v0.0.21

fix crashing on permission denied error

.github/FUNDING.yml

@@ -0,0 +1 @@
+github: asottile

.gitignore

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

.pre-commit-config.yaml

@@ -1,6 +1,6 @@
 repos:
 -   repo: https://github.com/pre-commit/pre-commit-hooks
-    rev: v2.5.0
+    rev: v3.4.0
     hooks:
     -   id: trailing-whitespace
     -   id: end-of-file-fixer
@@ -11,34 +11,34 @@ repos:
     -   id: name-tests-test
     -   id: requirements-txt-fixer
 -   repo: https://gitlab.com/pycqa/flake8
-    rev: 3.7.9
+    rev: 3.8.4
     hooks:
     -   id: flake8
         additional_dependencies: [flake8-typing-imports==1.7.0]
 -   repo: https://github.com/pre-commit/mirrors-autopep8
-    rev: v1.5
+    rev: v1.5.4
     hooks:
     -   id: autopep8
 -   repo: https://github.com/asottile/reorder_python_imports
-    rev: v2.1.0
+    rev: v2.3.6
     hooks:
     -   id: reorder-python-imports
         args: [--py3-plus]
 -   repo: https://github.com/asottile/add-trailing-comma
-    rev: v2.0.1
+    rev: v2.1.0
     hooks:
     -   id: add-trailing-comma
         args: [--py36-plus]
 -   repo: https://github.com/asottile/pyupgrade
-    rev: v2.1.0
+    rev: v2.7.4
     hooks:
     -   id: pyupgrade
         args: [--py36-plus]
 -   repo: https://github.com/asottile/setup-cfg-fmt
-    rev: v1.7.0
+    rev: v1.16.0
     hooks:
     -   id: setup-cfg-fmt
 -   repo: https://github.com/pre-commit/mirrors-mypy
-    rev: v0.770
+    rev: v0.800
     hooks:
     -   id: mypy

README.md

@@ -1,16 +1,30 @@
 [![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)
+[![pre-commit.ci status](https://results.pre-commit.ci/badge/github/asottile/babi/master.svg)](https://results.pre-commit.ci/latest/github/asottile/babi/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`.
 
+### babi vs. nano
+
+here is a youtube video where I discuss the motivation for creating and using
+`babi` instead of `nano`:
+
+[![youtube video about babi](https://img.youtube.com/vi/WyR1hAGmR3g/mqdefault.jpg)](https://youtu.be/WyR1hAGmR3g)
+
 ### quitting babi
 
 currently you can quit `babi` by using <kbd>^X</kbd> (or via <kbd>esc</kbd> +
@@ -41,7 +55,7 @@ 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
@@ -63,12 +77,16 @@ 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
-   to install some additional packages to download them (`pip install cson`)
-2. find a visual studio code theme, convert it to json (if it is not already
+1. find a visual studio code theme, convert it to json (if it is not already
    json) and put it at `~/.config/babi/theme.json`.  a helper script is
    provided to make this easier: `./bin/download-theme NAME URL`
 
+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
 
 most things work!  here's a few screenshots
@@ -79,7 +97,7 @@ 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 +10,10 @@ 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/buf.py

@@ -0,0 +1,315 @@
+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.expandtabs = True
+        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
+
+    @property
+    def tab_string(self) -> str:
+        if self.expandtabs:
+            return ' ' * self.tab_size
+        else:
+            return '\t'
+
+    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_manager.py

@@ -1,4 +1,3 @@
-import contextlib
 import curses
 from typing import Dict
 from typing import NamedTuple
@@ -34,12 +33,17 @@ class ColorManager(NamedTuple):
         return self.raw_color_pair(fg_i, bg_i)
 
     def raw_color_pair(self, fg: int, bg: int) -> int:
-        with contextlib.suppress(KeyError):
-            return self.raw_pairs[(fg, bg)]
+        if curses.COLORS > 0:
+            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)
-        return n
+            n = self.raw_pairs[(fg, bg)] = len(self.raw_pairs) + 1
+            curses.init_pair(n, fg, bg)
+            return n
+        else:
+            return 0
 
     @classmethod
     def make(cls) -> 'ColorManager':

babi/fdict.py

@@ -3,8 +3,10 @@ from typing import Iterable
 from typing import Mapping
 from typing import TypeVar
 
-TKey = TypeVar('TKey')
-TValue = TypeVar('TValue')
+from babi._types import Protocol
+
+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

@@ -1,21 +1,20 @@
-import contextlib
 import functools
 import json
 import os.path
 from typing import Any
 from typing import Dict
+from typing import FrozenSet
 from typing import List
 from typing import Match
 from typing import NamedTuple
 from typing import Optional
-from typing import Sequence
 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
@@ -68,6 +67,8 @@ class _Rule(Protocol):
     def include(self) -> Optional[str]: ...
     @property
     def patterns(self) -> 'Tuple[_Rule, ...]': ...
+    @property
+    def repository(self) -> 'FChainMap[str, _Rule]': ...
 
 
 @uniquely_constructed
@@ -84,9 +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')
@@ -96,7 +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:
@@ -104,7 +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:
@@ -112,7 +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:
@@ -120,7 +136,7 @@ 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:
@@ -142,7 +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 = ()
 
@@ -159,29 +175,33 @@ class Rule(NamedTuple):
             while_captures=while_captures,
             include=include,
             patterns=patterns,
+            repository=repository,
         )
 
 
 @uniquely_constructed
 class Grammar(NamedTuple):
     scope_name: str
+    repository: FChainMap[str, _Rule]
     patterns: Tuple[_Rule, ...]
-    repository: FDict[str, _Rule]
 
     @classmethod
-    def from_data(cls, data: Dict[str, Any]) -> 'Grammar':
+    def make(cls, data: Dict[str, Any]) -> 'Grammar':
         scope_name = data['scopeName']
-        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()
-            })
+            # 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(repository_dct)
+            for k, dct in data['repository'].items():
+                repository_dct[k] = Rule.make(dct, repository)
         else:
-            repository = FDict({})
+            repository = FChainMap()
+        patterns = tuple(Rule.make(d, repository) for d in data['patterns'])
         return cls(
             scope_name=scope_name,
-            patterns=patterns,
             repository=repository,
+            patterns=patterns,
         )
 
 
@@ -253,6 +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
 
@@ -264,7 +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
@@ -420,7 +441,8 @@ class EndRule(NamedTuple):
 
         boundary = match.end() == len(match.string)
         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
 
