Merge in upstream changes #1
Merged
Conversation
This file contains 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
…r latest dev version
…one for correct JavaScript.
[JavaScript] Remove Mustache templating from JavaScript syntax.
use the same scope for parameter separators in lambdas as for normal methods, for consistent highlighting, and fix bug with lambdas with multiple parameters and no curly braces
Add `border-box`, `text-before-edge`, `text-after-edge`, `ruby-base`, `ruby-text`, `run-in`, `ruby-base-container`, and `ruby-text-container` to CSS syntax for correct highlighting and add `user-select` to completions
…or more consistent color schemes
Numeric accessors used to be matched as constants. Now all invalid accessors (e.g. `abc.123` or `abc.+-`) are highlighted as such.
Define {{digitpart}} as described in the lexical analytics section of
the Python documentation.
Specifically: Don't allow trailing underscores.
Add support for "repeated N times" in array preamble and improve robustness by matching unexpected nested braces.
...by correcting `((example)+)*` style patterns to `((example)+)?`
[Markdown] improve backtracking on non-sregex engines like Oniguruma
[LaTeX] Improve support for array package
[Python] Do better at matching numbers
[XML] fix comments in doctype international subset
[C#] support object/collection initializers on a separate line
[HTML] support self-closing other/custom tags
[C#] correctly scope casts with namespaces
[CSS] Update missing values/properties
[php] Adds @api to phpdoc highlighting
[C#] make lambda parameter separator scope more specific
[XML] support xml-model preprocessor instructions
[C#] add support for composite string placeholder formatting
[C#] fix catch with generic type and no variable name
[R] major update
[CSS] support nested parens in calc function call
[JavaScript][HTML] Removed HTML comments from JavaScript.
…mpilation [JavaScript] Removed IE conditional compilation. Added test to cover old bug.
[Diff] Handle non-existant files
[CSS][Add] Scope for RGBA hexadecimal
Thom1729
pushed a commit
that referenced
this pull request
Nov 12, 2020
* [CSS] Update to sublime-syntax version 2
This commit contains the minimal required changes to update the syntax
definition to version 2 including proposed order of main contexts in the
header section and replacement of `pop: true` by `pop: 1`
According to existing syntax tests it seems there are no backward
compatibility issues.
* [CSS] Reorganize color values contexts
This commit...
1. creates a CONSTANTS section
2. moves color-values context into the new section
3. organizes color value patterns in variables
Note: May need further steps as `color-functions` included by
`color-values` feels not logical at this state.
* [CSS] Rename color constant scopes
This commit renames color-constant scopes to improve consistency.
Before this commit color constants are scoped inconsistently:
- `w3c-standard-color-name` => name
- `w3c-extended-color-keywords` => keywords
- `w3c-special-color-keyword` => keyword
A "keyword" in a "constant" scope is odd.
Mixed plural and singular ... .
All of those colors can now be addressed by `support.constant.color.w3c`
and further distinguished by [standard|extended|special] sub scopes.
* [CSS] Add operators and punctuation sections
Adding logical structures goes on.
* [CSS] Reorganize property values in variable
All pattern lists are organized in variables step by step to be independent from indentation and don't let such lists disturb
readability of normal pattern rules.
* [CSS] Move unicode-range context
Adding logical structures goes on.
It is a constant.
* [CSS] Reorganize property names in variable
All pattern lists are organized in variables step by step to be independent from indentation and don't let such lists disturb
readability of normal pattern rules.
* [CSS] Move and rename custom-property context
Adding logical structures goes on.
* [CSS] Refactor selectors section
This commit...
1. creates a SELECTORS section
2. moves related contexts into that section
3. organizes constant lists in variables
4. renames the `selector` context to plural to express non-popping
behavior
5. cleans unnecessary quotes from patterns
6. turns some anonymous into named contexts
* [CSS] Create builtin functions section
Just adds the header and fixes comment alignment. The contexts itself
stay unchanged at this point.
* [CSS] Reorganize property list and values
This commit creates two sections
1. PROERTY LIST
2. PROPERTY VALUES
to organize property related contexts.
