Js #11
Open
Js #11
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Author
|
Go |
ghost
approved these changes
Jan 5, 2023
Speedytats
approved these changes
Mar 8, 2023
Speedytats
approved these changes
Mar 8, 2023
harshsingh32
approved these changes
Apr 18, 2023
cesarapascual
approved these changes
Apr 19, 2023
* tools/preview_new_release.sh: Simplify and style changes - Make the `-p` not optional. The default value was harmful and I was not happy with uncertain values being passed to `rm -rf`. - Take URL prefixes as `-u` and `-y` instead of user names. The constructed URL doesn't work in every setup and is better provided directly. - Build ocamlformat. - Push the result to the fork repository. - Arguments are quoted when needed and other style changes. More reproducible than depending on installing it first in Opam. * Prune some mirage projects to reduce noise These projects have similar codebase and are maintained by a small number of people. * test-extra: Add ppxlib Ppxlib has unusual code and is a good target for the tests. * test-extra: Prune some projects Remove owl, which was not tested, sub-libraries of core, which all have similar code and index, which has similar code to irmin. * tools/projects.data: Remove flow It uses an old version of OCamlformat, the diff is too big to be useful. It's also not upgradable automatically due to its Dune config. * tools/projects.data: Remove owl The project hasn't been upgraded since the removal of `align-cases` and cannot be upgraded.
Printast fails to build since 5612a0f
These diffs were used to monitor the changes made to the parser but are getting in the way of making more changes. The patch files were out of sync.
* ocp-indent-compat: Don't unindent unwrapped docstrings In ocp-indent-compat mode, the indentation of unwrapped docstrings is based on the indentation of the first line. This applies to the janestreet profile.
* Preserve line break between operator and let `@@ let` used to always break while `@@ let+` used to always align. This was changed in 913f754 to always break in both cases for consistency. Now, the break is preserved except in the case of a plain `let`. Preserving the line break on `let` do not reduce the regressions with the previous version and unecessarily give more use cases to this unwanted feature. This applies to `let<op>`, `let exception` and `let module`.
- Preserve the formatting of the version option, this is required by Dune. - Update the .git-blame-ignore-revs file. This seems welcome, except in Dune where it must be disabled. - Add tarides/ocaml-platform-installer to the list - The list is sorted.
A let-binding can have a newtype and a coercion at the same time. This is fixed by removing the coupling between the two.
* Restore indentation of fun arrow
This is the formatting of 0.25.1.
* Improve indent of fun args preceded by a label
The arguments must be indented more to avoid them aligning with the
`fun` and `->` in case of a break:
let () =
very_long_function_name
~very_long_argument_label:(* foo *)
(fun
very_long_argument_name_one
very_long_argument_name_two
very_long_argument_name_three
-> () )
* Editorialize the changelog - Non-bug entries are moved to the "Changes" section. - Lists are sorted by PR number. - Similar entries are merged. - Spelling is improved * Bump version numbers * Mark breaking changes in the changelog with `*`
* Update the release procedure Missing mention of ocamlformat-lib. * Move the comment in ocamlformat.opam.template At the previous position, dune-release would generate more fields just before it, moving it to an unrelated place. * CHANGES: Standalone '*' that renders in HTML The previous idea was not working when the markdown was rendered by Github as every items were rendered the same. The syntax is a bit heavier in plain text but making these changes more visible doesn't hurt. * CI: Fix Windows release build - Use setup-ocaml instead of installing Opam from scratch. Removes calls to the cygwin installer. - Move scripts into the action file. Allows building for past releases. The script is now small enough to embed.
instead of the extended parser (the default).
This was a missing AST rule.
The special formatting of a begin-end in a match case was dropping attributes.
* Add a Roadmap * Add authors and maintainers information * Improve Contributing Guide and README * Convert Hacking guide to Markdown * Rename LICENSE * Update Changelog tags * Add description of the project in opam file * Remove outdated bash completion for test_branch Co-authored-by: Guillaume Petiot <guillaume@tarides.com> Co-authored-by: Jules Aguillon <jules@j3s.fr>
* Test wrapping class arrow type Test both break-separators to before and after. * Separate `fmt_arrow_type` The plan is to re-use it to format class types arrows. Some cleanup. * Consistent formatting for arrow class types Use the same indentation and breaks for arrows in class types as for arrows in core types. The main challenge is that class types contain class signatures, which are docked after an arrow. The `fmt_class_type` is rewritten in the "pro" style, which also remove unnecessary space in object poly types
* vendored parsers: Backport dead-code as comments Dead code in parser-extended have been removed in the past. I reverted that and commented it out instead. This makes the code more similar with parser-standard and helps recognize new code from previously removed code. * Backport 5.1 error printing code This slightly change some error messages.
* Backport 5.1.0-beta1 standard parser changes - New `Pmod_apply_unit` module type constructor. - New `value_constraint` field in value bindings. This makes the parser less permissive about bugs in the formatting of type annotation on let-bindings. - Some changes in `Location` and `Ast_mapper` are not necessary but are backported to reduce the diff. - String assignment operator has been removed upstream but remains supported by commenting out the change. * Backport Pmod_apply_unit to extended AST The extended AST already had a similar constructor. This is just a rename. * Backport value_binding change to parser-extended We already had a similar AST node but the `value_constraint` field is new and helps removing complex code in Sugar. The `pvb_is_pun` field is kept from the previous extended representation. * parser-extended: Rename 'let_bindings' to 'value_bindings' For consistency with the new 'value_binding'.