@@ -435,7 +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,
@@ -481,7 +512,9 @@ class WhileRule(NamedTuple):
 
         boundary = match.end() == len(match.string)
         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
 
@@ -521,7 +554,7 @@ class Compiler:
         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
@@ -531,22 +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':
             grammar = self._grammars.grammar_for_scope(self._root_scope)
-            return self._include(grammar, '$self')
+            return self._include(grammar, grammar.repository, '$self')
         elif s.startswith('#'):
-            return self._patterns(grammar, (grammar.repository[s[1:]],))
+            return self._patterns(grammar, (repository[s[1:]],))
         elif '#' not in s:
             grammar = self._grammars.grammar_for_scope(s)
-            return self._include(grammar, '$self')
+            return self._include(grammar, grammar.repository, '$self')
         else:
             scope, _, s = s.partition('#')
             grammar = self._grammars.grammar_for_scope(scope)
-            return self._include(grammar, f'#{s}')
+            return self._include(grammar, grammar.repository, f'#{s}')
 
     @functools.lru_cache(maxsize=None)
     def _patterns(
@@ -558,7 +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:
@@ -618,8 +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)
@@ -627,41 +665,58 @@ class Compiler:
 
 
 class Grammars:
-    def __init__(self, grammars: Sequence[Dict[str, Any]]) -> None:
-        self._raw = {grammar['scopeName']: grammar for grammar in grammars}
-        self._find_scope = [
-            (
-                frozenset(grammar.get('fileTypes', ())),
-                make_reg(grammar.get('firstLineMatch', '$impossible^')),
-                grammar['scopeName'],
-            )
-            for grammar in grammars
-        ]
-        self._parsed: Dict[str, Grammar] = {}
-        self._compilers: Dict[str, Compiler] = {}
+    def __init__(self, *directories: str) -> None:
+        self._scope_to_files = {
+            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
-    def from_syntax_dir(cls, syntax_dir: str) -> 'Grammars':
-        grammars = [{'scopeName': 'source.unknown', 'patterns': []}]
-        if os.path.exists(syntax_dir):
-            for filename in os.listdir(syntax_dir):
-                with open(os.path.join(syntax_dir, filename)) as f:
-                    grammars.append(json.load(f))
-        return cls(grammars)
+        unknown_grammar = {'scopeName': 'source.unknown', 'patterns': []}
+        self._raw = {'source.unknown': unknown_grammar}
+        self._file_types: List[Tuple[FrozenSet[str], str]] = []
+        self._first_line: List[Tuple[_Reg, str]] = []
+        self._parsed: Dict[str, Grammar] = {}
+        self._compiled: Dict[str, Compiler] = {}
+
+    def _raw_for_scope(self, scope: str) -> Dict[str, Any]:
+        try:
+            return self._raw[scope]
+        except KeyError:
+            pass
+
+        grammar_path = self._scope_to_files.pop(scope)
+        with open(grammar_path, encoding='UTF-8') 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:
-        with contextlib.suppress(KeyError):
+        try:
             return self._parsed[scope]
+        except KeyError:
+            pass
 
-        ret = self._parsed[scope] = Grammar.from_data(self._raw[scope])
+        raw = self._raw_for_scope(scope)
+        ret = self._parsed[scope] = Grammar.make(raw)
         return ret
 
     def compiler_for_scope(self, scope: str) -> Compiler:
-        with contextlib.suppress(KeyError):
-            return self._compilers[scope]
+        try:
+            return self._compiled[scope]
+        except KeyError:
+            pass
 
         grammar = self.grammar_for_scope(scope)
-        ret = self._compilers[scope] = Compiler(grammar, self)
+        ret = self._compiled[scope] = Compiler(grammar, self)
         return ret
 
     def blank_compiler(self) -> Compiler:
@@ -669,20 +724,26 @@ class Grammars:
 
     def compiler_for_file(self, filename: str, first_line: str) -> Compiler:
         for tag in tags_from_filename(filename) - {'text'}:
-            with contextlib.suppress(KeyError):
+            try:
+                # TODO: this doesn't always match even if we detect it
                 return self.compiler_for_scope(f'source.{tag}')
+            except KeyError:
+                pass
+
+        # 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, first_line_match, scope_name in self._find_scope:
-            if (
-                    ext in extensions or
-                    first_line_match.match(
-                        first_line, 0, first_line=True, boundary=True,
-                    )
-            ):
-                return self.compiler_for_scope(scope_name)
-        else:
-            return self.compiler_for_scope('source.unknown')
+        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')
 
 
 def highlight_line(

babi/history.py

@@ -19,7 +19,8 @@ class History:
         history_dir = xdg_data('history')
         os.makedirs(history_dir, exist_ok=True)
         for filename in os.listdir(history_dir):
-            with open(os.path.join(history_dir, filename)) as f:
+            history_filename = os.path.join(history_dir, filename)
+            with open(history_filename, encoding='UTF-8') as f:
                 self.data[filename] = f.read().splitlines()
                 self._orig_len[filename] = len(self.data[filename])
         try:
@@ -28,5 +29,6 @@ class History:
             for k, v in self.data.items():
                 new_history = v[self._orig_len[k]:]
                 if new_history:
-                    with open(os.path.join(history_dir, k), 'a+') as f:
+                    history_filename = os.path.join(history_dir, k)
+                    with open(history_filename, 'a+', encoding='UTF-8') as f:
                         f.write('\n'.join(new_history) + '\n')

babi/hl/interface.py

@@ -2,7 +2,7 @@ from typing import NamedTuple
 from typing import Tuple
 
 from babi._types import Protocol
-from babi.list_spy import SequenceNoSlice
+from babi.buf import Buf
 
 
 class HL(NamedTuple):
@@ -23,8 +23,8 @@ class FileHL(Protocol):
     def include_edge(self) -> bool: ...
     @property
     def regions(self) -> RegionsMapping: ...
-    def highlight_until(self, lines: SequenceNoSlice, idx: int) -> None: ...
-    def touch(self, lineno: int) -> None: ...
+    def highlight_until(self, lines: Buf, idx: int) -> None: ...
+    def register_callbacks(self, buf: Buf) -> None: ...
 
 
 class HLFactory(Protocol):

babi/hl/replace.py

@@ -4,9 +4,9 @@ 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
-from babi.list_spy import SequenceNoSlice
 
 
 class Replace:
@@ -15,10 +15,10 @@ class Replace:
     def __init__(self) -> None:
         self.regions: Dict[int, HLs] = collections.defaultdict(tuple)
 
