directory)
+            for filename in sorted(os.listdir(directory))
+            if filename.endswith('.json')
+        }
 
-    @classmethod
+        unknown_grammar = {'scopeName': 'source.unknown', 'patterns': []}
-    def from_syntax_dir(cls, syntax_dir: str) -> 'Grammars':
+        self._raw = {'source.unknown': unknown_grammar}
-        grammars = [Grammar.blank()]
+        self._file_types: List[Tuple[FrozenSet[str], str]] = []
-        if os.path.exists(syntax_dir):
+        self._first_line: List[Tuple[_Reg, str]] = []
-            grammars.extend(
+        self._parsed: Dict[str, Grammar] = {}
-                Grammar.parse(os.path.join(syntax_dir, filename))
+        self._compiled: Dict[str, Compiler] = {}
-                for filename in os.listdir(syntax_dir)
-            )
-        return cls(grammars)
 
-    def _compiler_for_grammar(self, grammar: Grammar) -> Compiler:
+    def _raw_for_scope(self, scope: str) -> Dict[str, Any]:
-        with contextlib.suppress(KeyError):
+        try:
-            return self._compilers[grammar]
+            return self._raw[scope]
+        except KeyError:
+            pass
 
-        ret = self._compilers[grammar] = Compiler(grammar, self.grammars)
+        grammar_path = self._scope_to_files.pop(scope)
+        with open(grammar_path) as f:
+            ret = self._raw[scope] = json.load(f)
+
+        file_types = frozenset(ret.get('fileTypes', ()))
+        first_line = make_reg(ret.get('firstLineMatch', '$impossible^'))
+
+        self._file_types.append((file_types, scope))
+        self._first_line.append((first_line, scope))
+
+        return ret
+
+    def grammar_for_scope(self, scope: str) -> Grammar:
+        try:
+            return self._parsed[scope]
+        except KeyError:
+            pass
+
+        raw = self._raw_for_scope(scope)
+        ret = self._parsed[scope] = Grammar.make(raw)
         return ret
 
     def compiler_for_scope(self, scope: str) -> Compiler:
-        return self._compiler_for_grammar(self.grammars[scope])
+        try:
+            return self._compiled[scope]
+        except KeyError:
+            pass
+
+        grammar = self.grammar_for_scope(scope)
+        ret = self._compiled[scope] = Compiler(grammar, self)
+        return ret
 
     def blank_compiler(self) -> Compiler:
         return self.compiler_for_scope('source.unknown')
 
-    def compiler_for_file(self, filename: str) -> Compiler:
+    def compiler_for_file(self, filename: str, first_line: str) -> Compiler:
-        if os.path.exists(filename):
+        for tag in tags_from_filename(filename) - {'text'}:
-            with open(filename) as f:
+            try:
-                first_line = next(f, '')
+                # TODO: this doesn't always match even if we detect it
-        else:
+                return self.compiler_for_scope(f'source.{tag}')
-            first_line = ''
+            except KeyError:
-        for grammar in self.grammars.values():
+                pass
-            if grammar.matches_file(filename, first_line):
-                break
-        else:
-            grammar = self.grammars['source.unknown']
 
-        return self._compiler_for_grammar(grammar)
+        # didn't find it in the fast path, need to read all the json
+        for k in tuple(self._scope_to_files):
+            self._raw_for_scope(k)
+
+        _, _, ext = os.path.basename(filename).rpartition('.')
+        for extensions, scope in self._file_types:
+            if ext in extensions:
+                return self.compiler_for_scope(scope)
+
+        for reg, scope in self._first_line:
+            if reg.match(first_line, 0, first_line=True, boundary=True):
+                return self.compiler_for_scope(scope)
+
+        return self.compiler_for_scope('source.unknown')
 
 
-@functools.lru_cache(maxsize=None)
 def highlight_line(
         compiler: 'Compiler',
         state: State,

babi/hl/interface.py

@@ -1,27 +1,32 @@
-from typing import Sequence
+from typing import NamedTuple
 from typing import Tuple
 
 from babi._types import Protocol
-from babi._types import TypedDict
+from babi.buf import Buf
-from babi.list_spy import SequenceNoSlice
 
 
-class CursesRegion(TypedDict):
+class HL(NamedTuple):
     x: int
-    n: int
+    end: int
-    color: int
+    attr: int
 
 
-CursesRegions = Tuple[CursesRegion, ...]
+HLs = Tuple[HL, ...]
+
+
+class RegionsMapping(Protocol):
+    def __getitem__(self, idx: int) -> HLs: ...
 
 
 class FileHL(Protocol):
     @property
-    def regions(self) -> Sequence[CursesRegions]: ...
+    def include_edge(self) -> bool: ...
-    def highlight_until(self, lines: SequenceNoSlice, idx: int) -> None: ...
+    @property
-    def touch(self, lineno: int) -> None: ...
+    def regions(self) -> RegionsMapping: ...
+    def highlight_until(self, lines: Buf, idx: int) -> None: ...
+    def register_callbacks(self, buf: Buf) -> None: ...
 
 
 class HLFactory(Protocol):
-    def get_file_highlighter(self, filename: str) -> FileHL: ...
+    def file_highlighter(self, filename: str, first_line: str) -> FileHL: ...
-    def get_blank_file_highlighter(self) -> FileHL: ...
+    def blank_file_highlighter(self) -> FileHL: ...

babi/hl/replace.py

@@ -0,0 +1,32 @@
+import collections
+import contextlib
+import curses
+from typing import Dict
+from typing import Generator
+
+from babi.buf import Buf
+from babi.hl.interface import HL
+from babi.hl.interface import HLs
+
+
+class Replace:
+    include_edge = True
+
+    def __init__(self) -> None:
+        self.regions: Dict[int, HLs] = collections.defaultdict(tuple)
+
+    def highlight_until(self, lines: Buf, idx: int) -> None:
+        """our highlight regions are populated in other ways"""
+
+    def register_callbacks(self, buf: Buf) -> None:
+        """our highlight regions are populated in other ways"""
+
+    @contextlib.contextmanager
+    def region(self, y: int, x: int, end: int) -> Generator[None, None, None]:
+        # XXX: this assumes pair 1 is the background
+        attr = curses.A_REVERSE | curses.A_DIM | curses.color_pair(1)
+        self.regions[y] = (HL(x=x, end=end, attr=attr),)
+        try:
+            yield
+        finally:
+            del self.regions[y]

babi/hl/selection.py

@@ -0,0 +1,58 @@
+import collections
+import curses
+from typing import Dict
+from typing import Optional
+from typing import Tuple
+
+from babi.buf import Buf
+from babi.hl.interface import HL
+from babi.hl.interface import HLs
+
+
+class Selection:
+    include_edge = True
+
+    def __init__(self) -> None:
+        self.regions: Dict[int, HLs] = collections.defaultdict(tuple)
+        self.start: Optional[Tuple[int, int]] = None
+        self.end: Optional[Tuple[int, int]] = None
+
+    def register_callbacks(self, buf: Buf) -> None:
+        """our highlight regions are populated in other ways"""
+
+    def highlight_until(self, lines: Buf, idx: int) -> None:
+        if self.start is None or self.end is None:
+            return
+
+        # XXX: this assumes pair 1 is the background
+        attr = curses.A_REVERSE | curses.A_DIM | curses.color_pair(1)
+        (s_y, s_x), (e_y, e_x) = self.get()
+        if s_y == e_y:
+            self.regions[s_y] = (HL(x=s_x, end=e_x, attr=attr),)
+        else:
+            self.regions[s_y] = (
+                HL(x=s_x, end=len(lines[s_y]) + 1, attr=attr),
+            )
+            for l_y in range(s_y + 1, e_y):
+                self.regions[l_y] = (
+                    HL(x=0, end=len(lines[l_y]) + 1, attr=attr),
+                )
+            self.regions[e_y] = (HL(x=0, end=e_x, attr=attr),)
+
+    def get(self) -> Tuple[Tuple[int, int], Tuple[int, int]]:
+        assert self.start is not None and self.end is not None
+        if self.start < self.end:
+            return self.start, self.end
+        else:
+            return self.end, self.start
+
+    def clear(self) -> None:
+        if self.start is not None and self.end is not None:
+            (s_y, _), (e_y, _) = self.get()
+            for l_y in range(s_y, e_y + 1):
+                del self.regions[l_y]
+        self.start = self.end = None
+
+    def set(self, s_y: int, s_x: int, e_y: int, e_x: int) -> None:
+        self.clear()
+        self.start, self.end = (s_y, s_x), (e_y, e_x)

babi/hl/syntax.py

@@ -1,19 +1,23 @@
 import curses
-from typing import Dict
+import functools
+import math
+from typing import Callable
 from typing import List
 from typing import NamedTuple
+from typing import Optional
 from typing import Tuple
 
+from babi.buf import Buf
 from babi.color_manager import ColorManager
 from babi.highlight import Compiler
 from babi.highlight import Grammars
 from babi.highlight import highlight_line
 from babi.highlight import State
-from babi.hl.interface import CursesRegion
+from babi.hl.interface import HL
-from babi.hl.interface import CursesRegions
+from babi.hl.interface import HLs
-from babi.list_spy import SequenceNoSlice
 from babi.theme import Style
 from babi.theme import Theme
+from babi.user_data import prefix_data
 from babi.user_data import xdg_config
 from babi.user_data import xdg_data
 
@@ -21,6 +25,8 @@ A_ITALIC = getattr(curses, 'A_ITALIC', 0x80000000)  # new in py37
 
 
 class FileSyntax:
+    include_edge = False
+
     def __init__(
             self,
             compiler: Compiler,
@@ -31,11 +37,13 @@ class FileSyntax:
         self._theme = theme
         self._color_manager = color_manager
 
-        self.regions: List[CursesRegions] = []
+        self.regions: List[HLs] = []
         self._states: List[State] = []
 
-        self._hl_cache: Dict[str, Dict[State, Tuple[State, CursesRegions]]]
+        # this will be assigned a functools.lru_cache per instance for
-        self._hl_cache = {}
+        # better hit rate and memory usage
+        self._hl: Optional[Callable[[State, str, bool], Tuple[State, HLs]]]
+        self._hl = None
 
     def attr(self, style: Style) -> int:
         pair = self._color_manager.color_pair(style.fg, style.bg)
@@ -46,72 +54,83 @@ class FileSyntax:
             curses.A_UNDERLINE * style.u
         )
 
-    def _hl(
+    def _hl_uncached(
             self,
             state: State,
             line: str,
-            i: int,
+            first_line: bool,
-    ) -> Tuple[State, CursesRegions]:
+    ) -> Tuple[State, HLs]:
-        try:
-            return self._hl_cache[line][state]
-        except KeyError:
-            pass
-
         new_state, regions = highlight_line(
-            self._compiler, state, f'{line}\n', first_line=i == 0,
+            self._compiler, state, f'{line}\n', first_line=first_line,
         )
 
         # remove the trailing newline
         new_end = regions[-1]._replace(end=regions[-1].end - 1)
         regions = regions[:-1] + (new_end,)
 
-        regs: List[CursesRegion] = []
+        regs: List[HL] = []
         for r in regions:
             style = self._theme.select(r.scope)
             if style == self._theme.default:
                 continue
 
-            n = r.end - r.start
             attr = self.attr(style)
             if (
                     regs and
-                    regs[-1]['color'] == attr and
+                    regs[-1].attr == attr and
-                    regs[-1]['x'] + regs[-1]['n'] == r.start
+                    regs[-1].end == r.start
             ):
-                regs[-1]['n'] += n
+                regs[-1] = regs[-1]._replace(end=r.end)
             else:
-                regs.append(CursesRegion(x=r.start, n=n, color=attr))
+                regs.append(HL(x=r.start, end=r.end, attr=attr))
 
-        dct = self._hl_cache.setdefault(line, {})
+        return new_state, tuple(regs)
-        ret = dct[state] = (new_state, tuple(regs))
+
-        return ret
+    def _set_cb(self, lines: Buf, idx: int, victim: str) -> None:
+        del self.regions[idx:]
+        del self._states[idx:]
+
+    def _del_cb(self, lines: Buf, idx: int, victim: str) -> None:
+        del self.regions[idx:]
+        del self._states[idx:]
+
+    def _ins_cb(self, lines: Buf, idx: int) -> None:
+        del self.regions[idx:]
+        del self._states[idx:]
+
+    def register_callbacks(self, buf: Buf) -> None:
+        buf.add_set_callback(self._set_cb)
+        buf.add_del_callback(self._del_cb)
+        buf.add_ins_callback(self._ins_cb)
+
+    def highlight_until(self, lines: Buf, idx: int) -> None:
+        if self._hl is None:
+            # the docs claim better performance with power of two sizing
+            size = max(4096, 2 ** (int(math.log(len(lines), 2)) + 2))
+            self._hl = functools.lru_cache(maxsize=size)(self._hl_uncached)
 
-    def highlight_until(self, lines: SequenceNoSlice, idx: int) -> None:
         if not self._states:
             state = self._compiler.root_state
         else:
             state = self._states[-1]
 
         for i in range(len(self._states), idx):
-            state, regions = self._hl(state, lines[i], i)
+            # https://github.com/python/mypy/issues/8579
+            state, regions = self._hl(state, lines[i], i == 0)  # type: ignore
             self._states.append(state)
             self.regions.append(regions)
 
-    def touch(self, lineno: int) -> None:
-        del self._states[lineno:]
-        del self.regions[lineno:]
-
 
 class Syntax(NamedTuple):
     grammars: Grammars
     theme: Theme
     color_manager: ColorManager
 
-    def get_file_highlighter(self, filename: str) -> FileSyntax:
+    def file_highlighter(self, filename: str, first_line: str) -> FileSyntax:
-        compiler = self.grammars.compiler_for_file(filename)
+        compiler = self.grammars.compiler_for_file(filename, first_line)
         return FileSyntax(compiler, self.theme, self.color_manager)
 
-    def get_blank_file_highlighter(self) -> FileSyntax:
+    def blank_file_highlighter(self) -> FileSyntax:
         compiler = self.grammars.blank_compiler()
         return FileSyntax(compiler, self.theme, self.color_manager)
 
@@ -139,7 +158,7 @@ class Syntax(NamedTuple):
             stdscr: 'curses._CursesWindow',
             color_manager: ColorManager,
     ) -> 'Syntax':
-        grammars = Grammars.from_syntax_dir(xdg_data('textmate_syntax'))
+        grammars = Grammars(prefix_data('grammar_v1'), xdg_data('grammar_v1'))
         theme = Theme.from_filename(xdg_config('theme.json'))
         ret = cls(grammars, theme, color_manager)
         ret._init_screen(stdscr)

babi/hl/trailing_whitespace.py

@@ -1,20 +1,21 @@
 import curses
 from typing import List
-from typing import NamedTuple
 
+from babi.buf import Buf
 from babi.color_manager import ColorManager
-from babi.hl.interface import CursesRegion
+from babi.hl.interface import HL
-from babi.hl.interface import CursesRegions
+from babi.hl.interface import HLs
-from babi.list_spy import SequenceNoSlice
 
 
-class FileTrailingWhitespace:
+class TrailingWhitespace:
+    include_edge = False
+
     def __init__(self, color_manager: ColorManager) -> None:
         self._color_manager = color_manager
 
-        self.regions: List[CursesRegions] = []
+        self.regions: List[HLs] = []
 
-    def _trailing_ws(self, line: str) -> CursesRegions:
+    def _trailing_ws(self, line: str) -> HLs:
         if not line:
             return ()
 
@@ -27,22 +28,25 @@ class FileTrailingWhitespace:
         else:
             pair = self._color_manager.raw_color_pair(-1, curses.COLOR_RED)
             attr = curses.color_pair(pair)
-            return (CursesRegion(x=i, n=len(line) - i, color=attr),)
+            return (HL(x=i, end=len(line), attr=attr),)
 
