* Update `remark-parse` to v7
* fix `link` title and `inlineCode`
* fix `code` meta
* fix code meta
* fix ListItem print
* fix heading print result
* remove blank line before childList
* trim `inlineCode` ast
* use List.spread
* use ListItem.spread
* fix `example-550.md` snap
* update `example-522.md` snap
* update `example-508.md` snap
* update `example-509.md` snap
* update `example-222.md` snap
* Update listitem print
* lower case definition.label to match ast
* trim definition.label to match ast
* trim definition.label to match ast
* unescape Link.title
* filter link with no title
* fix link title
* fix link title
* restore preprocess
* fix ast massage
* Restore list
* Update to remark-parse 8.0.2
* Update snapshot tests after rebase
* Create legacy `loose` attribute from `spread`
Uses @duailibe's suggestion from here:
https://github.com/prettier/prettier/issues/6180#issuecomment-499498432
Also backtracks some other unnecessary changes.
* Update test case for shortcut reference-style links
The commonmark spec does not allow for whitespace between the left and right
parts of a full reference-style link.
https://spec.commonmark.org/dingus/?text=%5Bfoo%5D%20%5Bbar%5D%0A%0A%5Bbar%5D%3A%20%2Furl%20%22title%22%0A
Notice that the `[foo]` is parsed as plain text, while the `[bar]` is parsed
as a shortcut reference link.
When there is a space between the two, each part is parsed as a separate
shortcut style reference link node:
https://astexplorer.net/#/gist/d69f5a227e2cbde3f43f0a639bee165c/822b531c252c162faba78e38607db73ef72886d5
Previously they would have been parsed as a single full reference link.
Here I update the test-case to account for the shortcut style links.
* Add remark-footnotes plugin to replace legacy footnotes option
Remark removed the footnotes option in favor of a new remark-footnotes
plugin in https://github.com/remarkjs/remark/pull/483
Here I've simply switched to use the plugin instead of the removed
option and updated the snapshot tests.
* Add test case for #4369
* Do not add gap if has leading space
* Restore `inlineCode` ast compare
* Restore `isPrevNodeLooseListItem`
* Ignore some ast property
* Restore blank line, Fix `isPrevNodeLooseListItem` check
* Title compare
* Ignore `loose`
* Fix footnoteDefinition
* Fix tabWidth
* Style
* `footnoteDefinition` is stable now
* Fix `_` after link
* Revert workaround, url tests
* Fix idempotence issue with underscores and autolinks
* Fix idempotence issue with underscores and autolinks - attempt # 2
* Fix idempotence issue with underscores and autolinks - attempt # 3
* Fix idempotence issue with underscores and autolinks - part # 4
* More tests
* Code style
* Only test master passed tests
* Update list
* Update list
* Use code as test name
* Fix idempotence issue with underscores and autolinks - part # 5
* Fix idempotence issue with underscores and autolinks - part # 6
* Linting fix
* More autolink tests
* Fix tests
* Rename to `auto-link`
* More tests
* Disable new cases
* tests for issue 4122
* add changelog
Co-authored-by: fisker <lionkay@gmail.com>
Co-authored-by: Georgii Dolzhykov <thorn.mailbox@gmail.com>
* Quote and unquote number keys
* Add changelog entry
* Add tests for escapes
* Fix idempotency
* Only run flow-repo/ tests with flow
271 jsfmt.spec.js in flow-repo use only the flow parser.
18 use flow and babel.
Now, all of them use only flow.
* Don’t unquote negative numbers
* Don’t quote/unquote numbers in TypeScript
* Quote BigInt
* Fix AST check
* Fix condition
* Fix BigInt to string
* Fix ESLint not recognizing BigInt
* Only quote/unquote “simple” numbers
* No need to clean BigIntLiteral anymore
* Add tests for 1. and .1
* Fix 1. and .1
* Remove unnecessary babel-ts from tests
* Only unquote numbers for babel
* Update changelog
* Modify to align binaryish expression with trailing comment
* Add tests
* Add comment
* Add test
* Add changelog
* Fix by lint
* Modify to move transformation to printBinaryishExpressions
* Remove unnecessary new line
* Fix hasTrailingCommentInBinaryish
* Add shouldReturnParts option to printComments
* Add tests from review
* Fix by lint
* Extract and use normalizeParts
* Group CallExpression when parent node is binaryish
* Fix by lint
* Remove unnecessary change
* Wrap the doc of most left node with group
* Add tests
* Update pr-8476.md
Co-authored-by: Georgii Dolzhykov <thorn.mailbox@gmail.com>