-    def highlight_until(self, lines: SequenceNoSlice, idx: int) -> None:
+    def highlight_until(self, lines: Buf, idx: int) -> None:
         """our highlight regions are populated in other ways"""
 
-    def touch(self, lineno: int) -> None:
+    def register_callbacks(self, buf: Buf) -> None:
         """our highlight regions are populated in other ways"""
 
     @contextlib.contextmanager

babi/hl/selection.py

@@ -4,9 +4,9 @@ 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
-from babi.list_spy import SequenceNoSlice
 
 
 class Selection:
@@ -17,7 +17,10 @@ class Selection:
         self.start: Optional[Tuple[int, int]] = None
         self.end: Optional[Tuple[int, int]] = None
 
-    def highlight_until(self, lines: SequenceNoSlice, idx: 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
 
@@ -36,9 +39,6 @@ class Selection:
                 )
             self.regions[e_y] = (HL(x=0, end=e_x, attr=attr),)
 
-    def touch(self, lineno: int) -> None:
-        """our highlight regions are populated in other ways"""
-
     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:

babi/hl/syntax.py

@@ -1,9 +1,13 @@
 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
@@ -11,9 +15,9 @@ from babi.highlight import highlight_line
 from babi.highlight import State
 from babi.hl.interface import HL
 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
 
@@ -36,8 +40,10 @@ class FileSyntax:
         self.regions: List[HLs] = []
         self._states: List[State] = []
 
-        self._hl_cache: Dict[str, Dict[State, Tuple[State, HLs]]]
-        self._hl_cache = {}
+        # this will be assigned a functools.lru_cache per instance for
+        # 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)
@@ -48,19 +54,14 @@ class FileSyntax:
             curses.A_UNDERLINE * style.u
         )
 
-    def _hl(
+    def _hl_uncached(
             self,
             state: State,
             line: str,
-            i: int,
+            first_line: bool,
     ) -> 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
@@ -83,25 +84,42 @@ class FileSyntax:
             else:
                 regs.append(HL(x=r.start, end=r.end, attr=attr))
 
-        dct = self._hl_cache.setdefault(line, {})
-        ret = dct[state] = (new_state, tuple(regs))
-        return ret
+        return new_state, tuple(regs)
+
+    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
@@ -140,7 +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

@@ -1,10 +1,10 @@
 import curses
 from typing import List
 
+from babi.buf import Buf
 from babi.color_manager import ColorManager
 from babi.hl.interface import HL
 from babi.hl.interface import HLs
-from babi.list_spy import SequenceNoSlice
 
 
 class TrailingWhitespace:
@@ -30,9 +30,23 @@ class TrailingWhitespace:
             attr = curses.color_pair(pair)
             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:]

babi/horizontal_scrolling.py

@@ -1,3 +1,8 @@
+import curses
+
+from babi.cached_property import cached_property
+
+
 def line_x(x: int, width: int) -> int:
     if x + 1 < width:
         return 0
@@ -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

@@ -1,10 +1,15 @@
 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
@@ -13,10 +18,11 @@ 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, stdin: str) -> EditResult:
-    screen.file.ensure_loaded(screen.status, stdin)
+    screen.file.ensure_loaded(screen.status, screen.margin, stdin)
 
     while True:
         screen.status.tick(screen.margin)
@@ -39,36 +45,37 @@ def _edit(screen: Screen, stdin: str) -> EditResult:
 
 def c_main(
         stdscr: 'curses._CursesWindow',
-        args: argparse.Namespace,
+        filenames: List[Optional[str]],
+        positions: List[int],
         stdin: str,
+        perf: Perf,
 ) -> int:
-    with perf_log(args.perf_log) as perf:
-        screen = Screen(stdscr, args.filenames or [None], perf)
-        with screen.history.save():
-            while screen.files:
-                screen.i = screen.i % len(screen.files)
-                res = _edit(screen, stdin)
-                if res == EditResult.EXIT:
-                    del screen.files[screen.i]
-                    # always go to the next file except at the end
-                    screen.i = min(screen.i, len(screen.files) - 1)
-                    screen.status.clear()
-                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}')
+    screen = Screen(stdscr, filenames, positions, perf)
+    with screen.history.save():
+        while screen.files:
+            screen.i = screen.i % len(screen.files)
+            res = _edit(screen, stdin)
+            if res == EditResult.EXIT:
+                del screen.files[screen.i]
+                # always go to the next file except at the end
+                screen.i = min(screen.i, len(screen.files) - 1)
+                screen.status.clear()
+            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') -> int:
-    screen = Screen(stdscr, ['<<key debug>>'], Perf())
-    screen.file.lines = ['']
+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')
@@ -76,7 +83,7 @@ def _key_debug(stdscr: 'curses._CursesWindow') -> int:
         screen.file.move_cursor(screen.stdscr, screen.margin)
 
         key = screen.get_char()
-        screen.file.lines.insert(-1, f'{key.wch!r} {key.keyname.decode()!r}')
+        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()
@@ -84,6 +91,37 @@ def _key_debug(stdscr: 'curses._CursesWindow') -> int:
             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='*')
@@ -95,17 +133,23 @@ def main(argv: Optional[Sequence[str]] = None) -> int:
 
     if '-' in args.filenames:
         print('reading stdin...', file=sys.stderr)
-        stdin = sys.stdin.read()
+        stdin = sys.stdin.buffer.read().decode()
         tty = os.open(CONSOLE, os.O_RDONLY)
         os.dup2(tty, sys.stdin.fileno())
     else:
         stdin = ''
 
-    with make_stdscr() as stdscr:
+    # 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)
+            return _key_debug(stdscr, perf)
         else:
-            return c_main(stdscr, args, stdin)
+            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
-    footer: bool
+    lines: int
+    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(header=False, footer=False)
-        elif curses.LINES == 2:
-            return cls(header=False, footer=True)
-        else:
-            return cls(header=True, footer=True)
+        return cls(curses.LINES, curses.COLS)

babi/perf.py

@@ -36,7 +36,7 @@ class Perf:
     def save_profiles(self, filename: str) -> None:
         assert self._prof is not None
         self._prof.dump_stats(f'{filename}.pstats')
-        with open(filename, 'w') as f:
+        with open(filename, 'w', encoding='UTF-8') as f:
             f.write('μs\tevent\n')
             for name, duration in self._records:
                 f.write(f'{int(duration * 1000 * 1000)}\t{name}\n')