-    def highlight_until(self, lines: SequenceNoSlice, idx: int) -> None:
+    def _set_cb(self, lines: Buf, idx: int, victim: str) -> None:
+        if idx < len(self.regions):
+            self.regions[idx] = self._trailing_ws(lines[idx])
+
+    def _del_cb(self, lines: Buf, idx: int, victim: str) -> None:
+        if idx < len(self.regions):
+            del self.regions[idx]
+
+    def _ins_cb(self, lines: Buf, idx: int) -> None:
+        if idx < len(self.regions):
+            self.regions.insert(idx, self._trailing_ws(lines[idx]))
+
+    def register_callbacks(self, buf: Buf) -> None:
+        buf.add_set_callback(self._set_cb)
+        buf.add_del_callback(self._del_cb)
+        buf.add_ins_callback(self._ins_cb)
+
+    def highlight_until(self, lines: Buf, idx: int) -> None:
         for i in range(len(self.regions), idx):
             self.regions.append(self._trailing_ws(lines[i]))
-
-    def touch(self, lineno: int) -> None:
-        del self.regions[lineno:]
-
-
-class TrailingWhitespace(NamedTuple):
-    color_manager: ColorManager
-
-    def get_file_highlighter(self, filename: str) -> FileTrailingWhitespace:
-        # no file-specific behaviour
-        return self.get_blank_file_highlighter()
-
-    def get_blank_file_highlighter(self) -> FileTrailingWhitespace:
-        return FileTrailingWhitespace(self.color_manager)

babi/horizontal_scrolling.py

@@ -1,10 +1,15 @@
+import curses
+
+from babi.cached_property import cached_property
+
+
 def line_x(x: int, width: int) -> int:
-    margin = min(width - 3, 6)
     if x + 1 < width:
         return 0
     elif width == 1:
         return x
     else:
+        margin = min(width - 3, 6)
         return (
             width - margin - 2 +
             (x + 1 - width) //
@@ -17,7 +22,7 @@ def scrolled_line(s: str, x: int, width: int) -> str:
     l_x = line_x(x, width)
     if l_x:
         s = f'«{s[l_x + 1:]}'
-        if l_x and len(s) > width:
+        if len(s) > width:
             return f'{s[:width - 1]}»'
         else:
             return s.ljust(width)
@@ -25,3 +30,17 @@ def scrolled_line(s: str, x: int, width: int) -> str:
         return f'{s[:width - 1]}»'
     else:
         return s.ljust(width)
+
+
+class _CalcWidth:
+    @cached_property
+    def _window(self) -> 'curses._CursesWindow':
+        return curses.newwin(1, 10)
+
+    def wcwidth(self, c: str) -> int:
+        self._window.addstr(0, 0, c)
+        return self._window.getyx()[1]
+
+
+wcwidth = _CalcWidth().wcwidth
+del _CalcWidth

babi/list_spy.py

@@ -1,85 +0,0 @@
-import functools
-import sys
-from typing import Callable
-from typing import Iterator
-from typing import List
-
-from babi._types import Protocol
-
-
-class SequenceNoSlice(Protocol):
-    def __len__(self) -> int: ...
-    def __getitem__(self, idx: int) -> str: ...
-
-    def __iter__(self) -> Iterator[str]:
-        for i in range(len(self)):
-            yield self[i]
-
-
-class MutableSequenceNoSlice(SequenceNoSlice, Protocol):
-    def __setitem__(self, idx: int, val: str) -> None: ...
-    def __delitem__(self, idx: int) -> None: ...
-    def insert(self, idx: int, val: str) -> None: ...
-
-    def append(self, val: str) -> None:
-        self.insert(len(self), val)
-
-    def pop(self, idx: int = -1) -> str:
-        victim = self[idx]
-        del self[idx]
-        return victim
-
-
-def _del(lst: MutableSequenceNoSlice, *, idx: int) -> None:
-    del lst[idx]
-
-
-def _set(lst: MutableSequenceNoSlice, *, idx: int, val: str) -> None:
-    lst[idx] = val
-
-
-def _ins(lst: MutableSequenceNoSlice, *, idx: int, val: str) -> None:
-    lst.insert(idx, val)
-
-
-class ListSpy(MutableSequenceNoSlice):
-    def __init__(self, lst: MutableSequenceNoSlice) -> None:
-        self._lst = lst
-        self._undo: List[Callable[[MutableSequenceNoSlice], None]] = []
-        self.min_line_touched = sys.maxsize
-
-    def __repr__(self) -> str:
-        return f'{type(self).__name__}({self._lst})'
-
-    def __len__(self) -> int:
-        return len(self._lst)
-
-    def __getitem__(self, idx: int) -> str:
-        return self._lst[idx]
-
-    def __setitem__(self, idx: int, val: str) -> None:
-        self._undo.append(functools.partial(_set, idx=idx, val=self._lst[idx]))
-        self.min_line_touched = min(idx, self.min_line_touched)
-        self._lst[idx] = val
-
-    def __delitem__(self, idx: int) -> None:
-        if idx < 0:
-            idx %= len(self)
-        self._undo.append(functools.partial(_ins, idx=idx, val=self._lst[idx]))
-        self.min_line_touched = min(idx, self.min_line_touched)
-        del self._lst[idx]
-
-    def insert(self, idx: int, val: str) -> None:
-        if idx < 0:
-            idx %= len(self)
-        self._undo.append(functools.partial(_del, idx=idx))
-        self.min_line_touched = min(idx, self.min_line_touched)
-        self._lst.insert(idx, val)
-
-    def undo(self, lst: MutableSequenceNoSlice) -> None:
-        for fn in reversed(self._undo):
-            fn(lst)
-
-    @property
-    def has_modifications(self) -> bool:
-        return bool(self._undo)

babi/main.py

@@ -1,17 +1,28 @@
 import argparse
 import curses
+import os
+import re
+import signal
+import sys
+from typing import List
 from typing import Optional
 from typing import Sequence
+from typing import Tuple
 
+from babi.buf import Buf
 from babi.file import File
+from babi.perf import Perf
 from babi.perf import perf_log
 from babi.screen import EditResult
 from babi.screen import make_stdscr
 from babi.screen import Screen
 
+CONSOLE = 'CONIN$' if sys.platform == 'win32' else '/dev/tty'
+POSITION_RE = re.compile(r'^\+-?\d+$')
 
-def _edit(screen: Screen) -> EditResult:
+
-    screen.file.ensure_loaded(screen.status)
+def _edit(screen: Screen, stdin: str) -> EditResult:
+    screen.file.ensure_loaded(screen.status, screen.margin, stdin)
 
     while True:
         screen.status.tick(screen.margin)
@@ -32,35 +43,113 @@ def _edit(screen: Screen) -> EditResult:
             screen.status.update(f'unknown key: {key}')
 
 
-def c_main(stdscr: 'curses._CursesWindow', args: argparse.Namespace) -> int:
+def c_main(
-    with perf_log(args.perf_log) as perf:
+        stdscr: 'curses._CursesWindow',
-        screen = Screen(stdscr, args.filenames or [None], perf)
+        filenames: List[Optional[str]],
-        with screen.history.save():
+        positions: List[int],
-            while screen.files:
+        stdin: str,
-                screen.i = screen.i % len(screen.files)
+        perf: Perf,
-                res = _edit(screen)
+) -> int:
-                if res == EditResult.EXIT:
+    screen = Screen(stdscr, filenames, positions, perf)
-                    del screen.files[screen.i]
+    with screen.history.save():
-                    screen.status.clear()
+        while screen.files:
-                elif res == EditResult.NEXT:
+            screen.i = screen.i % len(screen.files)
-                    screen.i += 1
+            res = _edit(screen, stdin)
-                    screen.status.clear()
+            if res == EditResult.EXIT:
-                elif res == EditResult.PREV:
+                del screen.files[screen.i]
-                    screen.i -= 1
+                # always go to the next file except at the end
-                    screen.status.clear()
+                screen.i = min(screen.i, len(screen.files) - 1)
-                else:
+                screen.status.clear()
-                    raise AssertionError(f'unreachable {res}')
+            elif res == EditResult.NEXT:
+                screen.i += 1
+                screen.status.clear()
+            elif res == EditResult.PREV:
+                screen.i -= 1
+                screen.status.clear()
+            elif res == EditResult.OPEN:
+                screen.i = len(screen.files) - 1
+            else:
+                raise AssertionError(f'unreachable {res}')
     return 0
 
 
+def _key_debug(stdscr: 'curses._CursesWindow', perf: Perf) -> int:
+    screen = Screen(stdscr, ['<<key debug>>'], [0], perf)
+    screen.file.buf = Buf([''])
+
+    while True:
+        screen.status.update('press q to quit')
+        screen.draw()
+        screen.file.move_cursor(screen.stdscr, screen.margin)
+
+        key = screen.get_char()
+        screen.file.buf.insert(-1, f'{key.wch!r} {key.keyname.decode()!r}')
+        screen.file.down(screen.margin)
+        if key.wch == curses.KEY_RESIZE:
+            screen.resize()
+        if key.wch == 'q':
+            return 0
+
+
+def _filenames(filenames: List[str]) -> Tuple[List[Optional[str]], List[int]]:
+    if not filenames:
+        return [None], [0]
+
+    ret_filenames: List[Optional[str]] = []
+    ret_positions = []
+
+    filenames_iter = iter(filenames)
+    for filename in filenames_iter:
+        if POSITION_RE.match(filename):
+            # in the success case we get:
+            #
+            # position_s = +...
+            # filename = (the next thing)
+            #
+            # in the error case we only need to reset `position_s` as
+            # `filename` is already correct
+            position_s = filename
+            try:
+                filename = next(filenames_iter)
+            except StopIteration:
+                position_s = '+0'
+            ret_positions.append(int(position_s[1:]))
+            ret_filenames.append(filename)
+        else:
+            ret_positions.append(0)
+            ret_filenames.append(filename)
+
+    return ret_filenames, ret_positions
+
+
 def main(argv: Optional[Sequence[str]] = None) -> int:
     parser = argparse.ArgumentParser()
     parser.add_argument('filenames', metavar='filename', nargs='*')
     parser.add_argument('--perf-log')
+    parser.add_argument(
+        '--key-debug', action='store_true', help=argparse.SUPPRESS,
+    )
     args = parser.parse_args(argv)
 
-    with make_stdscr() as stdscr:
+    if '-' in args.filenames:
-        return c_main(stdscr, args)
+        print('reading stdin...', file=sys.stderr)
+        stdin = sys.stdin.read()
+        tty = os.open(CONSOLE, os.O_RDONLY)
+        os.dup2(tty, sys.stdin.fileno())
+    else:
+        stdin = ''
+
+    # ignore backgrounding signals, we'll handle those in curses
+    # fixes a problem with ^Z on termination which would break the terminal
+    if sys.platform != 'win32':  # pragma: win32 no cover  # pragma: no branch
+        signal.signal(signal.SIGTSTP, signal.SIG_IGN)
+
+    with perf_log(args.perf_log) as perf, make_stdscr() as stdscr:
+        if args.key_debug:
+            return _key_debug(stdscr, perf)
+        else:
+            filenames, positions = _filenames(args.filenames)
+            return c_main(stdscr, filenames, positions, stdin, perf)
 
 
 if __name__ == '__main__':

babi/margin.py

@@ -3,12 +3,20 @@ from typing import NamedTuple
 
 
 class Margin(NamedTuple):
-    header: bool
+    lines: int
-    footer: bool
+    cols: int
+
+    @property
+    def header(self) -> bool:
+        return self.lines > 2
+
+    @property
+    def footer(self) -> bool:
+        return self.lines > 1
 
     @property
     def body_lines(self) -> int:
-        return curses.LINES - self.header - self.footer
+        return self.lines - self.header - self.footer
 
     @property
     def page_size(self) -> int:
@@ -17,11 +25,11 @@ class Margin(NamedTuple):
         else:
             return self.body_lines - 2
 
+    @property
+    def scroll_amount(self) -> int:
+        # integer round up without banker's rounding (so 1/2 => 1 instead of 0)
+        return int(self.lines / 2 + .5)
+
     @classmethod
     def from_current_screen(cls) -> 'Margin':
-        if curses.LINES == 1:
+        return cls(curses.LINES, curses.COLS)
-            return cls(header=False, footer=False)
-        elif curses.LINES == 2:
-            return cls(header=False, footer=True)
-        else:
-            return cls(header=True, footer=True)

babi/prompt.py

@@ -33,18 +33,19 @@ class Prompt:
 
     def _render_prompt(self, *, base: Optional[str] = None) -> None:
         base = base or self._prompt
-        if not base or curses.COLS < 7:
+        if not base or self._screen.margin.cols < 7:
             prompt_s = ''
-        elif len(base) > curses.COLS - 6:
+        elif len(base) > self._screen.margin.cols - 6:
-            prompt_s = f'{base[:curses.COLS - 7]}…: '
+            prompt_s = f'{base[:self._screen.margin.cols - 7]}…: '
         else:
             prompt_s = f'{base}: '
-        width = curses.COLS - len(prompt_s)
+        width = self._screen.margin.cols - len(prompt_s)
         line = scrolled_line(self._s, self._x, width)
         cmd = f'{prompt_s}{line}'
-        self._screen.stdscr.insstr(curses.LINES - 1, 0, cmd, curses.A_REVERSE)
+        prompt_line = self._screen.margin.lines - 1
+        self._screen.stdscr.insstr(prompt_line, 0, cmd, curses.A_REVERSE)
         x = len(prompt_s) + self._x - line_x(self._x, width)
-        self._screen.stdscr.move(curses.LINES - 1, x)
+        self._screen.stdscr.move(prompt_line, x)
 
     def _up(self) -> None:
         self._y = max(0, self._y - 1)
@@ -126,7 +127,7 @@ class Prompt:
             key = self._screen.get_char()
             if key.keyname == b'KEY_RESIZE':
                 self._screen.resize()
-            elif key.keyname == b'KEY_BACKSPACE' or key.keyname == b'^H':
+            elif key.keyname == b'KEY_BACKSPACE':
                 reverse_s = reverse_s[:-1]
             elif key.keyname == b'^R':
                 idx = max(0, idx - 1)
@@ -163,7 +164,6 @@ class Prompt:
         b'kLFT5': _ctrl_left,
         # editing
         b'KEY_BACKSPACE': _backspace,
-        b'^H': _backspace,  # ^Backspace
         b'KEY_DC': _delete,
         b'^K': _cut_to_end,
         # misc

babi/reg.py

@@ -1,82 +1,41 @@
 import functools
+import re
 from typing import Match
 from typing import Optional
 from typing import Tuple
 
 import onigurumacffi
 
-from babi.cached_property import cached_property
+_BACKREF_RE = re.compile(r'((?<!\\)(?:\\\\)*)\\([0-9]+)')
 
 
-def _replace_esc(s: str, chars: str) -> str:
+_FLAGS = {
-    """replace the given escape sequences of `chars` with \\uffff"""
+    # (first_line, boundary)
-    for c in chars:
+    (False, False): (
-        if f'\\{c}' in s:
+        onigurumacffi.OnigSearchOption.NOT_END_STRING |
-            break
+        onigurumacffi.OnigSearchOption.NOT_BEGIN_STRING |
-    else:
+        onigurumacffi.OnigSearchOption.NOT_BEGIN_POSITION
-        return s
+    ),
-
+    (False, True): (
-    b = []
+        onigurumacffi.OnigSearchOption.NOT_END_STRING |
-    i = 0
+        onigurumacffi.OnigSearchOption.NOT_BEGIN_STRING
-    length = len(s)
+    ),
-    while i < length:
+    (True, False): (
-        try:
+        onigurumacffi.OnigSearchOption.NOT_END_STRING |
-            sbi = s.index('\\', i)
+        onigurumacffi.OnigSearchOption.NOT_BEGIN_POSITION
-        except ValueError:
+    ),
-            b.append(s[i:])
+    (True, True): onigurumacffi.OnigSearchOption.NOT_END_STRING,
-            break
+}
-        if sbi > i:
-            b.append(s[i:sbi])
-        b.append('\\')
-        i = sbi + 1
-        if i < length:
-            if s[i] in chars:
-                b.append('\uffff')
-            else:
-                b.append(s[i])
-        i += 1
-    return ''.join(b)
 
 
 class _Reg:
     def __init__(self, s: str) -> None:
         self._pattern = s
+        self._reg = onigurumacffi.compile(self._pattern)
 
     def __repr__(self) -> str:
         return f'{type(self).__name__}({self._pattern!r})'
 
-    @cached_property
-    def _reg(self) -> onigurumacffi._Pattern:
-        return onigurumacffi.compile(self._pattern)
-
-    @cached_property
-    def _reg_no_A(self) -> onigurumacffi._Pattern:
-        return onigurumacffi.compile(_replace_esc(self._pattern, 'A'))
-
-    @cached_property
-    def _reg_no_G(self) -> onigurumacffi._Pattern:
-        return onigurumacffi.compile(_replace_esc(self._pattern, 'G'))
-
-    @cached_property
-    def _reg_no_A_no_G(self) -> onigurumacffi._Pattern:
-        return onigurumacffi.compile(_replace_esc(self._pattern, 'AG'))
-
-    def _get_reg(
-            self,
-            first_line: bool,
-            boundary: bool,
-    ) -> onigurumacffi._Pattern:
-        if boundary:
-            if first_line:
-                return self._reg
-            else:
-                return self._reg_no_A
-        else:
-            if first_line:
-                return self._reg_no_G
-            else:
-                return self._reg_no_A_no_G
-
     def search(
             self,
             line: str,
@@ -84,7 +43,7 @@ class _Reg:
             first_line: bool,
             boundary: bool,
     ) -> Optional[Match[str]]:
-        return self._get_reg(first_line, boundary).search(line, pos)
+        return self._reg.search(line, pos, flags=_FLAGS[first_line, boundary])
 
     def match(
             self,
@@ -93,36 +52,18 @@ class _Reg:
             first_line: bool,
             boundary: bool,
     ) -> Optional[Match[str]]:
-        return self._get_reg(first_line, boundary).match(line, pos)
+        return self._reg.match(line, pos, flags=_FLAGS[first_line, boundary])
 
 
 class _RegSet:
     def __init__(self, *s: str) -> None:
         self._patterns = s
+        self._set = onigurumacffi.compile_regset(*self._patterns)
 
     def __repr__(self) -> str:
         args = ', '.join(repr(s) for s in self._patterns)
         return f'{type(self).__name__}({args})'
 