* De-indent the `object` keyword in class types The previous indentation of the keyword was ambiguous as it was aligned with the previous and next lines. This is similar to the formatting of module `struct`. * Ensure comments, doc, attrs and parens are formatted for class types
A comment just before a constructor in a type declaration was causing the argument of the constructor to be formatted in vertical style. The constructor is now formatted the same as if there were no comment.
* test_branch: Print tested versions When looking at the diffs on Github, it can be reassuring to see that the versions being tested are right. * Don't checkout the merge commit
* fix a crash with type%e nonrec * add attributes test * changelog * fmt
* showcase issue * solve issue
This doesn't make OCamlformat able to format 5.4 code.
* Add module-indent argument * Use module-indent option to indent structs and sigs * Add tests for module-indent option * Update documentation for module-indent option
* test_branch: Use `--no-comment-check` A test case added a misplaced doc-string, causing a fatal warning and failing `test_branch` in CI. This passes `--no-comment-check` to ignore this error.
* Basic OCaml 5.4 support - update vendored parsers to mirror upstream at 5.4: * introduce locations for Longident.t components * distinguish (module M:S) and ((module M):(module S)) for expressions - support for new syntaxes: * bivariance * labelled tuples * Add test for labeled tuples * review: lax rule for variance annotation * review: test comments inside tuple types * review: track labels locations in tuple types * Locations for labels * Concrete node for tuple element punning
* parser-standard: Reduce diffs with upstream Promote small differences that do not change the parser's behavior to parser-standard. This will make future updates easier. * parser-extended: Reduce diffs with parser-standard This slightly reduce the difference between the two vendored parser in the hope of making the future update easier.
* Bin_conf: don't pass info to update_using_cmdline There is only a single instance anyway. * Bin_conf: parse cmdline only once and cache result Otherwise the config parser would run for every argument, resulting in quadratic parsing.
Set 'TERM=dumb' to workaround unstable output from Cmdliner
Cmdliner output makes the tests unstable and cf3d2e7 did not fix the issue. For example, the lower-bounds test in ocaml-ci fails.
* Fix precedence of tuples with labels
It was formatting:
let labeled_tuple () = ~x:1, ~y:(1 + 2)
into:
let labeled_tuple () = ~x:1, ~y:1 + 2
* Update CHANGES
* Fix dropped comment in 'if then begin .. end' * Update CHANGES
* Add configuration to control whether let-punning is used * Add tests for let punning configuration * Update manpage in documentation * CHANGES entry * Rename to letop-punning * More documentation with example * Fix CHANGES
* Fix redundancy in letop-punning tests * Notes on adding a test in CONTRIBUTING * CHANGES entry * Edits per @EmileTrotignon * Typo repair * Edits per @Julow
* letop-punning for extension nodes * Update CHANGES.md Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> * Fix comments getting dropped by letop-punning=always --------- Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Currently, Cmdliner is given a formatter that drops any text, but Cmdliner can (and does) print whitespace via the non-overridden out_spaces, out_indent and out_newlines.
These locations, which are new in 5.4, are sometimes traversed (e.g. Pexp_setfield) and
sometimes not (e.g. Pexp_field). So clearly a bug.
The fix is concretely:
- change map_loc to map over both fields `{ loc; txt }` of `Location.loc`
- propagate the change, ie change `map_loc` to `map_string_loc`,
`map_string_opt_loc` or `map_loc_lid` depending on the type. All occurrences
of the latter are bug fixes.
* Support the new functor type syntax `MT -> MT`
Support the unnamed functor parameters in module types:
module type F = ARG -> S
The extended parser is changed to preserve the concrete syntax of
functor arguments in module types. Notably, these three lines are no
longer equivalent and the first two are no longer turned into the third:
module M : (_ : S) -> (_ : S) -> S = N
module M : S -> S -> S = N
module M : (_ : S) (_ : S) -> S = N
Update to the new Odoc features: - @toc_status and @order_category - Structured code block tags syntax - Indentation of code and verbatim blocks Also backported formatting changes and code changes. * odoc-parser: Code block and verbatim blocks indentation change Odoc now strips the indentation from code and verbatim blocks. It doesn't use the indentation of the least indented line, like OCamlformat, but instead use the indentation of the block opening. * odoc: Normalize `\\n` into newline In code blocks, newlines can be turned into `\\n` (or the opposite) when formatting string literals that need to break. Update the normalization function to avoid crashing when that happens. * odoc: Remove extra indentation in code blocks The updated Odoc parser considers that code blocks horizontally start at the opening bracket. It no longer uses the indentation of the least indented line for that. As a result, OCamlformat cannot indent code blocks without changing their content. The indentation is also removed in code blocks that are formatted, to avoid adding visible indentation in rendered documentation. * odoc: Don't fail for removed whitespaces in code blocks During normalisation, allow whitespaces to disappear from code blocks. This happens when code if formatted.
* Missing parens in module type `(M with type ..) -> N` Fix the parenthesing rules for `with type` inside functors and functors inside `with type`. * Fix short-syntax functor formatting
* use special case beginend syntax in match cases when possible * also allow `if then begin match` * changelog * changelog * fmt
Co-authored-by: Jules Aguillon <jules@j3s.fr>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
10 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Me gustaría probar