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:
-            prompt_s = f'{base[:curses.COLS - 7]}…: '
+        elif len(base) > self._screen.margin.cols - 6:
+            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

@@ -6,80 +6,36 @@ 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:
-    """replace the given escape sequences of `chars` with \\uffff"""
-    for c in chars:
-        if f'\\{c}' in s:
-            break
-    else:
-        return s
-
-    b = []
-    i = 0
-    length = len(s)
-    while i < length:
-        try:
-            sbi = s.index('\\', i)
-        except ValueError:
-            b.append(s[i:])
-            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)
+_FLAGS = {
+    # (first_line, boundary)
+    (False, False): (
+        onigurumacffi.OnigSearchOption.NOT_END_STRING |
+        onigurumacffi.OnigSearchOption.NOT_BEGIN_STRING |
+        onigurumacffi.OnigSearchOption.NOT_BEGIN_POSITION
+    ),
+    (False, True): (
+        onigurumacffi.OnigSearchOption.NOT_END_STRING |
+        onigurumacffi.OnigSearchOption.NOT_BEGIN_STRING
+    ),
+    (True, False): (
+        onigurumacffi.OnigSearchOption.NOT_END_STRING |
+        onigurumacffi.OnigSearchOption.NOT_BEGIN_POSITION
+    ),
+    (True, True): onigurumacffi.OnigSearchOption.NOT_END_STRING,
+}
 
 
 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,
@@ -87,7 +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,
@@ -96,36 +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,
@@ -133,16 +71,7 @@ class _RegSet:
             first_line: bool,
             boundary: bool,
     ) -> Tuple[int, Optional[Match[str]]]:
-        if boundary:
-            if first_line:
-                return self._set.search(line, pos)
-            else:
-                return self._set_no_A.search(line, pos)
-        else:
-            if first_line:
-                return self._set_no_G.search(line, pos)
-            else:
-                return self._set_no_A_no_G.search(line, pos)
+        return self._set.search(line, pos, flags=_FLAGS[first_line, boundary])
 
 
 def expand_escaped(match: Match[str], s: str) -> str:
@@ -151,4 +80,4 @@ def expand_escaped(match: Match[str], s: str) -> str:
 
 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

@@ -38,6 +38,8 @@ EditResult = enum.Enum('EditResult', 'EXIT NEXT PREV OPEN')
 SEQUENCE_KEYNAME = {
     '\x1bOH': b'KEY_HOME',
     '\x1bOF': b'KEY_END',
+    '\x1b[1~': b'KEY_HOME',
+    '\x1b[4~': b'KEY_END',
     '\x1b[1;2A': b'KEY_SR',
     '\x1b[1;2B': b'KEY_SF',
     '\x1b[1;2C': b'KEY_SRIGHT',
@@ -60,6 +62,42 @@ SEQUENCE_KEYNAME = {
     '\x1b[1;6D': b'kLFT6',  # Shift + ^Left
     '\x1b[1;6H': b'kHOM6',  # Shift + ^Home
     '\x1b[1;6F': b'kEND6',  # Shift + ^End
+    '\x1b[~': b'KEY_BTAB',  # Shift + Tab
+}
+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'^D': b'KEY_DC',
+    b'PADENTER': b'^M',  # Enter on numpad
 }
 
 
@@ -73,14 +111,15 @@ class Screen:
             self,
             stdscr: 'curses._CursesWindow',
             filenames: List[Optional[str]],
+            initial_lines: List[int],
             perf: Perf,
     ) -> None:
         self.stdscr = stdscr
         self.color_manager = ColorManager.make()
         self.hl_factories = (Syntax.from_screen(stdscr, self.color_manager),)
         self.files = [
-            File(filename, self.color_manager, self.hl_factories)
-            for filename in filenames
+            File(filename, line, self.color_manager, self.hl_factories)
+            for filename, line in zip(filenames, initial_lines)
         ]
         self.i = 0
         self.history = History()
@@ -105,7 +144,7 @@ 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)
 
@@ -191,7 +230,10 @@ 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:
@@ -202,12 +244,10 @@ 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:
@@ -225,7 +265,7 @@ class Screen:
     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()
 
     def quick_prompt(
@@ -233,16 +273,17 @@ class Screen:
             prompt: str,
             opt_strs: Tuple[str, ...],
     ) -> Union[str, PromptResult]:
-        opts = [opt[0] for opt in opt_strs]
+        opts = {opt[0] for opt in opt_strs}
         while True:
             x = 0
+            prompt_line = self.margin.lines - 1
 
             def _write(s: str, *, attr: int = curses.A_REVERSE) -> None:
                 nonlocal x
 
-                if x >= curses.COLS:
+                if x >= self.margin.cols:
                     return
-                self.stdscr.insstr(curses.LINES - 1, x, s, attr)
+                self.stdscr.insstr(prompt_line, x, s, attr)
                 x += len(s)
 
             _write(prompt)
@@ -254,23 +295,23 @@ class Screen:
                     _write(', ')
             _write(']?')
 
-            if x < curses.COLS - 1:
-                s = ' ' * (curses.COLS - x)
-                self.stdscr.insstr(curses.LINES - 1, x, s, curses.A_REVERSE)
+            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 = curses.COLS - 1
-                self.stdscr.insstr(curses.LINES - 1, x, '…', curses.A_REVERSE)
+                x = self.margin.cols - 1
+                self.stdscr.insstr(prompt_line, x, '…', curses.A_REVERSE)
 
-            self.stdscr.move(curses.LINES - 1, x)
+            self.stdscr.move(prompt_line, x)
 
             key = self.get_char()
             if key.keyname == b'KEY_RESIZE':
                 self.resize()
             elif key.keyname == b'^C':
                 return self.status.cancelled()
-            elif isinstance(key.wch, str) and key.wch in opts:
-                return key.wch
+            elif isinstance(key.wch, str) and key.wch.lower() in opts:
+                return key.wch.lower()
 
     def prompt(
             self,
@@ -317,9 +358,9 @@ class Screen:
                 self.file.go_to_line(lineno, self.margin)
 
     def current_position(self) -> None:
-        line = f'line {self.file.y + 1}'
-        col = f'col {self.file.x + 1}'
-        line_count = max(len(self.file.lines) - 1, 1)
+        line = f'line {self.file.buf.y + 1}'
+        col = f'col {self.file.buf.x + 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})')
 
@@ -358,7 +399,7 @@ 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()
 