-    @cached_property
-    def _set(self) -> onigurumacffi._RegSet:
-        return onigurumacffi.compile_regset(*self._patterns)
-
-    @cached_property
-    def _set_no_A(self) -> onigurumacffi._RegSet:
-        patterns = (_replace_esc(p, 'A') for p in self._patterns)
-        return onigurumacffi.compile_regset(*patterns)
-
-    @cached_property
-    def _set_no_G(self) -> onigurumacffi._RegSet:
-        patterns = (_replace_esc(p, 'G') for p in self._patterns)
-        return onigurumacffi.compile_regset(*patterns)
-
-    @cached_property
-    def _set_no_A_no_G(self) -> onigurumacffi._RegSet:
-        patterns = (_replace_esc(p, 'AG') for p in self._patterns)
-        return onigurumacffi.compile_regset(*patterns)
-
     def search(
             self,
             line: str,
@@ -130,18 +71,13 @@ class _RegSet:
             first_line: bool,
             boundary: bool,
     ) -> Tuple[int, Optional[Match[str]]]:
-        if boundary:
+        return self._set.search(line, pos, flags=_FLAGS[first_line, boundary])
-            if first_line:
+
-                return self._set.search(line, pos)
+
-            else:
+def expand_escaped(match: Match[str], s: str) -> str:
-                return self._set_no_A.search(line, pos)
+    return _BACKREF_RE.sub(lambda m: f'{m[1]}{re.escape(match[int(m[2])])}', s)
-        else:
-            if first_line:
-                return self._set_no_G.search(line, pos)
-            else:
-                return self._set_no_A_no_G.search(line, pos)
 
 
 make_reg = functools.lru_cache(maxsize=None)(_Reg)
 make_regset = functools.lru_cache(maxsize=None)(_RegSet)
-ERR_REG = make_reg(')this pattern always triggers an error when used(')
+ERR_REG = make_reg('$ ^')

babi/screen.py

@@ -6,7 +6,6 @@ import os
 import re
 import signal
 import sys
-from typing import Callable
 from typing import Generator
 from typing import List
 from typing import NamedTuple
@@ -21,15 +20,19 @@ from babi.file import File
 from babi.file import get_lines
 from babi.history import History
 from babi.hl.syntax import Syntax
-from babi.hl.trailing_whitespace import TrailingWhitespace
 from babi.margin import Margin
 from babi.perf import Perf
 from babi.prompt import Prompt
 from babi.prompt import PromptResult
 from babi.status import Status
 
-VERSION_STR = 'babi v0'
+if sys.version_info >= (3, 8):  # pragma: no cover (py38+)
-EditResult = enum.Enum('EditResult', 'EXIT NEXT PREV')
+    import importlib.metadata as importlib_metadata
+else:  # pragma: no cover (<py38)
+    import importlib_metadata
+
+VERSION_STR = f'babi v{importlib_metadata.version("babi")}'
+EditResult = enum.Enum('EditResult', 'EXIT NEXT PREV OPEN')
 
 # TODO: find a place to populate these, surely there's a database somewhere
 SEQUENCE_KEYNAME = {
@@ -58,6 +61,40 @@ SEQUENCE_KEYNAME = {
     '\x1b[1;6H': b'kHOM6',  # Shift + ^Home
     '\x1b[1;6F': b'kEND6',  # Shift + ^End
 }
+KEYNAME_REWRITE = {
+    # windows-curses: numeric pad arrow keys
+    # - some overlay keyboards pick these as well
+    # - in xterm it seems these are mapped automatically
+    b'KEY_A2': b'KEY_UP',
+    b'KEY_C2': b'KEY_DOWN',
+    b'KEY_B3': b'KEY_RIGHT',
+    b'KEY_B1': b'KEY_LEFT',
+    b'PADSTOP': b'KEY_DC',
+    b'KEY_A3': b'KEY_PPAGE',
+    b'KEY_C3': b'KEY_NPAGE',
+    b'KEY_A1': b'KEY_HOME',
+    b'KEY_C1': b'KEY_END',
+    # windows-curses: map to our M- names
+    b'ALT_U': b'M-u',
+    # windows-curses: arguably these names are better than the xterm names
+    b'CTL_UP': b'kUP5',
+    b'CTL_DOWN': b'kDN5',
+    b'CTL_RIGHT': b'kRIT5',
+    b'CTL_LEFT': b'kLFT5',
+    b'CTL_HOME': b'kHOM5',
+    b'CTL_END': b'kEND5',
+    b'ALT_RIGHT': b'kRIT3',
+    b'ALT_LEFT': b'kLFT3',
+    b'ALT_E': b'M-e',
+    # windows-curses: idk why these are different
+    b'KEY_SUP': b'KEY_SR',
+    b'KEY_SDOWN': b'KEY_SF',
+    # macos: (sends this for backspace key, others interpret this as well)
+    b'^?': b'KEY_BACKSPACE',
+    # linux, perhaps others
+    b'^H': b'KEY_BACKSPACE',  # ^Backspace on my keyboard
+    b'PADENTER': b'^M',  # Enter on numpad
+}
 
 
 class Key(NamedTuple):
@@ -70,15 +107,16 @@ class Screen:
             self,
             stdscr: 'curses._CursesWindow',
             filenames: List[Optional[str]],
+            initial_lines: List[int],
             perf: Perf,
     ) -> None:
         self.stdscr = stdscr
-        color_manager = ColorManager.make()
+        self.color_manager = ColorManager.make()
-        hl_factories = (
+        self.hl_factories = (Syntax.from_screen(stdscr, self.color_manager),)
-            Syntax.from_screen(stdscr, color_manager),
+        self.files = [
-            TrailingWhitespace(color_manager),
+            File(filename, line, self.color_manager, self.hl_factories)
-        )
+            for filename, line in zip(filenames, initial_lines)
-        self.files = [File(f, hl_factories) for f in filenames]
+        ]
         self.i = 0
         self.history = History()
         self.perf = perf
@@ -86,7 +124,6 @@ class Screen:
         self.margin = Margin.from_current_screen()
         self.cut_buffer: Tuple[str, ...] = ()
         self.cut_selection = False
-        self._resize_cb: Optional[Callable[[], None]] = None
         self._buffered_input: Union[int, str, None] = None
 
     @property
@@ -103,7 +140,7 @@ class Screen:
         else:
             files = ''
             version_width = len(VERSION_STR) + 2
-        centered = filename.center(curses.COLS)[version_width:]
+        centered = filename.center(self.margin.cols)[version_width:]
         s = f' {VERSION_STR} {files}{centered}{files}'
         self.stdscr.insstr(0, 0, s, curses.A_REVERSE)
 
@@ -189,7 +226,10 @@ class Screen:
         if self._buffered_input is not None:
             wch, self._buffered_input = self._buffered_input, None
         else:
-            wch = self.stdscr.get_wch()
+            try:
+                wch = self.stdscr.get_wch()
+            except curses.error:  # pragma: no cover (macos bug?)
+                wch = self.stdscr.get_wch()
         if isinstance(wch, str) and wch == '\x1b':
             wch = self._get_sequence(wch)
             if len(wch) == 2:
@@ -200,12 +240,10 @@ class Screen:
         elif isinstance(wch, str) and wch.isprintable():
             wch = self._get_string(wch)
             return Key(wch, b'STRING')
-        elif wch == '\x7f':  # pragma: no cover (macos)
-            keyname = curses.keyname(curses.KEY_BACKSPACE)
-            return Key(wch, keyname)
 
         key = wch if isinstance(wch, int) else ord(wch)
         keyname = curses.keyname(key)
+        keyname = KEYNAME_REWRITE.get(keyname, keyname)
         return Key(wch, keyname)
 
     def get_char(self) -> Key:
@@ -220,31 +258,48 @@ class Screen:
         self.file.draw(self.stdscr, self.margin)
         self.status.draw(self.stdscr, self.margin)
 
-    @contextlib.contextmanager
-    def resize_cb(self, f: Callable[[], None]) -> Generator[None, None, None]:
-        assert self._resize_cb is None, self._resize_cb
-        self._resize_cb = f
-        try:
-            yield
-        finally:
-            self._resize_cb = None
-
     def resize(self) -> None:
         curses.update_lines_cols()
         self.margin = Margin.from_current_screen()
-        self.file.scroll_screen_if_needed(self.margin)
+        self.file.buf.scroll_screen_if_needed(self.margin)
         self.draw()
-        if self._resize_cb is not None:
-            self._resize_cb()
 
-    def quick_prompt(self, prompt: str, opts: str) -> Union[str, PromptResult]:
+    def quick_prompt(
+            self,
+            prompt: str,
+            opt_strs: Tuple[str, ...],
+    ) -> Union[str, PromptResult]:
+        opts = [opt[0] for opt in opt_strs]
         while True:
-            s = prompt.ljust(curses.COLS)
+            x = 0
-            if len(s) > curses.COLS:
+            prompt_line = self.margin.lines - 1
-                s = f'{s[:curses.COLS - 1]}…'
+
-            self.stdscr.insstr(curses.LINES - 1, 0, s, curses.A_REVERSE)
+            def _write(s: str, *, attr: int = curses.A_REVERSE) -> None:
-            x = min(curses.COLS - 1, len(prompt) + 1)
+                nonlocal x
-            self.stdscr.move(curses.LINES - 1, x)
+
+                if x >= self.margin.cols:
+                    return
+                self.stdscr.insstr(prompt_line, x, s, attr)
+                x += len(s)
+
+            _write(prompt)
+            _write(' [')
+            for i, opt_str in enumerate(opt_strs):
+                _write(opt_str[0], attr=curses.A_REVERSE | curses.A_BOLD)
+                _write(opt_str[1:])
+                if i != len(opt_strs) - 1:
+                    _write(', ')
+            _write(']?')
+
+            if x < self.margin.cols - 1:
+                s = ' ' * (self.margin.cols - x)
+                self.stdscr.insstr(prompt_line, x, s, curses.A_REVERSE)
+                x += 1
+            else:
+                x = self.margin.cols - 1
+                self.stdscr.insstr(prompt_line, x, '…', curses.A_REVERSE)
+
+            self.stdscr.move(prompt_line, x)
 
             key = self.get_char()
             if key.keyname == b'KEY_RESIZE':
@@ -299,14 +354,14 @@ class Screen:
                 self.file.go_to_line(lineno, self.margin)
 
     def current_position(self) -> None:
-        line = f'line {self.file.y + 1}'
+        line = f'line {self.file.buf.y + 1}'
-        col = f'col {self.file.x + 1}'
+        col = f'col {self.file.buf.x + 1}'
-        line_count = max(len(self.file.lines) - 1, 1)
+        line_count = max(len(self.file.buf) - 1, 1)
         lines_word = 'line' if line_count == 1 else 'lines'
         self.status.update(f'{line}, {col} (of {line_count} {lines_word})')
 
     def cut(self) -> None:
-        if self.file.select_start:
+        if self.file.selection.start:
             self.cut_buffer = self.file.cut_selection(self.margin)
             self.cut_selection = True
         else:
@@ -340,8 +395,9 @@ class Screen:
         else:
             action = from_stack.pop()
             to_stack.append(action.apply(self.file))
-            self.file.scroll_screen_if_needed(self.margin)
+            self.file.buf.scroll_screen_if_needed(self.margin)
             self.status.update(f'{op}: {action.name}')
+            self.file.selection.clear()
 
     def undo(self) -> None:
         self._undo_redo('undo', self.file.undo_stack, self.file.redo_stack)
@@ -365,7 +421,11 @@ class Screen:
 
     def command(self) -> Optional[EditResult]:
         response = self.prompt('', history='command')
-        if response == ':q':
+        if response is PromptResult.CANCELLED:
+            pass
+        elif response == ':q':
+            return self.quit_save_modified()
+        elif response == ':q!':
             return EditResult.EXIT
         elif response == ':w':
             self.save()
@@ -373,12 +433,38 @@ class Screen:
             self.save()
             return EditResult.EXIT
         elif response == ':sort':
-            if self.file.select_start:
+            if self.file.selection.start:
                 self.file.sort_selection(self.margin)
             else:
                 self.file.sort(self.margin)
             self.status.update('sorted!')
-        elif response is not PromptResult.CANCELLED:
+        elif response == ':sort!':
+            if self.file.selection.start:
+                self.file.sort_selection(self.margin, reverse=True)
+            else:
+                self.file.sort(self.margin, reverse=True)
+            self.status.update('sorted!')
+        elif response.startswith((':tabstop ', ':tabsize ')):
+            _, _, tab_size = response.partition(' ')
+            try:
+                parsed_tab_size = int(tab_size)
+            except ValueError:
+                self.status.update(f'invalid size: {tab_size}')
+            else:
+                if parsed_tab_size <= 0:
+                    self.status.update(f'invalid size: {parsed_tab_size}')
+                else:
+                    for file in self.files:
+                        file.buf.set_tab_size(parsed_tab_size)
+                    self.status.update('updated!')
+        elif response == ':comment' or response.startswith(':comment '):
+            _, _, comment = response.partition(' ')
+            comment = (comment or '#').strip()
+            if self.file.selection.start:
+                self.file.toggle_comment_selection(comment)
+            else:
+                self.file.toggle_comment(comment)
+        else:
             self.status.update(f'invalid command: {response}')
         return None
 
@@ -397,12 +483,12 @@ class Screen:
                 self.file.filename = filename
 
         if os.path.isfile(self.file.filename):
-            with open(self.file.filename) as f:
+            with open(self.file.filename, newline='') as f:
                 *_, sha256 = get_lines(f)
         else:
             sha256 = hashlib.sha256(b'').hexdigest()
 
-        contents = self.file.nl.join(self.file.lines)
+        contents = self.file.nl.join(self.file.buf)
         sha256_to_save = hashlib.sha256(contents.encode()).hexdigest()
 
         # the file on disk is the same as when we opened it
@@ -410,12 +496,12 @@ class Screen:
             self.status.update('(file changed on disk, not implemented)')
             return PromptResult.CANCELLED
 
-        with open(self.file.filename, 'w') as f:
+        with open(self.file.filename, 'w', newline='') as f:
             f.write(contents)
 
         self.file.modified = False
         self.file.sha256 = sha256_to_save
-        num_lines = len(self.file.lines) - 1
+        num_lines = len(self.file.buf) - 1
         lines = 'lines' if num_lines != 1 else 'line'
         self.status.update(f'saved! ({num_lines} {lines} written)')
 
@@ -436,10 +522,19 @@ class Screen:
             self.file.filename = response
             return self.save()
 
+    def open_file(self) -> Optional[EditResult]:
+        response = self.prompt('enter filename', history='open')
+        if response is not PromptResult.CANCELLED:
+            opened = File(response, 0, self.color_manager, self.hl_factories)
+            self.files.append(opened)
+            return EditResult.OPEN
+        else:
+            return None
+
     def quit_save_modified(self) -> Optional[EditResult]:
         if self.file.modified:
             response = self.quick_prompt(
-                'file is modified - save [y(es), n(o)]?', 'yn',
+                'file is modified - save', ('yes', 'no'),
             )
             if response == 'y':
                 if self.save_filename() is not PromptResult.CANCELLED:
@@ -454,10 +549,13 @@ class Screen:
         return EditResult.EXIT
 
     def background(self) -> None:
-        curses.endwin()
+        if sys.platform == 'win32':  # pragma: win32 cover
-        os.kill(os.getpid(), signal.SIGSTOP)
+            self.status.update('cannot run babi in background on Windows')
-        self.stdscr = _init_screen()
+        else:  # pragma: win32 no cover
-        self.resize()
+            curses.endwin()
+            os.kill(os.getpid(), signal.SIGSTOP)
+            self.stdscr = _init_screen()
+            self.resize()
 