* [CSS] Create at-rules section
* [CSS] Update comments
1. Create section
2. add `begin` and `end` sub-scopes
3. Include all `*` into punctuation for consistent highlighting.
* [CSS] Move selectors section
at-rules and selectors are siblings and therefore are organized next to
each other.
* [CSS] Cleanup at-rules contexts
1. Remove quotes
2. Remove `\s*` where appropriate
3. Remove dangling captures
4. Sort contexts
* [CSS] Add named at-rules-content contexts
* [CSS] Add missing comments includes
* [CSS] Refactor pseudo-class and pseudo-elements
This commit ...
1. interprets pseudo-classes and -elements as member-access:
div:nth-child(odd)
^ accessor
^^^^^^^^^^^^^^ member
a) replace `punctuation.definition.entity` by `punctuation.accessor`
b) accessors are no longer scoped `meta.function-call`
2. introduces meta.function-call.[identifier|arguments] scopes
3. optimizes function-call arguments context switches using `set` and
by avoiding lookaheads.
* [CSS] Refactor selector context part 2
This commit...
1. Improves matching incomplete `#id` and `.class` selectors in cases
when only `#` or `.` are present.
Note: Pushes a context for identifier matching onto stack, because of
a) better performance
b) help inherit syntax definitions to inject tags into the
selector such as `.<% print(element); %>` without loosing
meta scopes.
2. Sorts patterns logically
a) common stuff
b) html tags
c) attribute selectors
d) pseudo classes and elements
e) combining operators
3. Scope html tags as `entity.name.tag.html.css`
* [CSS] Rename qualified-name and vendor-prefix
Use plural to express non-popping behavior.
* [CSS] Refactor strings and attribute-selector-value
This commit ...
1. renames `literal-string` context to `quoted-strings`
2. adds a popping `quoted-string`
3. organizes common quoted-string content in named contexts.
4. uses `quoted-string` in `attribute-selector-value` context in order
to support leading whitespace in selector values.
* [CSS] Add illegal highlighting in attribute-selectors
This commit extends attribute-selector contexts in order to highlight
illegal patterns rather than failing with weird broken highlighting.
* [CSS] Rename attribute-selector's logical operator scope
This commit intents to introduce a more common operator scope, which
complies with scope naming guidelines.
* [CSS] Rename at-supports logical operator scopes
This commit renames the `.logic` scope to `.logical` for consistency
reasons and to make it comply with scope naming guidelines.
* [CSS] Rename combinators scope
This commit turns `combinators` into `keyword.operators` because:
a) they express some kind of logical relation to form a query
b) such operators are scoped `keyword.operator` in PackageDev's selector
syntax as well, which may be compared to CSS selectors.
c) it seems more logical as attribute-selectors use this scope, too.
* [CSS] Add column-combinator
Adds `||` operator
see: https://drafts.csswg.org/selectors-4/#the-column-combinator
* [CSS] Rename attribute-selector punctuation scopes
This commit scopes `[` and `]` as punctuation.section.attribute-selector
to comply with scope naming guidelines.
* [CSS] Refactor property identifiers
This commit...
1. converts several property identifier related contexts into named ones
2. makes use of push->set->set->pop strategy to organize contexts
3. simplifies some patterns
* [CSS] Refactor property values
This commit...
1. organizes all property-values (counter, font, other) in the property
values section
2. simplifies patterns in property value related contexts
3. fixes an issue which caused predefined font-names not to be
highlighted correctly, when followed by `!important`.
* [CSS] Rename counter name scope
This comment changes counter identifier scope to
`entity.other.counter-name.css` for consistency reasons as it is
already being used in counter functions.
* [CSS] Refactor builtin function contexts
This commit...
1. applies the same push->set->set-pop strategy for function arguments
after the scheme of pseudo-classes.
2. scopes function calls `meta.function-call.[identifier|arguments]`.
3. removes unnecessary quotes
4. renames all function contexts to plural for consistency reasons and
to express their non-popping behavior.
* [CSS] Update comma-delimiters context
This commit...
1. removes unnecessary whitespace patterns and capture groups from the
`comma-delimiters` context.
2. renames the `comma-delimiter` to `comma-delimiters` to express
non-popping behavior.
* [CSS] Tweak illegal-operator scope
* [CSS] Rename group punctuation scopes
This commit renames `punctuation.definition.group` to
`punctuation.section.group` to comply with scope naming guidelines.