@@ -384,7 +425,11 @@ 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()
@@ -397,7 +442,41 @@ class Screen:
             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.startswith(':expandtabs'):
+            for file in self.files:
+                file.buf.expandtabs = True
+            self.status.update('updated!')
+        elif response.startswith(':noexpandtabs'):
+            for file in self.files:
+                file.buf.expandtabs = False
+            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
 
@@ -416,12 +495,12 @@ 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, encoding='UTF-8', 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
@@ -429,12 +508,18 @@ class Screen:
             self.status.update('(file changed on disk, not implemented)')
             return PromptResult.CANCELLED
 
-        with open(self.file.filename, 'w') as f:
-            f.write(contents)
+        try:
+            with open(
+                self.file.filename, 'w', encoding='UTF-8', newline='',
+            ) as f:
+                f.write(contents)
+        except OSError as e:
+            self.status.update(f'cannot save file: {e}')
+            return PromptResult.CANCELLED
 
         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)')
 
@@ -458,7 +543,7 @@ class Screen:
     def open_file(self) -> Optional[EditResult]:
         response = self.prompt('enter filename', history='open')
         if response is not PromptResult.CANCELLED:
-            opened = File(response, self.color_manager, self.hl_factories)
+            opened = File(response, 0, self.color_manager, self.hl_factories)
             self.files.append(opened)
             return EditResult.OPEN
         else:
@@ -482,10 +567,13 @@ class Screen:
         return EditResult.EXIT
 
     def background(self) -> None:
-        curses.endwin()
-        os.kill(os.getpid(), signal.SIGSTOP)
-        self.stdscr = _init_screen()
-        self.resize()
+        if sys.platform == 'win32':  # pragma: win32 cover
+            self.status.update('cannot run babi in background on Windows')
+        else:  # pragma: win32 no cover
+            curses.endwin()
+            os.kill(os.getpid(), signal.SIGSTOP)
+            self.stdscr = _init_screen()
+            self.resize()
 
     DISPATCH = {
         b'KEY_RESIZE': resize,
@@ -495,6 +583,7 @@ class Screen:
         b'^U': uncut,
         b'M-u': undo,
         b'M-U': redo,
+        b'M-e': redo,
         b'^W': search,
         b'^\\': replace,
         b'^[': command,
@@ -510,7 +599,10 @@ class Screen:
 
 def _init_screen() -> 'curses._CursesWindow':
     # set the escape delay so curses does not pause waiting for sequences
-    if sys.version_info >= (3, 9):  # pragma: no cover
+    if (
+            sys.version_info >= (3, 9) and
+            hasattr(curses, 'set_escdelay')
+    ):  # pragma: no cover
         curses.set_escdelay(25)
     else:  # pragma: no cover
         os.environ.setdefault('ESCDELAY', '25')

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, encoding='UTF-8') 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, encoding='UTF-8') 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

@@ -147,5 +147,5 @@ class Theme(NamedTuple):
         if not os.path.exists(filename):
             return cls.blank()
         else:
-            with open(filename) as f:
+            with open(filename, encoding='UTF-8') as f:
                 return cls.from_dct(json.load(f))

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

@@ -1,86 +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('docker', Ext.PLIST, 'https://raw.githubusercontent.com/moby/moby/c7ad2b866/contrib/syntax/textmate/Docker.tmbundle/Syntaxes/Dockerfile.tmLanguage'),  # 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('make', Ext.PLIST, 'https://raw.githubusercontent.com/fadeevab/make.tmbundle/fd57c0552/Syntaxes/Makefile.plist'),  # 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('puppet', Ext.PLIST, 'https://raw.githubusercontent.com/lingua-pupuli/puppet-editor-syntax/dc414b8a/syntaxes/puppet.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
-    # TODO: https://github.com/atom/language-xml/pull/99
-    Syntax('xml', Ext.CSON, 'https://raw.githubusercontent.com/asottile/language-xml/2d76bc1f/grammars/xml.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

@@ -39,7 +39,6 @@ def json_with_comments(s: bytes) -> Any:
         idx = match.end()
         match = TOKEN.search(s, idx)
 
-    print(bio.getvalue())
     bio.seek(0)
     return json.load(bio)
 

setup.cfg

@@ -1,6 +1,6 @@
 [metadata]
 name = babi
-version = 0.0.2
+version = 0.0.21
 description = a text editor
 long_description = file: README.md
 long_description_content_type = text/markdown
@@ -16,14 +16,16 @@ classifiers =
     Programming Language :: Python :: 3.6
     Programming Language :: Python :: 3.7
     Programming Language :: Python :: 3.8
+    Programming Language :: Python :: 3.9
     Programming Language :: Python :: Implementation :: CPython
     Programming Language :: Python :: Implementation :: PyPy
 