     DISPATCH = {
         b'KEY_RESIZE': resize,
@@ -467,12 +565,14 @@ class Screen:
         b'^U': uncut,
         b'M-u': undo,
         b'M-U': redo,
+        b'M-e': redo,
         b'^W': search,
         b'^\\': replace,
         b'^[': command,
         b'^S': save,
         b'^O': save_filename,
         b'^X': quit_save_modified,
+        b'^P': open_file,
         b'kLFT3': lambda screen: EditResult.PREV,
         b'kRIT3': lambda screen: EditResult.NEXT,
         b'^Z': background,

babi/status.py

@@ -18,14 +18,14 @@ class Status:
 
     def draw(self, stdscr: 'curses._CursesWindow', margin: Margin) -> None:
         if margin.footer or self._status:
-            stdscr.insstr(curses.LINES - 1, 0, ' ' * curses.COLS)
+            stdscr.insstr(margin.lines - 1, 0, ' ' * margin.cols)
             if self._status:
                 status = f' {self._status} '
-                x = (curses.COLS - len(status)) // 2
+                x = (margin.cols - len(status)) // 2
                 if x < 0:
                     x = 0
                     status = status.strip()
-                stdscr.insstr(curses.LINES - 1, x, status, curses.A_REVERSE)
+                stdscr.insstr(margin.lines - 1, x, status, curses.A_REVERSE)
 
     def tick(self, margin: Margin) -> None:
         # when the window is only 1-tall, hide the status quicker

babi/textmate_demo.py

@@ -0,0 +1,69 @@
+import argparse
+from typing import Optional
+from typing import Sequence
+
+from babi.highlight import Compiler
+from babi.highlight import Grammars
+from babi.highlight import highlight_line
+from babi.theme import Style
+from babi.theme import Theme
+from babi.user_data import prefix_data
+from babi.user_data import xdg_config
+
+
+def print_styled(s: str, style: Style) -> None:
+    color_s = ''
+    undo_s = ''
+    if style.fg is not None:
+        color_s += '\x1b[38;2;{r};{g};{b}m'.format(**style.fg._asdict())
+        undo_s += '\x1b[39m'
+    if style.bg is not None:
+        color_s += '\x1b[48;2;{r};{g};{b}m'.format(**style.bg._asdict())
+        undo_s += '\x1b[49m'
+    if style.b:
+        color_s += '\x1b[1m'
+        undo_s += '\x1b[22m'
+    if style.i:
+        color_s += '\x1b[3m'
+        undo_s += '\x1b[23m'
+    if style.u:
+        color_s += '\x1b[4m'
+        undo_s += '\x1b[24m'
+    print(f'{color_s}{s}{undo_s}', end='', flush=True)
+
+
+def _highlight_output(theme: Theme, compiler: Compiler, filename: str) -> int:
+    state = compiler.root_state
+
+    if theme.default.bg is not None:
+        print('\x1b[48;2;{r};{g};{b}m'.format(**theme.default.bg._asdict()))
+    with open(filename) as f:
+        for line_idx, line in enumerate(f):
+            first_line = line_idx == 0
+            state, regions = highlight_line(compiler, state, line, first_line)
+            for start, end, scope in regions:
+                print_styled(line[start:end], theme.select(scope))
+    print('\x1b[m', end='')
+    return 0
+
+
+def main(argv: Optional[Sequence[str]] = None) -> int:
+    parser = argparse.ArgumentParser()
+    parser.add_argument('--theme', default=xdg_config('theme.json'))
+    parser.add_argument('--grammar-dir', default=prefix_data('grammar_v1'))
+    parser.add_argument('filename')
+    args = parser.parse_args(argv)
+
+    with open(args.filename) as f:
+        first_line = next(f, '')
+
+    theme = Theme.from_filename(args.theme)
+
+    grammars = Grammars(args.grammar_dir)
+    compiler = grammars.compiler_for_file(args.filename, first_line)
+
+    return _highlight_output(theme, compiler, args.filename)
+
+
+if __name__ == '__main__':
+    exit(main())

babi/theme.py

@@ -1,7 +1,6 @@
 import functools
 import json
 import os.path
-import re
 from typing import Any
 from typing import Dict
 from typing import NamedTuple
@@ -12,9 +11,6 @@ from babi._types import Protocol
 from babi.color import Color
 from babi.fdict import FDict
 
-# yes I know this is wrong, but it's good enough for now
-UN_COMMENT = re.compile(r'^\s*//.*$', re.MULTILINE)
-
 
 class Style(NamedTuple):
     fg: Optional[Color]
@@ -114,10 +110,14 @@ class Theme(NamedTuple):
         for rule in rules:
             if 'scope' not in rule:
                 scopes = ['']
+            elif rule['scope'] == '':
+                scopes = ['']
             elif isinstance(rule['scope'], str):
                 scopes = [
-                    # some themes have a buggy trailing comma
+                    s.strip()
-                    s.strip() for s in rule['scope'].strip(',').split(',')
+                    # some themes have a buggy trailing/leading comma
+                    for s in rule['scope'].strip().strip(',').split(',')
+                    if s.strip()
                 ]
             else:
                 scopes = rule['scope']
@@ -148,5 +148,4 @@ class Theme(NamedTuple):
             return cls.blank()
         else:
             with open(filename) as f:
-                contents = UN_COMMENT.sub('', f.read())
+                return cls.from_dct(json.load(f))
-                return cls.from_dct(json.loads(contents))

babi/user_data.py

@@ -1,4 +1,5 @@
 import os.path
+import sys
 
 
 def _xdg(*path: str, env: str, default: str) -> str:
@@ -14,3 +15,7 @@ def xdg_data(*path: str) -> str:
 
 def xdg_config(*path: str) -> str:
     return _xdg(*path, env='XDG_CONFIG_HOME', default='~/.config')
+
+
+def prefix_data(*path: str) -> str:
+    return os.path.join(sys.prefix, 'share/babi', *path)

bin/download-syntax

@@ -1,81 +0,0 @@
-#!/usr/bin/env python3
-import argparse
-import enum
-import json
-import os.path
-import plistlib
-import urllib.request
-from typing import NamedTuple
-
-import cson  # pip install cson
-
-
-DEFAULT_DIR = os.path.join(
-    os.environ.get('XDG_DATA_HOME') or
-    os.path.expanduser('~/.local/share'),
-    'babi/textmate_syntax',
-)
-Ext = enum.Enum('Ext', 'CSON PLIST JSON')
-
-
-def _convert_cson(src: bytes) -> str:
-    return json.dumps(cson.loads(src))
-
-
-def _convert_json(src: bytes) -> str:
-    return json.dumps(json.loads(src))
-
-
-def _convert_plist(src: bytes) -> str:
-    return json.dumps(plistlib.loads(src))
-
-
-EXT_CONVERT = {
-    Ext.CSON: _convert_cson,
-    Ext.JSON: _convert_json,
-    Ext.PLIST: _convert_plist,
-}
-
-
-class Syntax(NamedTuple):
-    name: str
-    ext: Ext
-    url: str
-
-
-SYNTAXES = (
-    Syntax('c', Ext.JSON, 'https://raw.githubusercontent.com/jeff-hykin/cpp-textmate-grammar/53e39b1c/syntaxes/c.tmLanguage.json'),  # noqa: E501
-    Syntax('css', Ext.CSON, 'https://raw.githubusercontent.com/atom/language-css/9feb69c081308b63f78bb0d6a2af2ff5eb7d869b/grammars/css.cson'),  # noqa: E501
-    Syntax('diff', Ext.PLIST, 'https://raw.githubusercontent.com/textmate/diff.tmbundle/0593bb77/Syntaxes/Diff.plist'),  # noqa: E501
-    Syntax('html', Ext.PLIST, 'https://raw.githubusercontent.com/textmate/html.tmbundle/0c3d5ee5/Syntaxes/HTML.plist'),  # noqa: E501
-    Syntax('html-derivative', Ext.PLIST, 'https://raw.githubusercontent.com/textmate/html.tmbundle/0c3d5ee54de3a993f747f54186b73a4d2d3c44a2/Syntaxes/HTML%20(Derivative).tmLanguage'),  # noqa: E501
-    Syntax('ini', Ext.PLIST, 'https://raw.githubusercontent.com/textmate/ini.tmbundle/7d8c7b55/Syntaxes/Ini.plist'),  # noqa: E501
-    Syntax('json', Ext.PLIST, 'https://raw.githubusercontent.com/microsoft/vscode-JSON.tmLanguage/d113e90937ed3ecc31ac54750aac2e8efa08d784/JSON.tmLanguage'),  # noqa: E501
-    Syntax('markdown', Ext.PLIST, 'https://raw.githubusercontent.com/microsoft/vscode-markdown-tm-grammar/59a5962/syntaxes/markdown.tmLanguage'),  # noqa: E501
-    Syntax('powershell', Ext.PLIST, 'https://raw.githubusercontent.com/PowerShell/EditorSyntax/4a0a0766/PowerShellSyntax.tmLanguage'),  # noqa: E501
-    Syntax('python', Ext.PLIST, 'https://raw.githubusercontent.com/MagicStack/MagicPython/c9b3409d/grammars/MagicPython.tmLanguage'),  # noqa: E501
-    # TODO: https://github.com/zargony/atom-language-rust/pull/149
-    Syntax('rust', Ext.CSON, 'https://raw.githubusercontent.com/asottile/atom-language-rust/e113ca67/grammars/rust.cson'),  # noqa: E501
-    Syntax('shell', Ext.CSON, 'https://raw.githubusercontent.com/atom/language-shellscript/7008ea926867d8a231003e78094091471c4fccf8/grammars/shell-unix-bash.cson'),  # noqa: E501
-    Syntax('yaml', Ext.PLIST, 'https://raw.githubusercontent.com/textmate/yaml.tmbundle/e54ceae3/Syntaxes/YAML.tmLanguage'),  # noqa: E501
-)
-
-
-def main() -> int:
-    parser = argparse.ArgumentParser()
-    parser.add_argument('--dest', default=DEFAULT_DIR)
-    args = parser.parse_args()
-
-    os.makedirs(args.dest, exist_ok=True)
-    for syntax in SYNTAXES:
-        print(f'downloading {syntax.name}...')
-        resp = urllib.request.urlopen(syntax.url).read()
-        converted = EXT_CONVERT[syntax.ext](resp)
-        with open(os.path.join(args.dest, f'{syntax.name}.json'), 'w') as f:
-            f.write(converted)
-
-    return 0
-
-
-if __name__ == '__main__':
-    exit(main())

bin/download-theme

@@ -0,0 +1,88 @@
+#!/usr/bin/env python3
+import argparse
+import io
+import json
+import os.path
+import plistlib
+import re
+import urllib.request
+from typing import Any
+
+import cson  # pip install cson
+
+TOKEN = re.compile(br'(\\\\|\\"|"|//|\n)')
+
+
+def json_with_comments(s: bytes) -> Any:
+    bio = io.BytesIO()
+
+    idx = 0
+    in_string = False
+    in_comment = False
+
+    match = TOKEN.search(s, idx)
+    while match:
+        if not in_comment:
+            bio.write(s[idx:match.start()])
+
+        tok = match[0]
+        if not in_comment and tok == b'"':
+            in_string = not in_string
+        elif in_comment and tok == b'\n':
+            in_comment = False
+        elif not in_string and tok == b'//':
+            in_comment = True
+
+        if not in_comment:
+            bio.write(tok)
+
+        idx = match.end()
+        match = TOKEN.search(s, idx)
+
+    bio.seek(0)
+    return json.load(bio)
+
+
+STRATEGIES = (json.loads, plistlib.loads, cson.loads, json_with_comments)
+
+
+def main() -> int:
+    parser = argparse.ArgumentParser()
+    parser.add_argument('name')
+    parser.add_argument('url')
+    args = parser.parse_args()
+
+    if '/blob/' in args.url:
+        url = args.url.replace('/blob/', '/raw/')
+    else:
+        url = args.url
+
+    contents = urllib.request.urlopen(url).read()
+
+    errors = []
+    for strategy in STRATEGIES:
+        try:
+            loaded = strategy(contents)
+        except Exception as e:
+            errors.append((f'{strategy.__module__}.{strategy.__name__}', e))
+        else:
+            break
+    else:
+        errors_s = '\n'.join(f'\t{name}: {error}' for name, error in errors)
+        raise AssertionError(f'could not load as json/plist/cson:\n{errors_s}')
+
+    config_dir = os.path.expanduser('~/.config/babi')
+    os.makedirs(config_dir, exist_ok=True)
+    dest = os.path.join(config_dir, f'{args.name}.json')
+    with open(dest, 'w') as f:
+        json.dump(loaded, f)
+
+    theme_json = os.path.join(config_dir, 'theme.json')
+    if os.path.lexists(theme_json):
+        os.remove(theme_json)
+    os.symlink(dest, theme_json)
+    return 0
+
+
+if __name__ == '__main__':
+    exit(main())

requirements-dev.txt

@@ -1,5 +1,5 @@
 covdefaults
 coverage
-git+https://github.com/asottile/hecate@ebe6dfb
+git+https://github.com/asottile/hecate@875567f
 pytest
 remote-pdb

setup.cfg

@@ -1,6 +1,6 @@
 [metadata]
 name = babi
-version = 0.0.1
+version = 0.0.14
 description = a text editor
 long_description = file: README.md
 long_description_content_type = text/markdown
@@ -22,12 +22,17 @@ classifiers =
 [options]
 packages = find:
 install_requires =
-    onigurumacffi>=0.0.10
+    babi-grammars
+    identify
+    onigurumacffi>=0.0.18
+    importlib_metadata>=1;python_version<"3.8"
+    windows-curses;sys_platform=="win32"
 python_requires = >=3.6.1
 
 [options.entry_points]
 console_scripts =
     babi = babi.main:main
+    babi-textmate-demo = babi.textmate_demo:main
 
 [options.packages.find]
 exclude =

testing/vsc_test/.gitignore

@@ -0,0 +1,2 @@
+/node_modules
+/package-lock.json

testing/vsc_test/package.json

@@ -0,0 +1,5 @@
+{
+  "dependencies": [
+    "vscode-textmate"
+  ]
+}

testing/vsc_test/vsc.js

@@ -0,0 +1,51 @@
+const fs = require('fs');
+const vsctm = require('vscode-textmate');
+
+if (process.argv.length < 4) {
+    console.log('usage: t.js GRAMMAR FILE');
+    process.exit(1);
+}
+
+const grammar = process.argv[2];
+const file = process.argv[3];
+
+const scope = JSON.parse(fs.readFileSync(grammar, {encoding: 'UTF-8'})).scopeName;
+
+/**
+ * Utility to read a file as a promise
+ */
+function readFile(path) {
+    return new Promise((resolve, reject) => {
+        fs.readFile(path, (error, data) => error ? reject(error) : resolve(data));
+    })
+}
+
+// Create a registry that can create a grammar from a scope name.
+const registry = new vsctm.Registry({
+    loadGrammar: (scopeName) => {
+        if (scopeName === scope) {
+            return readFile(grammar).then(data => vsctm.parseRawGrammar(data.toString(), grammar))
+        }
+        console.log(`Unknown scope name: ${scopeName}`);
+        return null;
+    }
+});
+
+// Load the JavaScript grammar and any other grammars included by it async.
+registry.loadGrammar(scope).then(grammar => {
+    const text = fs.readFileSync(file, {encoding: 'UTF-8'}).trimEnd('\n').split(/\n/);
+    let ruleStack = vsctm.INITIAL;
+    for (let i = 0; i < text.length; i++) {
+        const line = text[i];
+        const lineTokens = grammar.tokenizeLine(line, ruleStack);
+        console.log(`\nTokenizing line: ${line}`);
+        for (let j = 0; j < lineTokens.tokens.length; j++) {
+            const token = lineTokens.tokens[j];
+            console.log(` - token from ${token.startIndex} to ${token.endIndex} ` +
+              `(${line.substring(token.startIndex, token.endIndex)}) ` +
+              `with scopes ${token.scopes.join(', ')}`
+            );
+        }
+        ruleStack = lineTokens.ruleStack;
+    }
+});