* [CSS] Rename block punctuation scopes
This commit ...
1. adds `meta.block` to `meta.property-lists`
2. renames all braces scopes to `punctuation.section.block.[begin|end]`
The goal is to ensure consistent highlighting of all braces.
Note: Maybe replacing `meta.property-list` by `meta.mapping` would be
more appropriate here, but as that change would also require
changes to completions etc. it feels like out of scope at this
point.
* [CSS] Consistent comma scopes
* [CSS] Organize vendor-prefix patterns in a variable
* [CSS] Fix custom element chars pattern
According to HTML.sublime-syntax and whatwg.org custom elements may
contain `.` chars. The current implementation contains an `\\\.` which
makes no sense and was fixed in HTML.sublime-syntax before.
As fixing the pattern (copy from HTML.sublime-syntax) breaks selectors,
this commit simply removes the faulty escaped dot.
* [CSS] Rename line-names scopes
This commit...
1. removes unnecessary quotes from patterns
2. adds `meta.line-names.css meta.brackets.css` scope
3. renames punctuation to `punctuation.section.brackets`
* [CSS] Refactor custom-property contexts
This commit ...
1. splits custom-property definitions and references
2. scopes definitions as `entity.other.custom-property` as
`support.type` is not appropriate because these are user defined
variables.
3. scopes references as `variable.other.custom-property` as this is what
they are.
4. removes stacked scopes (except `invalid.deprecated`).
3. renames `keyword.other...prefix` to `entity.other...prefix` as the
whole word should be addressable by `entity.other.custom-property`.
The editor uses invalid.deprecated only anyway, so this shouldn't
change much.
5. Enable 'Goto Definition' and 'Goto Reference' for custom properties.
* [CSS] Remove dash from word_separators
Many selectors contain `-` thus it feels like removing it is logical.
* [CSS] Add meta.string.css scopes
This commit adds `meta.string.css` in order to support interpolation.
* [CSS] Rename unicode-range context
Use plural to express non-popping behavior.
* [CSS] Reorganize values/functions/constants
This commit ...
1. renames `-type` contexts to `-constants` as this is what those
contexts contain. Specs name it `type`, which is correct as those
constants are of certain kind, but it reads as those would contain
type definitions.
2. adds sections for string- and numeric- constants
3. Contexts which may contain functions and constants are named
`-values`.
4. All non-popping contexts use plural names.
5. Sort some context imports to achieve the same order everywhere.
This commit doesn't change the overall behavior of the syntax.
* [CSS] Sort constants sections alphabetically
* [CSS] Rename at-supports-parens to -groups
This commit...
1. renames at-supports-parens to match the meta.group scope name
2. swaps groups and operators includes as it is a habit to always have
groups and blocks first in a context.
* [CSS] Fix HTML and PHP syntax tests
* [CSS] Use meta.brackets in attribute-selectors
This commit...
1. stacks `meta.brackets` onto `meta.attribute.selector`
2. renames `punctuation.section` accordingly.
The goals:
1. Common scopes for brackets to ensure consistent highlighting.
2. Improve compliance with scope naming guidelines.
* [CSS] Rename qualified-names context to namespace-prefixes
This commit renames the context as it doesn't contain patterns for
qualified names, but only namespaces used to form the former ones.
* [CSS] Tweak attr function
This commit introduces a popping `var-function` context being used in
`attr-functions` arguments. It is to ensure to correctly pop the first
context after matching the first pattern.
* [CSS] Move operators/punctuation sections
This commit moves operators and punctuation section to the end of the
document and adds a prototypes section to follow the scheme of other
syntax definitions.
* [CSS] Remove quotes from variables
* [CSS] Add some bailouts to help with incomplete groups
* [CSS] Add missing meta.group in at-media context
* [CSS] Simplify at-keyframe-names context
Remove unnecessary context switches.
* [CSS] Simplify media-query context
Remove unnecessary context switches.