 [options]
 packages = find:
 install_requires =
+    babi-grammars
     identify
-    onigurumacffi>=0.0.10
+    onigurumacffi>=0.0.18
     importlib_metadata>=1;python_version<"3.8"
     windows-curses;sys_platform=="win32"
 python_requires = >=3.6.1
@@ -31,6 +33,7 @@ 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/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 +1,205 @@
+import pytest
+
+from testing.runner import and_exit
+from testing.runner import trigger_command_mode
+
+
+@pytest.fixture
+def three_lines_with_indentation(tmpdir):
+    f = tmpdir.join('f')
+    f.write('line_0\n    line_1\n    line_2')
+    return f
+
+
+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_empty_line_trailing_whitespace(run, tmpdir):
+    f = tmpdir.join('f')
+    f.write('1\n\n2\n')
+
+    with run(str(f)) as h, and_exit(h):
+        h.press('S-Down')
+        h.press('S-Down')
+
+        trigger_command_mode(h)
+        h.press_and_enter(':comment')
+
+        h.await_text('# 1\n#\n# 2')
+
+
+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')
+
+
+def test_comment_some_code_with_indentation(run, three_lines_with_indentation):
+    with run(str(three_lines_with_indentation)) 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\n    line_2\n')
+
+        h.press('S-Up')
+        trigger_command_mode(h)
+        h.press_and_enter(':comment')
+
+        h.await_text('line_0\n    line_1\n    line_2\n')
+
+
+def test_comment_some_code_on_indent_part(run, three_lines_with_indentation):
+    with run(str(three_lines_with_indentation)) as h, and_exit(h):
+        h.press('Down')
+        h.press('S-Down')
+
+        trigger_command_mode(h)
+        h.press_and_enter(':comment')
+
+        h.await_text('line_0\n    # line_1\n    # line_2\n')
+
+        h.press('S-Up')
+
+        trigger_command_mode(h)
+        h.press_and_enter(':comment')
+
+        h.await_text('line_0\n    line_1\n    line_2\n')
+
+
+def test_comment_some_code_on_tabs_part(run, tmpdir):
+    f = tmpdir.join('f')
+    f.write('line_0\n\tline_1\n\t\tline_2')
+
+    with run(str(f)) as h, and_exit(h):
+        h.await_text('line_0\n    line_1\n        line_2')
+        h.press('Down')
+        h.press('S-Down')
+
+        trigger_command_mode(h)
+        h.press_and_enter(':comment')
+
+        h.await_text('line_0\n    # line_1\n    #   line_2')
+
+        h.press('S-Up')
+
+        trigger_command_mode(h)
+        h.press_and_enter(':comment')
+
+        h.await_text('line_0\n    line_1\n        line_2')
+
+
+def test_comment_cursor_at_end_of_line(run, ten_lines):
+    with run(str(ten_lines)) as h, and_exit(h):
+        h.press('# ')
+        h.press('End')
+        h.await_cursor_position(x=8, y=1)
+
+        trigger_command_mode(h)
+        h.press_and_enter(':comment')
+
+        h.await_cursor_position(x=6, y=1)
+
+
+def test_add_comment_moves_cursor(run, ten_lines):
+    with run(str(ten_lines)) as h, and_exit(h):
+        h.press('End')
+
+        h.await_cursor_position(x=6, y=1)
+
+        trigger_command_mode(h)
+        h.press_and_enter(':comment')
+
+        h.await_cursor_position(x=8, y=1)
+
+
+def test_do_not_move_if_cursor_before_comment(run, tmpdir):
+    f = tmpdir.join('f')
+    f.write('\t\tfoo')
+
+    with run(str(f)) as h, and_exit(h):
+        h.press('Right')
+
+        h.await_cursor_position(x=4, y=1)
+
+        trigger_command_mode(h)
+        h.press_and_enter(':comment')
+
+        h.await_cursor_position(x=4, y=1)
+
+
+@pytest.mark.parametrize('comment', ('# ', '#'))
+def test_remove_comment_with_comment_elsewhere_in_line(run, tmpdir, comment):
+    f = tmpdir.join('f')
+    f.write(f'{comment}print("not a # comment here!")\n')
+
+    with run(str(f)) as h, and_exit(h):
+        trigger_command_mode(h)
+        h.press_and_enter(':comment')
+
+        h.await_text('\nprint("not a # comment here!")\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 * ' '
-        self.insstr(self._runner.screen.y, self._runner.screen.x, s)
+        s = self._screen.width * ' '
+        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()
@@ -266,6 +290,7 @@ KEYS = [
     Key('^_', b'^_', '\x1f'),
     Key('^\\', b'^\\', '\x1c'),
     Key('!resize', b'KEY_RESIZE', curses.KEY_RESIZE),
+    Key('^D', b'^D', '\x04'),
 ]
 KEYS_TMUX = {k.tmux: k.wch for k in KEYS}
 KEYS_CURSES = {k.value: k.curses for k in KEYS}
@@ -279,6 +304,7 @@ class DeferredRunner:
         self.color_pairs = {0: (7, 0)}
         self.screen = Screen(width, height)
         self._n_colors, self._can_change_color = {
+            'xterm-mono': (0, False),
             'screen': (8, False),
             'screen-256color': (256, False),
             'xterm-256color': (256, True),
@@ -365,8 +391,9 @@ class DeferredRunner:
     def _curses__noop(self, *_, **__):
         pass
 
-    _curses_cbreak = _curses_noecho = _curses_nonl = _curses__noop
-    _curses_raw = _curses_use_default_colors = _curses__noop
+    _curses_cbreak = _curses_endwin = _curses_noecho = _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
 
@@ -392,11 +419,10 @@ class DeferredRunner:
 
     def _curses_initscr(self):
         self._curses_update_lines_cols()
-        self.screen.disabled = False
-        return CursesScreen(self)
+        return CursesScreen(self.screen, self)
 
-    def _curses_endwin(self):
-        self.screen.disabled = True
+    def _curses_newwin(self, height, width):
+        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/expandtabs_test.py

@@ -0,0 +1,45 @@
+from testing.runner import and_exit
+from testing.runner import trigger_command_mode
+
+
+def test_set_expandtabs(run, tmpdir):
+    f = tmpdir.join('f')
+    f.write('a')
+
+    with run(str(f)) as h, and_exit(h):
+        h.press('Left')
+        trigger_command_mode(h)
+        h.press_and_enter(':expandtabs')
+        h.await_text('updated!')
+        h.press('Tab')
+        h.press('^S')
+    assert f.read() == '    a\n'
+
+
+def test_set_noexpandtabs(run, tmpdir):
+    f = tmpdir.join('f')
+    f.write('a')
+
+    with run(str(f)) as h, and_exit(h):
+        h.press('Left')
+        trigger_command_mode(h)
+        h.press_and_enter(':noexpandtabs')
+        h.await_text('updated!')
+        h.press('Tab')
+        h.press('^S')
+    assert f.read() == '\ta\n'
+
+
+def test_indent_with_expandtabs(run, tmpdir):
+    f = tmpdir.join('f')
+    f.write('a\nb\nc')
+
+    with run(str(f)) as h, and_exit(h):
+        trigger_command_mode(h)
+        h.press_and_enter(':noexpandtabs')
+        h.await_text('updated!')
+        for _ in range(3):
+            h.press('S-Down')
+        h.press('Tab')
+        h.press('^S')
+    assert f.read() == '\ta\n\tb\n\tc\n'

tests/features/indent_test.py

@@ -1,4 +1,5 @@
 from testing.runner import and_exit
+from testing.runner import trigger_command_mode
 
 
 def test_indent_at_beginning_of_line(run):
@@ -12,11 +13,12 @@ def test_indent_at_beginning_of_line(run):
 
 def test_indent_not_full_tab(run):
     with run() as h, and_exit(h):
-        h.press('h')
+        h.press('hello')
+        h.press('Home')
+        h.press('Right')
         h.press('Tab')
-        h.press('ello')
         h.await_text('h   ello')
-        h.await_cursor_position(x=8, y=1)
+        h.await_cursor_position(x=4, y=1)
 
 
 def test_indent_fixes_eof(run):
@@ -86,6 +88,20 @@ def test_dedent_selection(run, tmpdir):
         h.await_text('\n1\n2\n    3\n')
 