tests/buf_test.py

@@ -0,0 +1,178 @@
+import pytest
+
+from babi.buf import Buf
+
+
+def test_buf_repr():
+    ret = repr(Buf(['a', 'b', 'c']))
+    assert ret == "Buf(['a', 'b', 'c'], x=0, y=0, file_y=0)"
+
+
+def test_buf_item_retrieval():
+    buf = Buf(['a', 'b', 'c'])
+    assert buf[1] == 'b'
+    assert buf[-1] == 'c'
+    with pytest.raises(IndexError):
+        buf[3]
+
+
+def test_buf_del():
+    lst = ['a', 'b', 'c']
+
+    buf = Buf(lst)
+
+    with buf.record() as modifications:
+        del buf[1]
+
+    assert lst == ['a', 'c']
+
+    buf.apply(modifications)
+
+    assert lst == ['a', 'b', 'c']
+
+
+def test_buf_del_with_negative():
+    lst = ['a', 'b', 'c']
+
+    buf = Buf(lst)
+
+    with buf.record() as modifications:
+        del buf[-1]
+
+    assert lst == ['a', 'b']
+
+    buf.apply(modifications)
+
+    assert lst == ['a', 'b', 'c']
+
+
+def test_buf_insert():
+    lst = ['a', 'b', 'c']
+
+    buf = Buf(lst)
+
+    with buf.record() as modifications:
+        buf.insert(1, 'q')
+
+    assert lst == ['a', 'q', 'b', 'c']
+
+    buf.apply(modifications)
+
+    assert lst == ['a', 'b', 'c']
+
+
+def test_buf_insert_with_negative():
+    lst = ['a', 'b', 'c']
+
+    buf = Buf(lst)
+
+    with buf.record() as modifications:
+        buf.insert(-1, 'q')
+
+    assert lst == ['a', 'b', 'q', 'c']
+
+    buf.apply(modifications)
+
+    assert lst == ['a', 'b', 'c']
+
+
+def test_buf_set_value():
+    lst = ['a', 'b', 'c']
+
+    buf = Buf(lst)
+
+    with buf.record() as modifications:
+        buf[1] = 'hello'
+
+    assert lst == ['a', 'hello', 'c']
+
+    buf.apply(modifications)
+
+    assert lst == ['a', 'b', 'c']
+
+
+def test_buf_set_value_idx_negative():
+    lst = ['a', 'b', 'c']
+
+    buf = Buf(lst)
+
+    with buf.record() as modifications:
+        buf[-1] = 'hello'
+
+    assert lst == ['a', 'b', 'hello']
+
+    buf.apply(modifications)
+
+    assert lst == ['a', 'b', 'c']
+
+
+def test_buf_multiple_modifications():
+    lst = ['a', 'b', 'c']
+
+    buf = Buf(lst)
+
+    with buf.record() as modifications:
+        buf[1] = 'hello'
+        buf.insert(1, 'ohai')
+        del buf[0]
+
+    assert lst == ['ohai', 'hello', 'c']
+
+    buf.apply(modifications)
+
+    assert lst == ['a', 'b', 'c']
+
+
+def test_buf_iter():
+    buf = Buf(['a', 'b', 'c'])
+    buf_iter = iter(buf)
+    assert next(buf_iter) == 'a'
+    assert next(buf_iter) == 'b'
+    assert next(buf_iter) == 'c'
+    with pytest.raises(StopIteration):
+        next(buf_iter)
+
+
+def test_buf_append():
+    lst = ['a', 'b', 'c']
+
+    buf = Buf(lst)
+
+    with buf.record() as modifications:
+        buf.append('q')
+
+    assert lst == ['a', 'b', 'c', 'q']
+
+    buf.apply(modifications)
+
+    assert lst == ['a', 'b', 'c']
+
+
+def test_buf_pop_default():
+    lst = ['a', 'b', 'c']
+
+    buf = Buf(lst)
+
+    with buf.record() as modifications:
+        buf.pop()
+
+    assert lst == ['a', 'b']
+
+    buf.apply(modifications)
+
+    assert lst == ['a', 'b', 'c']
+
+
+def test_buf_pop_idx():
+    lst = ['a', 'b', 'c']
+
+    buf = Buf(lst)
+
+    with buf.record() as modifications:
+        buf.pop(1)
+
+    assert lst == ['a', 'c']
+
+    buf.apply(modifications)
+
+    assert lst == ['a', 'b', 'c']

tests/color_test.py

@@ -0,0 +1,16 @@
+import pytest
+
+from babi.color import Color
+
+
+@pytest.mark.parametrize(
+    ('s', 'expected'),
+    (
+        ('#1e77d3', Color(0x1e, 0x77, 0xd3)),
+        ('white', Color(0xff, 0xff, 0xff)),
+        ('black', Color(0x00, 0x00, 0x00)),
+        ('#ccc', Color(0xcc, 0xcc, 0xcc)),
+    ),
+)
+def test_color_parse(s, expected):
+    assert Color.parse(s) == expected

tests/conftest.py

@@ -0,0 +1,17 @@
+import json
+
+import pytest
+
+from babi.highlight import Grammars
+
+
+@pytest.fixture
+def make_grammars(tmpdir):
+    grammar_dir = tmpdir.join('grammars').ensure_dir()
+
+    def make_grammars(*grammar_dcts):
+        for grammar in grammar_dcts:
+            filename = f'{grammar["scopeName"]}.json'
+            grammar_dir.join(filename).write(json.dumps(grammar))
+        return Grammars(grammar_dir)
+    return make_grammars

tests/fdict_test.py

@@ -1,3 +1,6 @@
+import pytest
+
+from babi.fdict import FChainMap
 from babi.fdict import FDict
 
 
@@ -5,3 +8,21 @@ def test_fdict_repr():
     # mostly because this shouldn't get hit elsewhere but is uesful for
     # debugging purposes
     assert repr(FDict({1: 2, 3: 4})) == 'FDict({1: 2, 3: 4})'
+
+
+def test_f_chain_map():
+    chain_map = FChainMap({1: 2}, {3: 4}, FDict({1: 5}))
+    assert chain_map[1] == 5
+    assert chain_map[3] == 4
+
+    with pytest.raises(KeyError) as excinfo:
+        chain_map[2]
+    k, = excinfo.value.args
+    assert k == 2
+
+
+def test_f_chain_map_extend():
+    chain_map = FChainMap({1: 2})
+    assert chain_map[1] == 2
+    chain_map = FChainMap(chain_map, {1: 5})
+    assert chain_map[1] == 5

tests/features/comment_test.py

@@ -0,0 +1,75 @@
+from testing.runner import and_exit
+from testing.runner import trigger_command_mode
+
+
+def test_comment_some_code(run, ten_lines):
+    with run(str(ten_lines)) as h, and_exit(h):
+        h.press('S-Down')
+
+        trigger_command_mode(h)
+        h.press_and_enter(':comment')
+
+        h.await_text('# line_0\n# line_1\nline_2\n')
+
+
+def test_comment_some_code_with_alternate_comment_character(run, ten_lines):
+    with run(str(ten_lines)) as h, and_exit(h):
+        h.press('S-Down')
+
+        trigger_command_mode(h)
+        h.press_and_enter(':comment //')
+
+        h.await_text('// line_0\n// line_1\nline_2\n')
+
+
+def test_comment_partially_commented(run, ten_lines):
+    with run(str(ten_lines)) as h, and_exit(h):
+        h.press('#')
+        h.press('S-Down')
+        h.await_text('#line_0\nline_1\nline_2')
+
+        trigger_command_mode(h)
+        h.press_and_enter(':comment')
+
+        h.await_text('line_0\nline_1\nline_2\n')
+
+
+def test_comment_partially_uncommented(run, ten_lines):
+    with run(str(ten_lines)) as h, and_exit(h):
+        h.press('Down')
+        h.press('#')
+        h.press('Up')
+        h.press('S-Down')
+        h.await_text('line_0\n#line_1\nline_2')
+
+        trigger_command_mode(h)
+        h.press_and_enter(':comment')
+
+        h.await_text('# line_0\n# #line_1\nline_2\n')
+
+
+def test_comment_single_line(run, ten_lines):
+    with run(str(ten_lines)) as h, and_exit(h):
+        trigger_command_mode(h)
+        h.press_and_enter(':comment')
+
+        h.await_text('# line_0\nline_1\n')
+
+
+def test_uncomment_single_line(run, ten_lines):
+    with run(str(ten_lines)) as h, and_exit(h):
+        h.press('#')
+        h.await_text('#line_0\nline_1\n')
+
+        trigger_command_mode(h)
+        h.press_and_enter(':comment')
+
+        h.await_text('line_0\nline_1\n')
+
+
+def test_comment_with_trailing_whitespace(run, ten_lines):
+    with run(str(ten_lines)) as h, and_exit(h):
+        trigger_command_mode(h)
+        h.press_and_enter(':comment //   ')
+
+        h.await_text('// line_0\nline_1\n')

tests/features/conftest.py

@@ -16,6 +16,13 @@ from babi.screen import VERSION_STR
 from testing.runner import PrintsErrorRunner
 
 
+@pytest.fixture(autouse=True)
+def prefix_home(tmpdir):
+    prefix_home = tmpdir.join('prefix_home')
+    with mock.patch.object(sys, 'prefix', str(prefix_home)):
+        yield prefix_home
+
+
 @pytest.fixture(autouse=True)
 def xdg_data_home(tmpdir):
     data_home = tmpdir.join('data_home')
@@ -39,7 +46,6 @@ def ten_lines(tmpdir):
 
 class Screen:
     def __init__(self, width, height):
-        self.disabled = True
         self.nodelay = False
         self.width = width
         self.height = height
@@ -57,6 +63,16 @@ class Screen:
             self._prev_screenshot = ret
         return ret
 
+    def addstr(self, y, x, s, attr):
+        self.lines[y] = self.lines[y][:x] + s + self.lines[y][x + len(s):]
+
+        line_attr = self.attrs[y]
+        new = [attr] * len(s)
+        self.attrs[y] = line_attr[:x] + new + line_attr[x + len(s):]
+
+        self.y = y
+        self.x = x + len(s)
+
     def insstr(self, y, x, s, attr):
         line = self.lines[y]
         self.lines[y] = (line[:x] + s + line[x:])[:self.width]
@@ -66,6 +82,7 @@ class Screen:
         self.attrs[y] = (line_attr[:x] + new + line_attr[x:])[:self.width]
 
     def chgat(self, y, x, n, attr):
+        assert n >= 0  # TODO: switch to > 0, we should never do 0-length
         self.attrs[y][x:x + n] = [attr] * n
 
     def move(self, y, x):
@@ -166,7 +183,8 @@ class CursesError(NamedTuple):
 
 
 class CursesScreen:
-    def __init__(self, runner):
+    def __init__(self, screen, runner):
+        self._screen = screen
         self._runner = runner
         self._bkgd_attr = (-1, -1, 0)
 
@@ -190,20 +208,26 @@ class CursesScreen:
         pass
 
     def nodelay(self, val):
-        self._runner.screen.nodelay = val
+        self._screen.nodelay = val
+
+    def addstr(self, y, x, s, attr=0):
+        self._screen.addstr(y, x, s, self._to_attr(attr))
 
     def insstr(self, y, x, s, attr=0):
-        self._runner.screen.insstr(y, x, s, self._to_attr(attr))
+        self._screen.insstr(y, x, s, self._to_attr(attr))
 
     def clrtoeol(self):
-        s = self._runner.screen.width * ' '
+        s = self._screen.width * ' '
-        self.insstr(self._runner.screen.y, self._runner.screen.x, s)
+        self.insstr(self._screen.y, self._screen.x, s)
 
     def chgat(self, y, x, n, attr):
-        self._runner.screen.chgat(y, x, n, self._to_attr(attr))
+        self._screen.chgat(y, x, n, self._to_attr(attr))
 
     def move(self, y, x):
-        self._runner.screen.move(y, x)
+        self._screen.move(y, x)
+
+    def getyx(self):
+        return self._screen.y, self._screen.x
 
     def get_wch(self):
         return self._runner._get_wch()
@@ -254,6 +278,7 @@ KEYS = [
     Key('^E', b'^E', '\x05'),
     Key('^J', b'^J', '\n'),
     Key('^O', b'^O', '\x0f'),
+    Key('^P', b'^P', '\x10'),
     Key('^R', b'^R', '\x12'),
     Key('^S', b'^S', '\x13'),
     Key('^U', b'^U', '\x15'),
@@ -266,7 +291,7 @@ KEYS = [
     Key('^\\', b'^\\', '\x1c'),
     Key('!resize', b'KEY_RESIZE', curses.KEY_RESIZE),
 ]
-KEYS_TMUX = {k.tmux: k.value for k in KEYS}
+KEYS_TMUX = {k.tmux: k.wch for k in KEYS}
 KEYS_CURSES = {k.value: k.curses for k in KEYS}
 
 
@@ -297,7 +322,7 @@ class DeferredRunner:
         print(f'KEY: {keypress_event.wch!r}')
         return keypress_event.wch
 
-    def await_text(self, text):
+    def await_text(self, text, timeout=1):
         self._ops.append(AwaitText(text))
 
     def await_text_missing(self, text):
@@ -364,8 +389,9 @@ class DeferredRunner:
     def _curses__noop(self, *_, **__):
         pass
 
-    _curses_cbreak = _curses_noecho = _curses_nonl = _curses__noop
+    _curses_cbreak = _curses_endwin = _curses_noecho = _curses__noop
-    _curses_raw = _curses_use_default_colors = _curses__noop
+    _curses_nonl = _curses_raw = _curses_use_default_colors = _curses__noop
+    _curses_set_escdelay = _curses__noop
 
     _curses_error = curses.error  # so we don't mock the exception
 
@@ -391,11 +417,10 @@ class DeferredRunner:
 
     def _curses_initscr(self):
         self._curses_update_lines_cols()
-        self.screen.disabled = False
+        return CursesScreen(self.screen, self)
-        return CursesScreen(self)
 
-    def _curses_endwin(self):
+    def _curses_newwin(self, height, width):
-        self.screen.disabled = True
+        return CursesScreen(Screen(width, height), self)
 
     def _curses_not_implemented(self, fn):
         def fn_inner(*args, **kwargs):

tests/features/cut_uncut_test.py

@@ -132,3 +132,21 @@ def test_selection_cut_uncut_selection_offscreen_x(run):
         h.await_text_missing('hello')
         h.press('^K')
         h.await_text('hello\n')
+
+
+def test_selection_cut_uncut_at_end_of_file(run, ten_lines):
+    with run(str(ten_lines)) as h, and_exit(h):
+        h.press('S-Down')
+        h.press('S-Right')
+        h.press('^K')
+        h.await_text_missing('line_0')
+        h.await_text_missing('line_1')
+        h.await_text('ine_1')
+
+        h.press('^End')
+        h.press('^U')
+        h.await_text('line_0\nl\n')
+        h.await_cursor_position(x=1, y=11)
+
+        h.press('Down')
+        h.await_cursor_position(x=0, y=12)

tests/features/initial_position_test.py

@@ -0,0 +1,28 @@
+from testing.runner import and_exit
+
+
+def test_open_file_named_plus_something(run):
+    with run('+3') as h, and_exit(h):
+        h.await_text(' +3')
+
+
+def test_initial_position_one_file(run, tmpdir):
+    f = tmpdir.join('f')
+    f.write('hello\nworld\n')
+
+    with run('+2', str(f)) as h, and_exit(h):
+        h.await_cursor_position(x=0, y=2)
+
+
+def test_initial_position_multiple_files(run, tmpdir):
+    f = tmpdir.join('f')
+    f.write('1\n2\n3\n4\n')
+    g = tmpdir.join('g')
+    g.write('5\n6\n7\n8\n')
+
+    with run('+2', str(f), '+3', str(g)) as h, and_exit(h):
+        h.await_cursor_position(x=0, y=2)
+
+        h.press('^X')
+
+        h.await_cursor_position(x=0, y=3)

tests/features/key_debug_test.py

@@ -0,0 +1,22 @@
+import curses
+
+from babi.screen import VERSION_STR
+
+
+def test_key_debug(run):
+    with run('--key-debug') as h:
+        h.await_text(VERSION_STR, timeout=2)
+
+        h.await_text('press q to quit')
+
+        h.press('a')
+        h.await_text("'a' 'STRING'")
+
+        h.press('^X')
+        h.await_text(r"'\x18' '^X'")
+
+        with h.resize(width=20, height=20):
+            h.await_text(f"{curses.KEY_RESIZE} 'KEY_RESIZE'")
+
+        h.press('q')
+        h.await_exit()

tests/features/movement_test.py

@@ -411,3 +411,30 @@ def test_sequence_handling(run_only_fake):
         h.press(' test7')
         h.await_text('test1 test2 test3 test4 test5 test6 test7')
         h.await_text(r'\x1b[1;')
+
+
+def test_indentation_using_tabs(run, tmpdir):
+    f = tmpdir.join('f')
+    f.write(f'123456789\n\t12\t{"x" * 20}\n')
+
+    with run(str(f), width=20) as h, and_exit(h):
+        h.await_text('123456789\n    12  xxxxxxxxxxx»\n')
+
+        h.press('Down')
+        h.await_cursor_position(x=0, y=2)
+        h.press('Up')
+        h.await_cursor_position(x=0, y=1)
+
+        h.press('Right')
+        h.await_cursor_position(x=1, y=1)
+        h.press('Down')
+        h.await_cursor_position(x=0, y=2)
+        h.press('Up')
+        h.await_cursor_position(x=1, y=1)
+
+        h.press('Down')
+        h.await_cursor_position(x=0, y=2)
+        h.press('Right')
+        h.await_cursor_position(x=4, y=2)
+        h.press('Up')
+        h.await_cursor_position(x=4, y=1)

