- Remove explicit did_you_mean dependency (if you want the suggestions in ruby 2.5 & 2.6, you'll need to add did_you_mean to your Gemfile directly).
- require 'set' in Spellr::Reporter previously we were just were relying on FastIgnore requiring it, and i don't use it there any longer.
- Resolve fast_ignore follow_symlinks deprecation
- Drop ruby 2.4 support, to allow for...
- Spelling suggestions while using
spellr --interactive
- And a new, probably frequently wrong,
spellr --autocorrect
- Assume all files are utf8, more comprehensively. (Sets ::Encoding.default_external and default_internal while running)
- Recognize url with _ in query string and zero length path
- Assume all files are utf8
- output a suggested
spellr --interactive
command with filenames, when running this without --interactive
- Recognize URL with tilde in path
--suppress--file-rules
so you can check files that would usually be ignored
- Single line disable! use
spellr:disable-line
#25
- Update fast_ignore dependency, it's faster now (about 0.5s faster on rails codebase)
- Update fast_ignore dependency fixing symlinks to directories handling.
- Massive test refactor
- Spellr now only pays attention to Spellr.pwd. Dir.pwd can be whatever
- All output goes through Spellr.config.output. Now we can override it.
- tests are twice as fast. Fewer warnings
- upgrade fast_ignore dependency
- use refinements for backports so that RakeTask doesn't conflict with Rubocop::RakeTask
- add the ability to use spellr as a rake task
- relax fast_ignore requirement
- dry_run checks config validity
- require fast_ignore 0.6.0
- new interactive UI
- misc performance improvements
- add CSS wordlist from MDN
- improve html wordlist comprehensiveness from MDN and W3C
- add ruby 2.7 words
- update fast_ignore requirement. it's slightly faster.
- misc other performance improvements
- require Parallel dependency in gemspec (oops)
- Check files in parallel
- Lots of pure refactoring
- Capfile is now considered a ruby file by default
- Removed the fetch command it was just unnecessarily slow and awkward. instead use
locale: [US,AU]
- Added usage documentation
- Fixed an issue where file-specific wordlists couldn't be added to
- fix the private method 'touch' issue when generating wordlists
- fix the js/javascript defaults being named differently (now consistently is named javascript)
- add .jsx.snap and .tsx.snap as extensions using the javascript wordlists
- LOTS of performance improvements. it's about 4 times faster
- significantly better key heuristic matching, with configurable weight (
key_heuristic_weight
). - Update FastIgnore dependency.
- Change the yml format slightly.
ignore
is nowexcludes
.only
is nowincludes
I feel like this makes more sense for the way the config is merged. and the right time to do it is when you'll probably have to tweak it anyway because: - the
only
/includes
items are now parsed using FastIgnore's gitignore inspired allow list format (see https://github.com/robotdana/fast_ignore#using-an-includes-list) Mostly it's the same just more flexible, though there may need to be some small adjustments. - the cli arguments are now also managed using FastIgnore's rules, fixing issues with absolute paths and paths beginning with
./
also it's a LOT faster when just checking a single file, basically instant. so that's nice.
- add automatic rubygems and dockerhub deploy
- remove unnecessary files
- interactive add to wordlist uses consistent letters rather than random numbers as keys
- additional wordlists per language can no longer be defined
- Supports ruby 2.3 - 2.6
- Initial Release