 
+def test_dedent_selection_with_noexpandtabs(run, tmpdir):
+    f = tmpdir.join('f')
+    f.write('1\n\t2\n\t\t3\n')
+    with run(str(f)) as h, and_exit(h):
+        trigger_command_mode(h)
+        h.press_and_enter(':noexpandtabs')
+        h.await_text('updated!')
+        for _ in range(3):
+            h.press('S-Down')
+        h.press('BTab')
+        h.press('^S')
+    assert f.read() == '1\n2\n\t3\n'
+
+
 def test_dedent_beginning_of_line(run, tmpdir):
     f = tmpdir.join('f')
     f.write('    hi\n')

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/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/replace_test.py

@@ -30,7 +30,8 @@ def test_replace_cancel_at_replace_string(run):
         h.await_text('cancelled')
 
 
-def test_replace_actual_contents(run, ten_lines):
+@pytest.mark.parametrize('key', ('y', 'Y'))
+def test_replace_actual_contents(run, ten_lines, key):
     with run(str(ten_lines)) as h, and_exit(h):
         h.press('^\\')
         h.await_text('search (to replace):')
@@ -38,7 +39,7 @@ def test_replace_actual_contents(run, ten_lines):
         h.await_text('replace with:')
         h.press_and_enter('ohai')
         h.await_text('replace [yes, no, all]?')
-        h.press('y')
+        h.press(key)
         h.await_text_missing('line_0')
         h.await_text('ohai')
         h.await_text(' *')
@@ -272,3 +273,31 @@ def test_replace_separate_line_after_wrapping(run, ten_lines):
         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')
@@ -102,12 +128,24 @@ def test_save_file_when_it_did_not_exist(run, tmpdir):
     assert f.read() == 'hello world\n'
 
 
+def test_saving_file_permission_denied(run, tmpdir):
+    f = tmpdir.join('f').ensure()
+    f.chmod(0o400)
+
+    with run(str(f)) as h, and_exit(h):
+        h.press('hello world')
+        h.press('^S')
+        # the filename message is missing as it is too long to be captured
+        h.await_text('cannot save file: [Errno 13] Permission denied:')
+        h.await_text(' *')
+
+
 def test_save_via_ctrl_o(run, tmpdir):
     f = tmpdir.join('f')
     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'
@@ -189,3 +227,38 @@ def test_save_on_exit_resize(run, tmpdir):
         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/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
@@ -97,3 +99,62 @@ def test_syntax_highlighting_off_screen_does_not_crash(run, tmpdir):
         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')
+
+
+def test_does_not_crash_with_no_color_support(run):
+    with run(term='xterm-mono') as h, and_exit(h):
+        pass

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

@@ -50,6 +50,18 @@ def test_backspace_at_end_of_file_still_allows_scrolling_down(run, tmpdir):
         h.await_text_missing('*')
 
 
+def test_backspace_deletes_newline_at_end_of_file(run, tmpdir):
+    f = tmpdir.join('f')
+    f.write('foo\n\n')
+
+    with run(str(f)) as h, and_exit(h):
+        h.press('^End')
+        h.press('BSpace')
+        h.press('^S')
+
+    assert f.read() == 'foo\n'
+
+
 @pytest.mark.parametrize('key', ('BSpace', '^H'))
 def test_backspace_deletes_text(run, tmpdir, key):
     f = tmpdir.join('f')
@@ -72,14 +84,15 @@ def test_delete_at_end_of_file(run, tmpdir):
         h.await_text_missing('*')
 
 
-def test_delete_removes_character_afterwards(run, tmpdir):
+@pytest.mark.parametrize('key', ('DC', '^D'))
+def test_delete_removes_character_afterwards(run, tmpdir, key):
     f = tmpdir.join('f')
     f.write('hello world')
 
     with run(str(f)) as h, and_exit(h):
         h.await_text('hello world')
         h.press('Right')
-        h.press('DC')
+        h.press(key)
         h.await_text('hllo world')
         h.await_text('f *')
 
@@ -97,6 +110,24 @@ 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(' *')

tests/file_test.py

@@ -8,7 +8,7 @@ from babi.file import get_lines
 
 
 def test_position_repr():
-    ret = repr(File('f.txt', ColorManager.make(), ()))
+    ret = repr(File('f.txt', 0, ColorManager.make(), ()))
     assert ret == "<File 'f.txt'>"
 
 

tests/highlight_test.py

@@ -1,34 +1,37 @@
-from babi.highlight import Grammars
+import pytest
+
 from babi.highlight import highlight_line
 from babi.highlight import Region
 
 
-def test_grammar_matches_extension_only_name():
+def test_grammar_matches_extension_only_name(make_grammars):
     data = {'scopeName': 'shell', 'patterns': [], 'fileTypes': ['bashrc']}
-    grammars = Grammars([data])
+    grammars = make_grammars(data)
     compiler = grammars.compiler_for_file('.bashrc', 'alias nano=babi')
     assert compiler.root_state.entries[0].scope[0] == 'shell'
 
 
-def test_grammar_matches_via_identify_tag():
-    data = {'scopeName': 'source.ini', 'patterns': []}
-    grammars = Grammars([data])
+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'
 
 
-def _compiler_state(*grammar_dcts):
-    grammars = Grammars(grammar_dcts)
-    compiler = grammars.compiler_for_scope(grammar_dcts[0]['scopeName'])
-    return compiler, compiler.root_state
+@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():
+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)
@@ -51,8 +54,8 @@ BEGIN_END_NO_NL = {
 }
 
 
-def test_backslash_g_inline():
-    compiler, state = _compiler_state(BEGIN_END_NO_NL)
+def test_backslash_g_inline(compiler_state):
+    compiler, state = compiler_state(BEGIN_END_NO_NL)
 
     _, regions = highlight_line(compiler, state, 'xaax', True)
     assert regions == (
@@ -63,8 +66,8 @@ def test_backslash_g_inline():
     )
 
 
-def test_backslash_g_next_line():
-    compiler, state = _compiler_state(BEGIN_END_NO_NL)
+def test_backslash_g_next_line(compiler_state):
+    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)
@@ -81,8 +84,8 @@ def test_backslash_g_next_line():
     )
 
 
-def test_end_before_other_match():
-    compiler, state = _compiler_state(BEGIN_END_NO_NL)
+def test_end_before_other_match(compiler_state):
+    compiler, state = compiler_state(BEGIN_END_NO_NL)
 
     state, regions = highlight_line(compiler, state, 'xazzx', True)
 