tests/features/multiple_files_test.py

@@ -1,10 +1,19 @@
-def test_multiple_files(run, tmpdir):
+import pytest
+
+
+@pytest.fixture
+def abc(tmpdir):
     a = tmpdir.join('file_a')
     a.write('a text')
     b = tmpdir.join('file_b')
     b.write('b text')
     c = tmpdir.join('file_c')
     c.write('c text')
+    yield a, b, c
+
+
+def test_multiple_files(run, abc):
+    a, b, c = abc
 
     with run(str(a), str(b), str(c)) as h:
         h.await_text('file_a')
@@ -44,9 +53,36 @@ def test_multiple_files(run, tmpdir):
         h.press('^J')
         h.await_text('unknown key')
         h.press('^X')
-        h.await_text('file_a')
+        h.await_text('file_b')
         h.await_text_missing('unknown key')
+        h.press('^X')
+        h.await_text('file_a')
+        h.press('^X')
+        h.await_exit()
+
+
+def test_multiple_files_close_from_beginning(run, abc):
+    a, b, c = abc
+
+    with run(str(a), str(b), str(c)) as h:
         h.press('^X')
         h.await_text('file_b')
         h.press('^X')
+        h.await_text('file_c')
+        h.press('^X')
+        h.await_exit()
+
+
+def test_multiple_files_close_from_end(run, abc):
+    a, b, c = abc
+
+    with run(str(a), str(b), str(c)) as h:
+        h.press('M-Right')
+        h.await_text('file_b')
+
+        h.press('^X')
+        h.await_text('file_c')
+        h.press('^X')
+        h.await_text('file_a')
+        h.press('^X')
         h.await_exit()

tests/features/open_test.py

@@ -0,0 +1,38 @@
+from testing.runner import and_exit
+
+
+def test_open_cancelled(run, tmpdir):
+    f = tmpdir.join('f')
+    f.write('hello world')
+
+    with run(str(f)) as h, and_exit(h):
+        h.await_text('hello world')
+
+        h.press('^P')
+        h.await_text('enter filename:')
+        h.press('^C')
+
+        h.await_text('cancelled')
+        h.await_text('hello world')
+
+
+def test_open(run, tmpdir):
+    f = tmpdir.join('f')
+    f.write('hello world')
+    g = tmpdir.join('g')
+    g.write('goodbye world')
+
+    with run(str(f)) as h:
+        h.await_text('hello world')
+
+        h.press('^P')
+        h.press_and_enter(str(g))
+
+        h.await_text('[2/2]')
+        h.await_text('goodbye world')
+
+        h.press('^X')
+        h.await_text('hello world')
+
+        h.press('^X')
+        h.await_exit()

tests/features/replace_test.py

@@ -37,7 +37,7 @@ def test_replace_actual_contents(run, ten_lines):
         h.press_and_enter('line_0')
         h.await_text('replace with:')
         h.press_and_enter('ohai')
-        h.await_text('replace [y(es), n(o), a(ll)]?')
+        h.await_text('replace [yes, no, all]?')
         h.press('y')
         h.await_text_missing('line_0')
         h.await_text('ohai')
@@ -59,7 +59,7 @@ match me!
         h.press_and_enter('me!')
         h.await_text('replace with:')
         h.press_and_enter('youuuu')
-        h.await_text('replace [y(es), n(o), a(ll)]?')
+        h.await_text('replace [yes, no, all]?')
         h.press('y')
         h.await_cursor_position(x=6, y=3)
         h.press('Up')
@@ -74,7 +74,7 @@ def test_replace_cancel_at_individual_replace(run, ten_lines):
         h.press_and_enter(r'line_\d')
         h.await_text('replace with:')
         h.press_and_enter('ohai')
-        h.await_text('replace [y(es), n(o), a(ll)]?')
+        h.await_text('replace [yes, no, all]?')
         h.press('^C')
         h.await_text('cancelled')
 
@@ -86,7 +86,7 @@ def test_replace_unknown_characters_at_individual_replace(run, ten_lines):
         h.press_and_enter(r'line_\d')
         h.await_text('replace with:')
         h.press_and_enter('ohai')
-        h.await_text('replace [y(es), n(o), a(ll)]?')
+        h.await_text('replace [yes, no, all]?')
         h.press('?')
         h.press('^C')
         h.await_text('cancelled')
@@ -99,7 +99,7 @@ def test_replace_say_no_to_individual_replace(run, ten_lines):
         h.press_and_enter('line_[135]')
         h.await_text('replace with:')
         h.press_and_enter('ohai')
-        h.await_text('replace [y(es), n(o), a(ll)]?')
+        h.await_text('replace [yes, no, all]?')
         h.press('y')
         h.await_text_missing('line_1')
         h.press('n')
@@ -116,7 +116,7 @@ def test_replace_all(run, ten_lines):
         h.press_and_enter(r'line_(\d)')
         h.await_text('replace with:')
         h.press_and_enter(r'ohai+\1')
-        h.await_text('replace [y(es), n(o), a(ll)]?')
+        h.await_text('replace [yes, no, all]?')
         h.press('a')
         h.await_text_missing('line')
         h.await_text('ohai+1')
@@ -130,7 +130,7 @@ def test_replace_with_empty_string(run, ten_lines):
         h.press_and_enter('line_1')
         h.await_text('replace with:')
         h.press('Enter')
-        h.await_text('replace [y(es), n(o), a(ll)]?')
+        h.await_text('replace [yes, no, all]?')
         h.press('y')
         h.await_text_missing('line_1')
 
@@ -153,7 +153,7 @@ def test_replace_small_window_size(run, ten_lines):
         h.press_and_enter('line')
         h.await_text('replace with:')
         h.press_and_enter('wat')
-        h.await_text('replace [y(es), n(o), a(ll)]?')
+        h.await_text('replace [yes, no, all]?')
 
         with h.resize(width=8, height=24):
             h.await_text('replace…')
@@ -170,7 +170,7 @@ def test_replace_height_1_highlight(run, tmpdir):
         h.press_and_enter('^x+$')
         h.await_text('replace with:')
         h.press('Enter')
-        h.await_text('replace [y(es), n(o), a(ll)]?')
+        h.await_text('replace [yes, no, all]?')
 
         with h.resize(width=80, height=1):
             h.await_text_missing('xxxxx')
@@ -189,7 +189,7 @@ def test_replace_line_goes_off_screen(run):
         h.press_and_enter('b+')
         h.await_text('replace with:')
         h.press_and_enter('wat')
-        h.await_text('replace [y(es), n(o), a(ll)]?')
+        h.await_text('replace [yes, no, all]?')
         h.await_text(f'{"a" * 20}{"b" * 59}»')
         h.press('y')
         h.await_text(f'{"a" * 20}wat')
@@ -221,7 +221,7 @@ def test_replace_multiple_occurrences_in_line(run):
         h.press_and_enter('a+')
         h.await_text('replace with:')
         h.press_and_enter('q')
-        h.await_text('replace [y(es), n(o), a(ll)]?')
+        h.await_text('replace [yes, no, all]?')
         h.press('a')
         h.await_text('bqbq')
 
@@ -234,7 +234,7 @@ def test_replace_after_wrapping(run, ten_lines):
         h.press_and_enter('line_[02]')
         h.await_text('replace with:')
         h.press_and_enter('ohai')
-        h.await_text('replace [y(es), n(o), a(ll)]?')
+        h.await_text('replace [yes, no, all]?')
         h.press('y')
         h.await_text_missing('line_2')
         h.press('y')
@@ -251,7 +251,7 @@ def test_replace_after_cursor_after_wrapping(run):
         h.press_and_enter('b')
         h.await_text('replace with:')
         h.press_and_enter('q')
-        h.await_text('replace [y(es), n(o), a(ll)]?')
+        h.await_text('replace [yes, no, all]?')
         h.press('n')
         h.press('y')
         h.await_text('replaced 1 occurrence')
@@ -267,8 +267,36 @@ def test_replace_separate_line_after_wrapping(run, ten_lines):
         h.press_and_enter('line_[01]')
         h.await_text('replace with:')
         h.press_and_enter('_')
-        h.await_text('replace [y(es), n(o), a(ll)]?')
+        h.await_text('replace [yes, no, all]?')
         h.press('y')
         h.await_text_missing('line_0')
         h.press('y')
         h.await_text_missing('line_1')
+
+
+def test_replace_with_newline_characters(run, ten_lines):
+    with run(str(ten_lines)) as h, and_exit(h):
+        h.press('^\\')
+        h.await_text('search (to replace):')
+        h.press_and_enter('(line)_([01])')
+        h.await_text('replace with:')
+        h.press_and_enter(r'\1\n\2')
+        h.await_text('replace [yes, no, all]?')
+        h.press('a')
+        h.await_text_missing('line_0')
+        h.await_text_missing('line_1')
+        h.await_text('line\n0\nline\n1\n')
+
+
+def test_replace_with_multiple_newline_characters(run, ten_lines):
+    with run(str(ten_lines)) as h, and_exit(h):
+        h.press('^\\')
+        h.await_text('search (to replace):')
+        h.press_and_enter('(li)(ne)_(1)')
+        h.await_text('replace with:')
+        h.press_and_enter(r'\1\n\2\n\3\n')
+        h.await_text('replace [yes, no, all]?')
+        h.press('a')
+
+        h.await_text_missing('line_1')
+        h.await_text('li\nne\n1\n\nline_2')

tests/features/save_test.py

@@ -1,6 +1,7 @@
 import pytest
 
 from testing.runner import and_exit
+from testing.runner import trigger_command_mode
 
 
 def test_mixed_newlines(run, tmpdir):
@@ -12,6 +13,31 @@ def test_mixed_newlines(run, tmpdir):
         h.await_text(r"mixed newlines will be converted to '\n'")
 
 
+def test_modify_file_with_windows_newlines(run, tmpdir):
+    f = tmpdir.join('f')
+    f.write_binary(b'foo\r\nbar\r\n')
+    with run(str(f)) as h, and_exit(h):
+        # should not start modified
+        h.await_text_missing('*')
+        h.press('Enter')
+        h.await_text('*')
+        h.press('^S')
+        h.await_text('saved!')
+    assert f.read_binary() == b'\r\nfoo\r\nbar\r\n'
+
+
+def test_saving_file_with_multiple_lines_at_end_maintains_those(run, tmpdir):
+    f = tmpdir.join('f')
+    f.write('foo\n\n')
+    with run(str(f)) as h, and_exit(h):
+        h.press('a')
+        h.await_text('*')
+        h.press('^S')
+        h.await_text('saved!')
+
+    assert f.read() == 'afoo\n\n'
+
+
 def test_new_file(run):
     with run('this_is_a_new_file') as h, and_exit(h):
         h.await_text('this_is_a_new_file')
@@ -107,7 +133,7 @@ def test_save_via_ctrl_o(run, tmpdir):
     with run(str(f)) as h, and_exit(h):
         h.press('hello world')
         h.press('^O')
-        h.await_text(f'enter filename: ')
+        h.await_text('enter filename: ')
         h.press('Enter')
         h.await_text('saved! (1 line written)')
     assert f.read() == 'hello world\n'
@@ -148,7 +174,7 @@ def test_save_on_exit_cancel_yn(run):
         h.press('hello')
         h.await_text('hello')
         h.press('^X')
-        h.await_text('file is modified - save [y(es), n(o)]?')
+        h.await_text('file is modified - save [yes, no]?')
         h.press('^C')
         h.await_text('cancelled')
 
@@ -158,7 +184,7 @@ def test_save_on_exit_cancel_filename(run):
         h.press('hello')
         h.await_text('hello')
         h.press('^X')
-        h.await_text('file is modified - save [y(es), n(o)]?')
+        h.await_text('file is modified - save [yes, no]?')
         h.press('y')
         h.await_text('enter filename:')
         h.press('^C')
@@ -171,7 +197,7 @@ def test_save_on_exit(run, tmpdir):
         h.press('hello')
         h.await_text('hello')
         h.press('^X')
-        h.await_text('file is modified - save [y(es), n(o)]?')
+        h.await_text('file is modified - save [yes, no]?')
         h.press('y')
         h.await_text(f'enter filename: {f}')
         h.press('Enter')
@@ -183,9 +209,44 @@ def test_save_on_exit_resize(run, tmpdir):
         h.press('hello')
         h.await_text('hello')
         h.press('^X')
-        h.await_text('file is modified - save [y(es), n(o)]?')
+        h.await_text('file is modified - save [yes, no]?')
         with h.resize(width=10, height=24):
             h.await_text('file is m…')
-        h.await_text('file is modified - save [y(es), n(o)]?')
+        h.await_text('file is modified - save [yes, no]?')
         h.press('^C')
         h.await_text('cancelled')
+
+
+def test_vim_save_on_exit_cancel_yn(run):
+    with run() as h, and_exit(h):
+        h.press('hello')
+        h.await_text('hello')
+        trigger_command_mode(h)
+        h.press_and_enter(':q')
+        h.await_text('file is modified - save [yes, no]?')
+        h.press('^C')
+        h.await_text('cancelled')
+
+
+def test_vim_save_on_exit(run, tmpdir):
+    f = tmpdir.join('f')
+    with run(str(f)) as h:
+        h.press('hello')
+        h.await_text('hello')
+        trigger_command_mode(h)
+        h.press_and_enter(':q')
+        h.await_text('file is modified - save [yes, no]?')
+        h.press('y')
+        h.await_text('enter filename: ')
+        h.press('Enter')
+        h.await_exit()
+
+
+def test_vim_force_exit(run, tmpdir):
+    f = tmpdir.join('f')
+    with run(str(f)) as h:
+        h.press('hello')
+        h.await_text('hello')
+        trigger_command_mode(h)
+        h.press_and_enter(':q!')
+        h.await_exit()

tests/features/sort_test.py

@@ -21,6 +21,16 @@ def test_sort_entire_file(run, unsorted):
     assert unsorted.read() == 'a\nb\nc\nd\n'
 
 
+def test_reverse_sort_entire_file(run, unsorted):
+    with run(str(unsorted)) as h, and_exit(h):
+        trigger_command_mode(h)
+        h.press_and_enter(':sort!')
+        h.await_text('sorted!')
+        h.await_cursor_position(x=0, y=1)
+        h.press('^S')
+    assert unsorted.read() == 'd\nc\nb\na\n'
+
+
 def test_sort_selection(run, unsorted):
     with run(str(unsorted)) as h, and_exit(h):
         h.press('S-Down')
@@ -32,6 +42,18 @@ def test_sort_selection(run, unsorted):
     assert unsorted.read() == 'b\nd\nc\na\n'
 
 
+def test_reverse_sort_selection(run, unsorted):
+    with run(str(unsorted)) as h, and_exit(h):
+        h.press('Down')
+        h.press('S-Down')
+        trigger_command_mode(h)
+        h.press_and_enter(':sort!')
+        h.await_text('sorted!')
+        h.await_cursor_position(x=0, y=2)
+        h.press('^S')
+    assert unsorted.read() == 'd\nc\nb\na\n'
+
+
 def test_sort_selection_does_not_include_eof(run, unsorted):
     with run(str(unsorted)) as h, and_exit(h):
         for _ in range(5):

tests/features/stdin_test.py

@@ -0,0 +1,22 @@
+import shlex
+import sys
+
+from babi.screen import VERSION_STR
+from testing.runner import PrintsErrorRunner
+
+
+def test_open_from_stdin():
+    with PrintsErrorRunner('env', 'TERM=screen', 'bash', '--norc') as h:
+        cmd = (sys.executable, '-mcoverage', 'run', '-m', 'babi', '-')
+        babi_cmd = ' '.join(shlex.quote(part) for part in cmd)
+        h.press_and_enter(fr"echo $'hello\nworld' | {babi_cmd}")
+
+        h.await_text(VERSION_STR, timeout=2)
+        h.await_text('<<new file>> *')
+        h.await_text('hello\nworld')
+
+        h.press('^X')
+        h.press('n')
+        h.await_text_missing('<<new file>>')
+        h.press_and_enter('exit')
+        h.await_exit()

tests/features/syntax_highlight_test.py

@@ -15,6 +15,7 @@ THEME = json.dumps({
             'settings': {'foreground': '#5f0000', 'background': '#ff5f5f'},
         },
         {'scope': 'tqs', 'settings': {'foreground': '#00005f'}},
+        {'scope': 'qmark', 'settings': {'foreground': '#5f0000'}},
         {'scope': 'b', 'settings': {'fontStyle': 'bold'}},
         {'scope': 'i', 'settings': {'fontStyle': 'italic'}},
         {'scope': 'u', 'settings': {'fontStyle': 'underline'}},
@@ -28,6 +29,7 @@ SYNTAX = json.dumps({
         {'match': r'#.*$\n?', 'name': 'comment'},
         {'match': r'^-.*$\n?', 'name': 'diffremove'},
         {'begin': '"""', 'end': '"""', 'name': 'tqs'},
+        {'match': r'\?', 'name': 'qmark'},
     ],
 })
 DEMO_S = '''\
@@ -43,7 +45,7 @@ still more
 @pytest.fixture(autouse=True)
 def theme_and_grammar(xdg_data_home, xdg_config_home):
     xdg_config_home.join('babi/theme.json').ensure().write(THEME)
-    xdg_data_home.join('babi/textmate_syntax/demo.json').ensure().write(SYNTAX)
+    xdg_data_home.join('babi/grammar_v1/demo.json').ensure().write(SYNTAX)
 
