* Fix parsing of redirect tokens
In the parser, wrap redirect tokens such as ">", "a>", and "2>1" in a
tuple (along with the name of the redirect file, if applicable) to
differentiate them from regular command line arguments.
This ensures that a command like ![echo ">"] is parsed correctly.
We now need 2 types of IO-redirect tokens (IOREDIRECT1 and IOREDIRECT2)
so that the parser knows whether or not an IO-redirect has a file
argument. For example, "a>" has type IOREDIRECT1, whereas "2>1" has
type IOREDIRECT2.
* update ioredir tests
* add tests
* add news
* fix multiple redirections error message
* add tests
* Fix incorrect IOREDIRECT tokens in Python mode
* add news
* Update lexer used in CompletionContextParser
The completion context lexer should start in subprocess mode so that
operators like ``2>`` are recognized as IOREDIRECT tokens.
This fixes the test case "command 2>/dev/nul{X}" for
test_completion_context.py::test_command.
* refactor: remove usage of global variables in abbrevs.py
* chore: add flake8-mutable to prevent mutable defaults
* fix: abbrevs expand test
* refactor: add xonsh session singleton
* refactor: fix circular errors when using xonshSession as singleton
* refactor: remove black magicked builtin attributes
* style: black format tests as well
* refactor: update tests to use xonsh-session singleton
* refactor: update abbrevs to not use builtins
* test: remove DummyCommandsCache and patch orig class
* fix: failing test_command_completers
* test: use monkeypatch to update xession fixture
* fix: failing test_pipelines
* fix: failing test_main
* chore: run test suit as single invocation
* test: fix tests/test_xonsh.xsh
* refactor: remove builtins from docs/conf.py
* fix: mypy error in jobs
* fix: test error from test_main
* test: close xession error in test_command_completers
* chore: use pytest-cov for reporting coverage
this will include subprocess calls, and will increase coverage
* style:
* completion-context: Add CompletionContextParser placeholder
Implements the xonsh (tab-)completion context parser.
This parser is meant to parse a (possibly incomplete) command line.
* completers: tools: Implement ``contextual_completer`` decorator
This is used to mark completers that want to use the parsed completion context.
* completers: Enable using contextual completers in xonsh/completer.py
* completers: readline, ptk, jupyter: Enable using contextual completers
Pass ``multiline_text`` and ``cursor_index`` to ``Completer.complete()``
* parsers: base: Refactor out a ``raise_parse_error`` function
* tokenize: Enable ``tolerant`` mode
If ``tolerant`` is True, yield ERRORTOKEN instead of
throwing an exception when encountering an error.
* lexer: Enable ``tolerant`` mode
Tokenize without extra checks (e.g. paren matching).
When True, ERRORTOKEN contains the erroneous string instead of an error msg.
* tests: lexer: Test ``tolerant`` mode
* completion-context: Implement simple CommandContext parsing
* completion-context: tests: Test simple CommandContext parsing
* completion-context: Implement parsing sub-commands
* completion-context: tests: Test parsing sub-commands
* completion-context: Add news file
* completion-context: parser: Add parser table path to relevant locations
Code-coverage, mypy ignore list, etc.
* completion-context: Implement parsing partial strings and line continuations
* completion-context: tests: Test parsing partial strings and line continuations
* completion-context: Convert ``Span`` object to a ``slice``
* completion-context: Refactor out ``create_command`` and ``cursor_in_span``
* completion-context: Implement handling empty commands
* completion-context: tests: Test handling empty commands
* completion-context: Implement handling multiple commands
Separated by newlines, `;`, `and`, `or`, `|`, `&&`, `||`
* completion-context: tests: Test handling multiple commands
Separated by newlines, `;`, `and`, `or`, `|`, `&&`, `||`
* completion-context: Implement handling python context
* completion-context: tests: Test handling python context
* completers: tools: Add `contextual_command_completer`
* completers: Make `complete_skipper` contextual
* completers: Make `complete_from_man` contextual
* completers: Make `complete_from_bash` contextual and add test
* completers: Make `complete_pip` contextual and update tests
* completers: Keep opening string quote if it exists
* completion-context: Handle cursor after a closing quote
For example - cursor at the end of ``ls "/usr/"``.
1. The closing quote will be appended to all completions.
I.e the completion ``/usr/bin`` will turn into ``/usr/bin"``
2. If not specified, lprefix will cover the closing prefix.
I.e for ``ls "/usr/"``, the default lprefix will be 6 to include the closing quote.
* completion-context: tests: Test handling cursor after a closing quote
* completion-context: Fix bug with multiple empty commands
e.g. `;;;`
* completion-context: tests: Speed up tests
From ~15 seconds to ~500 ms
* completion-context: Expand commands and subcommands
* completion-context: Simplify `commands` rules
* completion-context: Simplify `sub_expression` rules
* completion-context: Simplify editing a multi-command token
* completion-context: Inline `create_command`
* completion-context: Implement `contextual_command_completer_for` helper
* completers: Make `complete_cd`/`complete_rmdir` contextual and add tests
* completers: path: Don't append a double-backslash in a raw string
When completing a path, if a raw string is used (e.g. `r"C:\Windows\"`),
there's no reason to append a double-backslash (e.g. `r"C:\Windows\\"`).
* completers: Make `complete_xonfig`/`complete_xontrib` contextual and add tests
* completers: Make `complete_completer` contextual and add tests
* completers: Make `complete_import` contextual and add tests
* completion-context: Add python `ctx` attribute
* completion: tools: Simplify `RichCompletion` attributes handling
* completers: Make `base`, `python`, and `commands` contextual
* Add tests
* No need for `python_mode` completer anymore
* completion: tools: Add `append_space` attribute to `RichCompletion`
* completion-context: Get all lines in a main python context
* xontrib: jedi: Make the `jedi` completer contextual
* completers: tools: Remove `get_ptk_completer` and `PromptToolkitCompleter.current_document`
These aren't needed anymore now that contextual completers can access the multiline code (via `PythonContext.multiline_code`).
* completion-context: ptk: Expand aliases
* completion-context: jupyter: Expand aliases and fix line handling
* completer: Preserve custom prefix after closing quote
* completers: bash: Ensure bash completion uses the complete prefix
* completers: pip: Append a space after a pip command
* completers: pip: Prevent bad package name completions
* completers: Remove a common prefix from `RichCompletion` if `display` wasn't provided
* completion-context: Treat cursor at edge of `&& || | ;` as normal args
This will be used for completing a space
* completers: Complete end proc keywords correctly
I wouldn't normally do something like this but issue #487 brought to
my attention the fact that too many of the python modules don't have
an encoding comment and of those that do there is a lot of pointless
inconsistency in the style of the comment.
