Pyspelling

- **FIX**: CSS4 allows `:not()`, `:has()`, and `:is()` to be nested in `:not()`.

- **NEW**: Add support for `div p`, `div>p`, `div+p`, `div~p` in the HTML/XML filter's CSS selectors. (51)
- **NEW**: Add support for the `:root` CSS selector. (57)
- **NEW**: Add support for experimental `:has()` selector. (54)
- **FIX**: According to CSS4 specification, `:is()` is the final name for `:matches()` but the `:matches()` is an allowed alias. (53)
- **FIX**: Allow `:not()` to be nested in `:is()`/`:matches()`. (56)

- **NEW**: (Breaking change) Task names should be unique and using `--name` from the command line will only target one `name` (the last task defined with that name). If you were not using `name` to run a group of tasks, you will not notice any changes.
- **NEW**: Task option `group` has been added to target multiple tasks with the `--group` command line option. `group` name can be shared across different tasks.
- **NEW**: Add XML filter (PySpelling now has a dependency on `lxml`).
- **NEW**: Add Open Document Format (ODF) filter for `.odt`, `.ods`, and `.odp` files.
- **NEW**: Add Office Open XML format (newer Microsoft document format) for `.docx`, `.xlsx`, and `.pptx` files.
- **NEW**: CSS selectors in XML and HTML filters now support `:not()` and `:matches()` pseudo class.
- **NEW**: CSS selectors now support `,` in patterns.
- **NEW**: CSS selectors now support `i` in attribute selectors: `[attr=value i]`.
- **NEW**: CSS selectors now support namespaces (some configuration required).
- **NEW**: For better HTML context, display a tag's ancestry (just tag name of parents).
- **NEW**: Captured tags are now configurable via `captures`, but tags that are not captured still have their children crawled unless they are under `ignores`.
- **NEW**: Support modes added for HTML filter: `html`, `html5`, and `xhtml`.
- **NEW**: `CHECK_BOM` plugin attribute has been deprecated in favor of overriding the exposed `has_bom` function.
- **NEW**: Tasks can be hidden with the `hidden` configuration option. Tasks with `hidden` enabled will only run if they are explicitly called by name.
- **NEW**: Add normal string support to Python filter.
- **NEW**: Add string and template literal support for JavaScript filter.
- **NEW**: Add string support for CPP filter.
- **NEW**: Add `generic_mode` option to CPP to allow for generic C/C++ comment style capture from non C/C++ file types.
- **NEW**: Context will normalize line endings before applying context (can be disabled).
- **NEW**: CPP, Stylesheet, and JavaScript plugins now normalize line endings of block comments.
- **NEW**: UTF-16 and UTF-32 is not really supported by Aspell and Hunspell, so at the end of the pipeline, Unicode strings that have the associated encoding of UTF-16 or UTF-32 will encoding with the compatible UTF-8. This does not apply to files being processed with a disabled pipeline. When the pipeline is disabled, files are sent directly to the spell checker with no modifications.
- **FIX**: Case related issues when comparing tags and attributes in HTML.
- **FIX**: CSS selectors should only compare case insensitive for ASCII characters A-Z and a-z.
- **FIX**: Allow CSS escapes in selectors.
- **FIX**: Don't send empty (or strings that are just whitespace) to spell checker to prevent Aspell 0.50 series from crashing (also to increase performance).
- **FIX**: Catch and bubble up errors better.
- **FIX**: Fix issue where Python module docstrings would not get spell checked if they followed a shebang.

- **NEW**: Add URL/email address filter. (30)
- **NEW**: If `pipeline` configuration key is set to `null`, do not use any filters, and send the filename, not the content, to the spell checker.
- **NEW**: Add `encoding` option to `dictionary` configuration for the purpose of communicating what encoding the main dictionary is when compiling wordlists (only Aspell takes advantage of this).
- **FIX**: Fix Hunspell `-O` option which was mistakenly `-o`. (31)

- **NEW**: Allow multiple names on command line via: `pyspelling -n name1 -n name2`.
- **FIX**: Fix empty HTML tags not properly having their attributes evaluated.
- **FIX**: Fix case where a deprecation warning for `filters` is shown when it shouldn't.
- **FIX**: Better docstring recognition in Python filter.
- **FIX**: Catch comments outside of the `<HTML>` tag.
- **FIX**: Filter out `Doctype`, `CData`, and other XML or non-content type information.

- **FIX**: Fix CPP comment regular expression.