 
 @pytest.fixture
@@ -78,3 +80,76 @@ def test_syntax_highlighting_does_not_highlight_arrows(run, tmpdir):
     with run(str(f), term='screen-256color', width=20) as h, and_exit(h):
         h.await_text('loooo')
         h.assert_screen_attr_equals(2, [(243, 40, 0)] * 19 + [(236, 40, 0)])
+
+        h.press('Down')
+        h.press('^E')
+        h.await_text_missing('loooo')
+        expected = [(236, 40, 0)] + [(243, 40, 0)] * 15 + [(236, 40, 0)] * 4
+        h.assert_screen_attr_equals(2, expected)
+
+
+def test_syntax_highlighting_off_screen_does_not_crash(run, tmpdir):
+    f = tmpdir.join('f.demo')
+    f.write(f'"""a"""{"x" * 40}"""b"""')
+
+    with run(str(f), term='screen-256color', width=20) as h, and_exit(h):
+        h.await_text('"""a"""')
+        h.assert_screen_attr_equals(1, [(17, 40, 0)] * 7 + [(236, 40, 0)] * 13)
+        h.press('^E')
+        h.await_text('"""b"""')
+        expected = [(236, 40, 0)] * 11 + [(17, 40, 0)] * 7 + [(236, 40, 0)] * 2
+        h.assert_screen_attr_equals(1, expected)
+
+
+def test_syntax_highlighting_one_off_left_of_screen(run, tmpdir):
+    f = tmpdir.join('f.demo')
+    f.write(f'{"x" * 11}?123456789')
+
+    with run(str(f), term='screen-256color', width=20) as h, and_exit(h):
+        h.await_text('xxx?123')
+        expected = [(236, 40, 0)] * 11 + [(52, 40, 0)] + [(236, 40, 0)] * 8
+        h.assert_screen_attr_equals(1, expected)
+
+        h.press('End')
+        h.await_text_missing('?')
+        h.assert_screen_attr_equals(1, [(236, 40, 0)] * 20)
+
+
+def test_syntax_highlighting_to_edge_of_screen(run, tmpdir):
+    f = tmpdir.join('f.demo')
+    f.write(f'# {"x" * 18}')
+
+    with run(str(f), term='screen-256color', width=20) as h, and_exit(h):
+        h.await_text('# xxx')
+        h.assert_screen_attr_equals(1, [(243, 40, 0)] * 20)
+
+
+def test_syntax_highlighting_with_tabs(run, tmpdir):
+    f = tmpdir.join('f.demo')
+    f.write('\t# 12345678901234567890\n')
+
+    with run(str(f), term='screen-256color', width=20) as h, and_exit(h):
+        h.await_text('1234567890')
+        expected = 4 * [(236, 40, 0)] + 15 * [(243, 40, 0)] + [(236, 40, 0)]
+        h.assert_screen_attr_equals(1, expected)
+
+
+def test_syntax_highlighting_tabs_after_line_creation(run, tmpdir):
+    f = tmpdir.join('f')
+    # trailing whitespace is used to trigger highlighting
+    f.write('foo\n\txx \ny    \n')
+
+    with run(str(f), term='screen-256color') as h, and_exit(h):
+        # this looks weird, but it populates the width cache
+        h.press('Down')
+        h.press('Down')
+        h.press('Down')
+
+        # press enter after the tab
+        h.press('Up')
+        h.press('Up')
+        h.press('Right')
+        h.press('Right')
+        h.press('Enter')
+
+        h.await_text('foo\n    x\nx\ny\n')

tests/features/tabsize_test.py

@@ -0,0 +1,30 @@
+import pytest
+
+from testing.runner import and_exit
+from testing.runner import trigger_command_mode
+
+
+@pytest.mark.parametrize('setting', ('tabsize', 'tabstop'))
+def test_set_tabstop(run, setting):
+    with run() as h, and_exit(h):
+        h.press('a')
+        h.press('Left')
+        trigger_command_mode(h)
+        h.press_and_enter(f':{setting} 2')
+        h.await_text('updated!')
+        h.press('Tab')
+        h.await_text('\n  a')
+        h.await_cursor_position(x=2, y=1)
+
+
+@pytest.mark.parametrize('tabsize', ('-1', '0', 'wat'))
+def test_set_invalid_tabstop(run, tabsize):
+    with run() as h, and_exit(h):
+        h.press('a')
+        h.press('Left')
+        trigger_command_mode(h)
+        h.press_and_enter(f':tabstop {tabsize}')
+        h.await_text(f'invalid size: {tabsize}')
+        h.press('Tab')
+        h.await_text('    a')
+        h.await_cursor_position(x=4, y=1)

tests/features/text_editing_test.py

@@ -97,6 +97,24 @@ def test_delete_at_end_of_line(run, tmpdir):
         h.await_text('f *')
 
 
+def test_delete_at_end_of_last_line(run, tmpdir):
+    f = tmpdir.join('f')
+    f.write('hello\n')
+
+    with run(str(f)) as h, and_exit(h):
+        h.await_text('hello')
+        h.press('End')
+        h.press('DC')
+        # should not make the file modified
+        h.await_text_missing('*')
+
+        # delete should still be functional
+        h.press('Left')
+        h.press('Left')
+        h.press('DC')
+        h.await_text('helo')
+
+
 def test_press_enter_beginning_of_file(run, tmpdir):
     f = tmpdir.join('f')
     f.write('hello world')

tests/features/undo_redo_test.py

@@ -1,3 +1,5 @@
+import pytest
+
 from testing.runner import and_exit
 
 
@@ -9,7 +11,8 @@ def test_nothing_to_undo_redo(run):
         h.await_text('nothing to redo!')
 
 
-def test_undo_redo(run):
+@pytest.mark.parametrize('r', ('M-U', 'M-e'))
+def test_undo_redo(run, r):
     with run() as h, and_exit(h):
         h.press('hello')
         h.await_text('hello')
@@ -17,7 +20,7 @@ def test_undo_redo(run):
         h.await_text('undo: text')
         h.await_text_missing('hello')
         h.await_text_missing(' *')
-        h.press('M-U')
+        h.press(r)
         h.await_text('redo: text')
         h.await_text('hello')
         h.await_text(' *')
@@ -127,3 +130,16 @@ def test_undo_redo_causes_scroll(run):
         h.await_cursor_position(x=0, y=1)
         h.press('M-U')
         h.await_cursor_position(x=0, y=4)
+
+
+def test_undo_redo_clears_selection(run, ten_lines):
+    # maintaining the selection across undo/redo is both difficult and not all
+    # that useful.  prior to this it was buggy anyway (a negative selection
+    # indented and then undone would highlight out of bounds)
+    with run(str(ten_lines), width=20) as h, and_exit(h):
+        h.press('S-Down')
+        h.press('Tab')
+        h.await_cursor_position(x=4, y=2)
+        h.press('M-u')
+        h.await_cursor_position(x=0, y=2)
+        h.assert_screen_attr_equals(1, [(-1, -1, 0)] * 20)

tests/file_test.py

@@ -2,12 +2,13 @@ import io
 
 import pytest
 
+from babi.color_manager import ColorManager
 from babi.file import File
 from babi.file import get_lines
 
 
 def test_position_repr():
-    ret = repr(File('f.txt', ()))
+    ret = repr(File('f.txt', 0, ColorManager.make(), ()))
     assert ret == "<File 'f.txt'>"
 
 

tests/highlight_test.py

@@ -1,22 +1,37 @@
-from babi.highlight import Grammar
+import pytest
-from babi.highlight import Grammars
+
 from babi.highlight import highlight_line
 from babi.highlight import Region
 
 
-def _compiler_state(grammar_dct, *others):
+def test_grammar_matches_extension_only_name(make_grammars):
-    grammar = Grammar.from_data(grammar_dct)
+    data = {'scopeName': 'shell', 'patterns': [], 'fileTypes': ['bashrc']}
-    grammars = [grammar, *(Grammar.from_data(dct) for dct in others)]
+    grammars = make_grammars(data)
-    compiler = Grammars(grammars).compiler_for_scope(grammar.scope_name)
+    compiler = grammars.compiler_for_file('.bashrc', 'alias nano=babi')
-    return compiler, compiler.root_state
+    assert compiler.root_state.entries[0].scope[0] == 'shell'
 
 
-def test_backslash_a():
+def test_grammar_matches_via_identify_tag(make_grammars):
+    grammars = make_grammars({'scopeName': 'source.ini', 'patterns': []})
+    compiler = grammars.compiler_for_file('setup.cfg', '')
+    assert compiler.root_state.entries[0].scope[0] == 'source.ini'
+
+
+@pytest.fixture
+def compiler_state(make_grammars):
+    def _compiler_state(*grammar_dcts):
+        grammars = make_grammars(*grammar_dcts)
+        compiler = grammars.compiler_for_scope(grammar_dcts[0]['scopeName'])
+        return compiler, compiler.root_state
+    return _compiler_state
+
+
+def test_backslash_a(compiler_state):
     grammar = {
         'scopeName': 'test',
         'patterns': [{'name': 'aaa', 'match': r'\Aa+'}],
     }
-    compiler, state = _compiler_state(grammar)
+    compiler, state = compiler_state(grammar)
 
     state, (region_0,) = highlight_line(compiler, state, 'aaa', True)
     state, (region_1,) = highlight_line(compiler, state, 'aaa', False)
@@ -39,8 +54,8 @@ BEGIN_END_NO_NL = {
 }
 
 
-def test_backslash_g_inline():
+def test_backslash_g_inline(compiler_state):
-    compiler, state = _compiler_state(BEGIN_END_NO_NL)
+    compiler, state = compiler_state(BEGIN_END_NO_NL)
 
     _, regions = highlight_line(compiler, state, 'xaax', True)
     assert regions == (
@@ -51,8 +66,8 @@ def test_backslash_g_inline():
     )
 
 
-def test_backslash_g_next_line():
+def test_backslash_g_next_line(compiler_state):
-    compiler, state = _compiler_state(BEGIN_END_NO_NL)
+    compiler, state = compiler_state(BEGIN_END_NO_NL)
 
     state, regions1 = highlight_line(compiler, state, 'x\n', True)
     state, regions2 = highlight_line(compiler, state, 'aax\n', False)
@@ -69,8 +84,8 @@ def test_backslash_g_next_line():
     )
 
 
-def test_end_before_other_match():
+def test_end_before_other_match(compiler_state):
-    compiler, state = _compiler_state(BEGIN_END_NO_NL)
+    compiler, state = compiler_state(BEGIN_END_NO_NL)
 
     state, regions = highlight_line(compiler, state, 'xazzx', True)
 
@@ -95,8 +110,8 @@ BEGIN_END_NL = {
 }
 
 
-def test_backslash_g_captures_nl():
+def test_backslash_g_captures_nl(compiler_state):
-    compiler, state = _compiler_state(BEGIN_END_NL)
+    compiler, state = compiler_state(BEGIN_END_NL)
 
     state, regions1 = highlight_line(compiler, state, 'x\n', True)
     state, regions2 = highlight_line(compiler, state, 'aax\n', False)
@@ -112,8 +127,8 @@ def test_backslash_g_captures_nl():
     )
 
 
-def test_backslash_g_captures_nl_next_line():
+def test_backslash_g_captures_nl_next_line(compiler_state):
-    compiler, state = _compiler_state(BEGIN_END_NL)
+    compiler, state = compiler_state(BEGIN_END_NL)
 
     state, regions1 = highlight_line(compiler, state, 'x\n', True)
     state, regions2 = highlight_line(compiler, state, 'aa\n', False)
@@ -135,8 +150,8 @@ def test_backslash_g_captures_nl_next_line():
     )
 
 
-def test_while_no_nl():
+def test_while_no_nl(compiler_state):
-    compiler, state = _compiler_state({
+    compiler, state = compiler_state({
         'scopeName': 'test',
         'patterns': [{
             'begin': '> ',
@@ -170,8 +185,8 @@ def test_while_no_nl():
     )
 
 
-def test_complex_captures():
+def test_complex_captures(compiler_state):
-    compiler, state = _compiler_state({
+    compiler, state = compiler_state({
         'scopeName': 'test',
         'patterns': [
             {
@@ -201,8 +216,8 @@ def test_complex_captures():
     )
 
 
-def test_captures_multiple_applied_to_same_capture():
+def test_captures_multiple_applied_to_same_capture(compiler_state):
-    compiler, state = _compiler_state({
+    compiler, state = compiler_state({
         'scopeName': 'test',
         'patterns': [
             {
@@ -244,8 +259,8 @@ def test_captures_multiple_applied_to_same_capture():
     )
 
 
-def test_captures_ignores_empty():
+def test_captures_ignores_empty(compiler_state):
-    compiler, state = _compiler_state({
+    compiler, state = compiler_state({
         'scopeName': 'test',
         'patterns': [{
             'match': '(.*) hi',
@@ -267,8 +282,8 @@ def test_captures_ignores_empty():
     )
 
 
-def test_captures_ignores_invalid_out_of_bounds():
+def test_captures_ignores_invalid_out_of_bounds(compiler_state):
-    compiler, state = _compiler_state({
+    compiler, state = compiler_state({
         'scopeName': 'test',
         'patterns': [{'match': '.', 'captures': {'1': {'name': 'oob'}}}],
     })
@@ -280,8 +295,8 @@ def test_captures_ignores_invalid_out_of_bounds():
     )
 
 
-def test_captures_begin_end():
+def test_captures_begin_end(compiler_state):
-    compiler, state = _compiler_state({
+    compiler, state = compiler_state({
         'scopeName': 'test',
         'patterns': [
             {
@@ -302,8 +317,8 @@ def test_captures_begin_end():
     )
 
 
-def test_captures_while_captures():
+def test_captures_while_captures(compiler_state):
-    compiler, state = _compiler_state({
+    compiler, state = compiler_state({
         'scopeName': 'test',
         'patterns': [
             {
@@ -331,8 +346,8 @@ def test_captures_while_captures():
     )
 
 
-def test_captures_implies_begin_end_captures():
+def test_captures_implies_begin_end_captures(compiler_state):
-    compiler, state = _compiler_state({
+    compiler, state = compiler_state({
         'scopeName': 'test',
         'patterns': [
             {
@@ -352,8 +367,8 @@ def test_captures_implies_begin_end_captures():
     )
 
 
-def test_captures_implies_begin_while_captures():
+def test_captures_implies_begin_while_captures(compiler_state):
-    compiler, state = _compiler_state({
+    compiler, state = compiler_state({
         'scopeName': 'test',
         'patterns': [
             {
@@ -380,8 +395,8 @@ def test_captures_implies_begin_while_captures():
     )
 
 
-def test_include_self():
+def test_include_self(compiler_state):
-    compiler, state = _compiler_state({
+    compiler, state = compiler_state({
         'scopeName': 'test',
         'patterns': [
             {
@@ -404,8 +419,8 @@ def test_include_self():
     )
 
 
-def test_include_repository_rule():
+def test_include_repository_rule(compiler_state):
-    compiler, state = _compiler_state({
+    compiler, state = compiler_state({
         'scopeName': 'test',
         'patterns': [{'include': '#impl'}],
         'repository': {
@@ -426,8 +441,40 @@ def test_include_repository_rule():
     )
 