* [CSS] Tweak unicode-ranges scope
This commit follows the CSS3 package and scopes unicode-ranges
`constant.numeric`
* [CSS] Add named context for line-names-content
* [CSS] Move unicode-ranges to numeric constants section
* [CSS] Workaround for core issue sublimehq#3666
see: sublimehq/sublime_text#3666
* [CSS] Add more named contexts
This commit turns remaining anonymous at-rule contexts into named ones
to be prepared for inheritance.
* [CSS] Fix at-rule termination
This commit uses `at-rule-end` as prototype in all at-rule contexts and
makes sure to pop only if `;`, `{` or `}` are matched. The last one is
additionally matched to handle incomplete code blocks.
Source:
https://developer.mozilla.org/en-US/docs/Web/CSS/At-rule
At-rules are CSS statements that instructs CSS how to behave.
They begin with an at sign, '@' (U+0040 COMMERCIAL AT), followed by an
identifier and includes everything up to the next semicolon,
';' (U+003B SEMICOLON), or the next CSS block, whichever comes first.
* [CSS] Add at-namespace-identifier context
Another workaround for sublimehq/sublime_text#3666
* [CSS] Add at-page-identifier context
Another workaround for sublimehq/sublime_text#3666
* [CSS] Remove capture group from at-font-face context
Makes it comply with the other contexts.
* [CSS] Add at-rule keyword punctuation tests
This commit is to ensure keyword and punctuation.definition scopes of
all at-rule keywords to be stacked in correct order.
* [CSS] Tweak at-rule-end
As `;` is matched in main context the capturing rule-terminator context
is replaced by a lookahead.
Hence `;` no longer receives `meta.at-rule` scope as it does not receive
`meta.property-value` in property lists. This is to increase consistency
of meta scopes.
* [CSS] Add at-rule meta scope tests
This commit adds various meta scope related tests to ensure rule, group
and block boundaries are matched correctly and to clarify the difference
between at-block and rule-list usages.
* [CSS] Remove rule-terminator context
Globally match all `;` in property lists using `rule-terminators` to
ensure to highlight them all properly. As a result the consuming
`rule-terminator` context is obsolete.
Notes:
1. It is not intended to highlight stray semicolons as they no real
syntax errors and don't prevent proper evaluation.
2. This commit doesn't change the behavior as `meta.property-value`
is assigned via `meta_content_scope` directive and thus wasn't
assigned to the `;` before.
* [CSS] Remove obsolete eol-pop context
* [CSS] Add various at-rule spec links
* [CSS] Improve media-queries context
This commit applies parsing rules from
https://drafts.csswg.org/mediaqueries-5/#media
It adds...
1. range operators <, >, <=, >= and = as keyword.operator.comparison
2. allows nested groups according to
https://drafts.csswg.org/mediaqueries-5/#mq-syntax
* [CSS] Add MEDIA QUERIES section
Media queries are used in several at-rules and consist of various
contexts. Both justify a dedicated section for logical grouping.
* [CSS] Add at-rule other fallback context
Completions may want to see a `meta.at-rule` and we might not want to
fail on unknown/experimental/new at-rules even though special constants
can't be highlighted then.
This commit therefore adds a general `at-other` context which fulfills
the least syntactical rules to highlight at-rules according to
https://www.w3.org/TR/css-syntax-3/#at-rule
Note:
Any `@<ident> ... [ {} | ; ]` is valid syntax even though the rule might
not (yet) be implemented. A syntax definition is to highlight proper
syntax. It should not be used for linting too much. This is what all the
various LSP plugins or SublimeLinter can do.
* [CSS] Fix #id tokens
This commit introduces a `name` variable and uses it instead of
`{{nmchar}}+` in various situations. This is to improve compliance with
CSS tokenization specification.
The `name` is defined at https://www.w3.org/TR/css3-selectors/#lex and
is matched in HASH (after `#`) and IDENT (after `--`).
The main goal is to fix `#id` selector highlighting if the id is of
numeric type.
* [CSS] Tweak nmstart and nmchar variables
This commit modifies the `nonascii` part of `nmstart` and `nmchar` to
make it match the specification
see: https://www.w3.org/TR/css3-selectors/#lex
The spec says:
"nonascii [^\0-\177]"
In short:
A name or identifier may consist of all unicode code points which are
not part of ASCII.