@@ -107,8 +110,8 @@ BEGIN_END_NL = {
 }
 
 
-def test_backslash_g_captures_nl():
-    compiler, state = _compiler_state(BEGIN_END_NL)
+def test_backslash_g_captures_nl(compiler_state):
+    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)
@@ -124,8 +127,8 @@ def test_backslash_g_captures_nl():
     )
 
 
-def test_backslash_g_captures_nl_next_line():
-    compiler, state = _compiler_state(BEGIN_END_NL)
+def test_backslash_g_captures_nl_next_line(compiler_state):
+    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)
@@ -147,8 +150,8 @@ def test_backslash_g_captures_nl_next_line():
     )
 
 
-def test_while_no_nl():
-    compiler, state = _compiler_state({
+def test_while_no_nl(compiler_state):
+    compiler, state = compiler_state({
         'scopeName': 'test',
         'patterns': [{
             'begin': '> ',
@@ -182,8 +185,8 @@ def test_while_no_nl():
     )
 
 
-def test_complex_captures():
-    compiler, state = _compiler_state({
+def test_complex_captures(compiler_state):
+    compiler, state = compiler_state({
         'scopeName': 'test',
         'patterns': [
             {
@@ -213,8 +216,8 @@ def test_complex_captures():
     )
 
 
-def test_captures_multiple_applied_to_same_capture():
-    compiler, state = _compiler_state({
+def test_captures_multiple_applied_to_same_capture(compiler_state):
+    compiler, state = compiler_state({
         'scopeName': 'test',
         'patterns': [
             {
@@ -256,8 +259,8 @@ def test_captures_multiple_applied_to_same_capture():
     )
 
 
-def test_captures_ignores_empty():
-    compiler, state = _compiler_state({
+def test_captures_ignores_empty(compiler_state):
+    compiler, state = compiler_state({
         'scopeName': 'test',
         'patterns': [{
             'match': '(.*) hi',
@@ -279,8 +282,8 @@ def test_captures_ignores_empty():
     )
 
 
-def test_captures_ignores_invalid_out_of_bounds():
-    compiler, state = _compiler_state({
+def test_captures_ignores_invalid_out_of_bounds(compiler_state):
+    compiler, state = compiler_state({
         'scopeName': 'test',
         'patterns': [{'match': '.', 'captures': {'1': {'name': 'oob'}}}],
     })
@@ -292,8 +295,8 @@ def test_captures_ignores_invalid_out_of_bounds():
     )
 
 
-def test_captures_begin_end():
-    compiler, state = _compiler_state({
+def test_captures_begin_end(compiler_state):
+    compiler, state = compiler_state({
         'scopeName': 'test',
         'patterns': [
             {
@@ -314,8 +317,8 @@ def test_captures_begin_end():
     )
 
 
-def test_captures_while_captures():
-    compiler, state = _compiler_state({
+def test_captures_while_captures(compiler_state):
+    compiler, state = compiler_state({
         'scopeName': 'test',
         'patterns': [
             {
@@ -343,8 +346,8 @@ def test_captures_while_captures():
     )
 
 
-def test_captures_implies_begin_end_captures():
-    compiler, state = _compiler_state({
+def test_captures_implies_begin_end_captures(compiler_state):
+    compiler, state = compiler_state({
         'scopeName': 'test',
         'patterns': [
             {
@@ -364,8 +367,8 @@ def test_captures_implies_begin_end_captures():
     )
 
 
-def test_captures_implies_begin_while_captures():
-    compiler, state = _compiler_state({
+def test_captures_implies_begin_while_captures(compiler_state):
+    compiler, state = compiler_state({
         'scopeName': 'test',
         'patterns': [
             {
@@ -392,8 +395,8 @@ def test_captures_implies_begin_while_captures():
     )
 
 
-def test_include_self():
-    compiler, state = _compiler_state({
+def test_include_self(compiler_state):
+    compiler, state = compiler_state({
         'scopeName': 'test',
         'patterns': [
             {
@@ -416,8 +419,8 @@ def test_include_self():
     )
 
 
-def test_include_repository_rule():
-    compiler, state = _compiler_state({
+def test_include_repository_rule(compiler_state):
+    compiler, state = compiler_state({
         'scopeName': 'test',
         'patterns': [{'include': '#impl'}],
         'repository': {
@@ -438,8 +441,40 @@ def test_include_repository_rule():
     )
 
 
-def test_include_other_grammar():
-    compiler, state = _compiler_state(
+def test_include_with_nested_repositories(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': [
@@ -494,8 +529,8 @@ def test_include_other_grammar():
     )
 
 
-def test_include_base():
-    compiler, state = _compiler_state(
+def test_include_base(compiler_state):
+    compiler, state = compiler_state(
         {
             'scopeName': 'test',
             'patterns': [
@@ -542,8 +577,8 @@ def test_include_base():
     )
 
 
-def test_rule_with_begin_and_no_end():
-    compiler, state = _compiler_state({
+def test_rule_with_begin_and_no_end(compiler_state):
+    compiler, state = compiler_state({
         'scopeName': 'test',
         'patterns': [
             {
@@ -566,8 +601,8 @@ def test_rule_with_begin_and_no_end():
     )
 
 
-def test_begin_end_substitute_special_chars():
-    compiler, state = _compiler_state({
+def test_begin_end_substitute_special_chars(compiler_state):
+    compiler, state = compiler_state({
         'scopeName': 'test',
         'patterns': [{'begin': r'(\*)', 'end': r'\1', 'name': 'italic'}],
     })
@@ -579,3 +614,48 @@ def test_begin_end_substitute_special_chars():
         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

@@ -4,8 +4,9 @@ from unittest import mock
 
 import pytest
 
+from babi.buf import Buf
 from babi.color_manager import ColorManager
-from babi.highlight import Grammars
+from babi.hl.interface import HL
 from babi.hl.syntax import Syntax
 from babi.theme import Color
 from babi.theme import Theme
@@ -71,8 +72,8 @@ THEME = Theme.from_dct({
 
 
 @pytest.fixture
-def syntax(tmpdir):
-    return Syntax(Grammars.from_syntax_dir(tmpdir), THEME, ColorManager.make())
+def syntax(make_grammars):
+    return Syntax(make_grammars(), THEME, ColorManager.make())
 
 
 def test_init_screen_low_color(stdscr, syntax):
@@ -149,3 +150,20 @@ def test_style_attributes_applied(stdscr, syntax):
         style = THEME.select(('keyword.python',))
         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'