 
-def test_include_other_grammar():
+def test_include_with_nested_repositories(compiler_state):
-    compiler, state = _compiler_state(
+    compiler, state = compiler_state({
+        'scopeName': 'test',
+        'patterns': [{
+            'begin': '<', 'end': '>', 'name': 'b',
+            'patterns': [
+                {'include': '#rule1'},
+                {'include': '#rule2'},
+                {'include': '#rule3'},
+            ],
+            'repository': {
+                'rule2': {'match': '2', 'name': 'inner2'},
+                'rule3': {'match': '3', 'name': 'inner3'},
+            },
+        }],
+        'repository': {
+            'rule1': {'match': '1', 'name': 'root1'},
+            'rule2': {'match': '2', 'name': 'root2'},
+        },
+    })
+
+    state, regions = highlight_line(compiler, state, '<123>', first_line=True)
+
+    assert regions == (
+        Region(0, 1, ('test', 'b')),
+        Region(1, 2, ('test', 'b', 'root1')),
+        Region(2, 3, ('test', 'b', 'inner2')),
+        Region(3, 4, ('test', 'b', 'inner3')),
+        Region(4, 5, ('test', 'b')),
+    )
+
+
+def test_include_other_grammar(compiler_state):
+    compiler, state = compiler_state(
         {
             'scopeName': 'test',
             'patterns': [
@@ -482,8 +529,8 @@ def test_include_other_grammar():
     )
 
 
-def test_include_base():
+def test_include_base(compiler_state):
-    compiler, state = _compiler_state(
+    compiler, state = compiler_state(
         {
             'scopeName': 'test',
             'patterns': [
@@ -528,3 +575,87 @@ def test_include_base():
         Region(2, 3, ('test', 'tick')),
         Region(3, 4, ('test',)),
     )
+
+
+def test_rule_with_begin_and_no_end(compiler_state):
+    compiler, state = compiler_state({
+        'scopeName': 'test',
+        'patterns': [
+            {
+                'begin': '!', 'end': '!', 'name': 'bang',
+                'patterns': [{'begin': '--', 'name': 'invalid'}],
+            },
+        ],
+    })
+
+    state, regions = highlight_line(compiler, state, '!x! !--!', True)
+
+    assert regions == (
+        Region(0, 1, ('test', 'bang')),
+        Region(1, 2, ('test', 'bang')),
+        Region(2, 3, ('test', 'bang')),
+        Region(3, 4, ('test',)),
+        Region(4, 5, ('test', 'bang')),
+        Region(5, 7, ('test', 'bang', 'invalid')),
+        Region(7, 8, ('test', 'bang', 'invalid')),
+    )
+
+
+def test_begin_end_substitute_special_chars(compiler_state):
+    compiler, state = compiler_state({
+        'scopeName': 'test',
+        'patterns': [{'begin': r'(\*)', 'end': r'\1', 'name': 'italic'}],
+    })
+
+    state, regions = highlight_line(compiler, state, '*italic*', True)
+
+    assert regions == (
+        Region(0, 1, ('test', 'italic')),
+        Region(1, 7, ('test', 'italic')),
+        Region(7, 8, ('test', 'italic')),
+    )
+
+
+def test_backslash_z(compiler_state):
+    # similar to text.git-commit grammar, \z matches nothing!
+    compiler, state = compiler_state({
+        'scopeName': 'test',
+        'patterns': [
+            {'begin': '#', 'end': r'\z', 'name': 'comment'},
+            {'name': 'other', 'match': '.'},
+        ],
+    })
+
+    state, regions1 = highlight_line(compiler, state, '# comment', True)
+    state, regions2 = highlight_line(compiler, state, 'other?', False)
+
+    assert regions1 == (
+        Region(0, 1, ('test', 'comment')),
+        Region(1, 9, ('test', 'comment')),
+    )
+
+    assert regions2 == (
+        Region(0, 6, ('test', 'comment')),
+    )
+
+
+def test_buggy_begin_end_grammar(compiler_state):
+    # before this would result in an infinite loop of start / end
+    compiler, state = compiler_state({
+        'scopeName': 'test',
+        'patterns': [
+            {
+                'begin': '(?=</style)',
+                'end': '(?=</style)',
+                'name': 'css',
+            },
+        ],
+    })
+
+    state, regions = highlight_line(compiler, state, 'test </style', True)
+
+    assert regions == (
+        Region(0, 5, ('test',)),
+        Region(5, 6, ('test', 'css')),
+        Region(6, 12, ('test',)),
+    )

tests/hl/syntax_test.py

@@ -4,9 +4,9 @@ from unittest import mock
 
 import pytest
 
+from babi.buf import Buf
 from babi.color_manager import ColorManager
-from babi.highlight import Grammar
+from babi.hl.interface import HL
-from babi.highlight import Grammars
 from babi.hl.syntax import Syntax
 from babi.theme import Color
 from babi.theme import Theme
@@ -72,18 +72,23 @@ THEME = Theme.from_dct({
 
 
 @pytest.fixture
-def syntax():
+def syntax(make_grammars):
-    return Syntax(Grammars([Grammar.blank()]), THEME, ColorManager.make())
+    return Syntax(make_grammars(), THEME, ColorManager.make())
 
 
 def test_init_screen_low_color(stdscr, syntax):
     with FakeCurses.patch(n_colors=16, can_change_color=False) as fake_curses:
         syntax._init_screen(stdscr)
-    assert syntax.color_manager.colors == {}
+    assert syntax.color_manager.colors == {
-    assert syntax.color_manager.raw_pairs == {}
+        Color.parse('#cccccc'): -1,
+        Color.parse('#333333'): -1,
+        Color.parse('#000000'): -1,
+        Color.parse('#009900'): -1,
+    }
+    assert syntax.color_manager.raw_pairs == {(-1, -1): 1}
     assert fake_curses.colors == {}
-    assert fake_curses.pairs == {}
+    assert fake_curses.pairs == {1: (-1, -1)}
-    assert stdscr.attr == 0
+    assert stdscr.attr == 1 << 8
 
 
 def test_init_screen_256_color(stdscr, syntax):
@@ -131,7 +136,7 @@ def test_lazily_instantiated_pairs(stdscr, syntax):
         assert len(fake_curses.pairs) == 1
 
         style = THEME.select(('string.python',))
-        attr = syntax.get_blank_file_highlighter().attr(style)
+        attr = syntax.blank_file_highlighter().attr(style)
         assert attr == 2 << 8
 
         assert len(syntax.color_manager.raw_pairs) == 2
@@ -143,5 +148,22 @@ def test_style_attributes_applied(stdscr, syntax):
         syntax._init_screen(stdscr)
 
         style = THEME.select(('keyword.python',))
-        attr = syntax.get_blank_file_highlighter().attr(style)
+        attr = syntax.blank_file_highlighter().attr(style)
         assert attr == 2 << 8 | curses.A_BOLD
+
+
+def test_syntax_highlight_cache_first_line(stdscr, make_grammars):
+    with FakeCurses.patch(n_colors=256, can_change_color=False):
+        grammars = make_grammars({
+            'scopeName': 'source.demo',
+            'fileTypes': ['demo'],
+            'patterns': [{'match': r'\Aint', 'name': 'keyword'}],
+        })
+        syntax = Syntax(grammars, THEME, ColorManager.make())
+        syntax._init_screen(stdscr)
+        file_hl = syntax.file_highlighter('foo.demo', '')
+        file_hl.highlight_until(Buf(['int', 'int']), 2)
+        assert file_hl.regions == [
+            (HL(0, 3, curses.A_BOLD | 2 << 8),),
+            (),
+        ]

tests/list_spy_test.py

@@ -1,144 +0,0 @@
-import pytest
-
-from babi.list_spy import ListSpy
-
-
-def test_list_spy_repr():
-    assert repr(ListSpy(['a', 'b', 'c'])) == "ListSpy(['a', 'b', 'c'])"
-
-
-def test_list_spy_item_retrieval():
-    spy = ListSpy(['a', 'b', 'c'])
-    assert spy[1] == 'b'
-    assert spy[-1] == 'c'
-    with pytest.raises(IndexError):
-        spy[3]
-
-
-def test_list_spy_del():
-    lst = ['a', 'b', 'c']
-
-    spy = ListSpy(lst)
-    del spy[1]
-
-    assert lst == ['a', 'c']
-
-    spy.undo(lst)
-
-    assert lst == ['a', 'b', 'c']
-
-
-def test_list_spy_del_with_negative():
-    lst = ['a', 'b', 'c']
-
-    spy = ListSpy(lst)
-    del spy[-1]
-
-    assert lst == ['a', 'b']
-
-    spy.undo(lst)
-
-    assert lst == ['a', 'b', 'c']
-
-
-def test_list_spy_insert():
-    lst = ['a', 'b', 'c']
-
-    spy = ListSpy(lst)
-    spy.insert(1, 'q')
-
-    assert lst == ['a', 'q', 'b', 'c']
-
-    spy.undo(lst)
-
-    assert lst == ['a', 'b', 'c']
-
-
-def test_list_spy_insert_with_negative():
-    lst = ['a', 'b', 'c']
-
-    spy = ListSpy(lst)
-    spy.insert(-1, 'q')
-
-    assert lst == ['a', 'b', 'q', 'c']
-
-    spy.undo(lst)
-
-    assert lst == ['a', 'b', 'c']
-
-
-def test_list_spy_set_value():
-    lst = ['a', 'b', 'c']
-
-    spy = ListSpy(lst)
-    spy[1] = 'hello'
-
-    assert lst == ['a', 'hello', 'c']
-
-    spy.undo(lst)
-
-    assert lst == ['a', 'b', 'c']
-
-
-def test_list_spy_multiple_modifications():
-    lst = ['a', 'b', 'c']
-
-    spy = ListSpy(lst)
-    spy[1] = 'hello'
-    spy.insert(1, 'ohai')
-    del spy[0]
-
-    assert lst == ['ohai', 'hello', 'c']
-
-    spy.undo(lst)
-
-    assert lst == ['a', 'b', 'c']
-
-
-def test_list_spy_iter():
-    spy = ListSpy(['a', 'b', 'c'])
-    spy_iter = iter(spy)
-    assert next(spy_iter) == 'a'
-    assert next(spy_iter) == 'b'
-    assert next(spy_iter) == 'c'
-    with pytest.raises(StopIteration):
-        next(spy_iter)
-
-
-def test_list_spy_append():
-    lst = ['a', 'b', 'c']
-
-    spy = ListSpy(lst)
-    spy.append('q')
-
-    assert lst == ['a', 'b', 'c', 'q']
-
-    spy.undo(lst)
-
-    assert lst == ['a', 'b', 'c']
-
-
-def test_list_spy_pop_default():
-    lst = ['a', 'b', 'c']
-
-    spy = ListSpy(lst)
-    spy.pop()
-
-    assert lst == ['a', 'b']
-
-    spy.undo(lst)
-
-    assert lst == ['a', 'b', 'c']
-
-
-def test_list_spy_pop_idx():
-    lst = ['a', 'b', 'c']
-
-    spy = ListSpy(lst)
-    spy.pop(1)
-
-    assert lst == ['a', 'c']
-
-    spy.undo(lst)
-
-    assert lst == ['a', 'b', 'c']

tests/main_test.py

@@ -0,0 +1,19 @@
+import pytest
+
+from babi import main
+
+
+@pytest.mark.parametrize(
+    ('in_filenames', 'expected_filenames', 'expected_positions'),
+    (
+        ([], [None], [0]),
+        (['+3'], ['+3'], [0]),
+        (['f'], ['f'], [0]),
+        (['+3', 'f'], ['f'], [3]),
+        (['+-3', 'f'], ['f'], [-3]),
+        (['+3', '+3'], ['+3'], [3]),
+        (['+2', 'f', '+5', 'g'], ['f', 'g'], [2, 5]),
+    ),
+)
+def test_filenames(in_filenames, expected_filenames, expected_positions):
+    filenames, positions = main._filenames(in_filenames)

tests/reg_test.py

@@ -35,9 +35,8 @@ def test_reg_other_escapes_left_untouched():
 def test_reg_not_out_of_bounds_at_end():
     # the only way this is triggerable is with an illegal regex, we'd rather
     # produce an error about the regex being wrong than an IndexError
-    reg = _Reg('\\A\\')
     with pytest.raises(onigurumacffi.OnigError) as excinfo:
-        reg.search('\\', 0, first_line=False, boundary=False)
+        _Reg('\\A\\')
     msg, = excinfo.value.args
     assert msg == 'end pattern at escape'
 

tests/textmate_demo_test.py

@@ -0,0 +1,84 @@
+import json
+
+import pytest
+
+from babi.textmate_demo import main
+
+THEME = {
+    'colors': {'foreground': '#ffffff', 'background': '#000000'},
+    'tokenColors': [
+        {'scope': 'bold', 'settings': {'fontStyle': 'bold'}},
+        {'scope': 'italic', 'settings': {'fontStyle': 'italic'}},
+        {'scope': 'underline', 'settings': {'fontStyle': 'underline'}},
+        {'scope': 'comment', 'settings': {'foreground': '#1e77d3'}},
+    ],
+}
+
+GRAMMAR = {
+    'scopeName': 'source.demo',
+    'fileTypes': ['demo'],
+    'patterns': [
+        {'match': r'\*[^*]*\*', 'name': 'bold'},
+        {'match': '/[^/]*/', 'name': 'italic'},
+        {'match': '_[^_]*_', 'name': 'underline'},
+        {'match': '#.*', 'name': 'comment'},
+    ],
+}
+
+
+@pytest.fixture
+def theme_grammars(tmpdir):
+    theme = tmpdir.join('config/theme.json').ensure()
+    theme.write(json.dumps(THEME))
+    grammars = tmpdir.join('grammar_v1').ensure_dir()
+    grammars.join('source.demo.json').write(json.dumps(GRAMMAR))
+    return theme, grammars
+
+
+def test_basic(theme_grammars, tmpdir, capsys):
+    theme, grammars = theme_grammars
+
+    f = tmpdir.join('f.demo')
+    f.write('*bold*/italic/_underline_# comment\n')
+
+    assert not main((
+        '--theme', str(theme), '--grammar-dir', str(grammars),
+        str(f),
+    ))
+
+    out, _ = capsys.readouterr()
+
+    assert out == (
+        '\x1b[48;2;0;0;0m\n'
+        '\x1b[38;2;255;255;255m\x1b[48;2;0;0;0m\x1b[1m'
+        '*bold*'
+        '\x1b[39m\x1b[49m\x1b[22m'
+        '\x1b[38;2;255;255;255m\x1b[48;2;0;0;0m\x1b[3m'
+        '/italic/'
+        '\x1b[39m\x1b[49m\x1b[23m'
+        '\x1b[38;2;255;255;255m\x1b[48;2;0;0;0m\x1b[4m'
+        '_underline_'
+        '\x1b[39m\x1b[49m\x1b[24m'
+        '\x1b[38;2;30;119;211m\x1b[48;2;0;0;0m'
+        '# comment'
+        '\x1b[39m\x1b[49m\x1b'
+        '[38;2;255;255;255m\x1b[48;2;0;0;0m\n\x1b[39m\x1b[49m'
+        '\x1b[m'
+    )
+
+
+def test_basic_with_blank_theme(theme_grammars, tmpdir, capsys):
+    theme, grammars = theme_grammars
+    theme.write('{}')
+
+    f = tmpdir.join('f.demo')
+    f.write('*bold*/italic/_underline_# comment\n')
+
+    assert not main((
+        '--theme', str(theme), '--grammar-dir', str(grammars),
+        str(f),
+    ))
+
+    out, _ = capsys.readouterr()
+
+    assert out == '*bold*/italic/_underline_# comment\n\x1b[m'

tests/theme_test.py

@@ -72,6 +72,31 @@ def test_theme_scope_split_by_commas():
     assert theme.select(('c',)).i is True
 
 
+def test_theme_scope_comma_at_beginning_and_end():
+    theme = Theme.from_dct({
+        'colors': {'foreground': '#cccccc', 'background': '#333333'},
+        'tokenColors': [
+            {'scope': '\n,a,b,\n', 'settings': {'fontStyle': 'italic'}},
+        ],
+    })
+    assert theme.select(('d',)).i is False
+    assert theme.select(('a',)).i is True
+    assert theme.select(('b',)).i is True
+
+
+def test_theme_scope_internal_newline_commas():
+    # this is arguably malformed, but `cobalt2` in the wild has this issue
+    theme = Theme.from_dct({
+        'colors': {'foreground': '#cccccc', 'background': '#333333'},
+        'tokenColors': [
+            {'scope': '\n,a,\n,b,\n', 'settings': {'fontStyle': 'italic'}},
+        ],
+    })
+    assert theme.select(('d',)).i is False
+    assert theme.select(('a',)).i is True
+    assert theme.select(('b',)).i is True
+
+
 def test_theme_scope_as_A_list():
     theme = Theme.from_dct({
         'colors': {'foreground': '#cccccc', 'background': '#333333'},