Related rules:
ident [-]?{nmstart}{nmchar}*
name {nmchar}+
nmstart [_a-z]|{nonascii}|{escape}
nonascii [^\0-\177]
unicode \\[0-9a-f]{1,6}(\r\n|[ \n\r\t\f])?
escape {unicode}|\\[^\n\r\f0-9a-f]
nmchar [_a-z0-9-]|{nonascii}|{escape}
* [CSS] Add custom-media-identifier
Adapted from CSS3 package as the token has existed in syntax tests
without scope.
* [CSS] Tweak at-charset pattern
This commit moves `scope:` to capture group `0:` as this is the way
patterns of all the other at-rules are implemented.
* [CSS] Enforce identifier break
With regards to {{nmchar}} variable \b is not a sufficient pattern to
match identifier or name boundaries.
This commit therefore replaces all trailing \b by a custom pattern.
As using (?!{{nmchar}}) doubles parsing time an optimized alternative is
implemented as {{break}}.
Note: To be able to remove leading \b we'd need to ensure to match all
tokens in all contexts, which look like names or identifiers.
* [CSS] Update known html tags
This commit updates the `html_tag_names` variable with categorized html
tags from https://developer.mozilla.org/en-US/docs/Web/HTML/Element
Note: Some of the existing tag names seem to be svg tags. They are maintained in a dedicated `svg_tags` variable to avoid breaking things.
* [CSS] Add at-page-margin contexts
This comment adds support for page margin at-rules.
see: https://www.w3.org/TR/css-page-3/#margin-at-rule
* [CSS] Add missing comments context includes
* [CSS] Fix attribute-selector wildcard attribute name
* [CSS] Add `disc` and css-wide keywords to invalid counter-style names
That's what the specs say about it.
* [CSS] Update counter-style-system property
This commit adds a dedicated context for @counter-style { system: ; }
in order to properly support `extends: <counter-style-name>`.
Note:
1. This may be useful for goto definition / goto reference in future.
2. goto definition/reference may need to scope the identifier as
variable. I kept `entity.other` unchanged for compatibility reasons
at this point. Scope names may need further discussions.
* [CSS] Reorganize counter-property in dedicated section
* [CSS] Reorganize font-property in dedicated section
* [CSS] Add counter-style-speak-as property
* [CSS] Reorganize font-name-constants in a variable
* [CSS] Consolidate counter-style test cases
Move and extends counter-style attribute tests to the existing block
which already contains a bunch of those and remove the obsolete blacknwhite counter-style rule.
* [CSS] Add var() highlighting to speak-as values
* [CSS] Reorganize builtin property values
* [CSS] Handle illegal blocks in property lists
Highlight illegal braces according to
https://www.w3.org/TR/CSS22/syndata.html#parsing-errors
* [CSS] Handle unknown at-rules
Unknown at-rule keywords are to be ignored, while maintaining bracket
and terminator balancing;
see: https://www.w3.org/TR/CSS22/syndata.html#parsing-errors
At-rules with unknown at-keywords. User agents must ignore an invalid
at-keyword together with everything following it, up to the end of the
block that contains the invalid at-keyword, or up to and including the
next semicolon (;), or up to and including the next block ({...}),
whichever comes first.
* [CSS] Reusable block-end context
* [CSS] Update media queries
This commit...
1. adds support for `var()` and `calc()` functions
2. creates a dedicated `ration-constants` context
3. highlights vendor prefixes the same way as in other places without
stacking `support.type.property-name`
4. adds missing level 4 and 5 property names
5. adds missing level 4 and 5 property values and some comments about
the properties they belong to
* [CSS] Compact unicode-range pattern
* [CSS] Consistent case insensitive patterns
This commit turns all into case insensitive patterns. Some patterns
already were, others not. According to CSS specs all tokens are to be
matched case insensitive. May not be the best style but doesn't create
parse errors in browsers.
Note: This causes lexing times to increase, so we are where we started.
The original test file requires 16ms (was 17ms).
* [CSS] Fix counter-style name highlighting
Only in context of a counter-style prelude several constants and
keywords are considered illegal. This is fixed here.
This commit also
1. removes `counter_styles` from common `property-values` as those
tokens are already matched in all proper contexts.
2. adds `none` to the list of predefined `counter_styles` to simplify
word boundary handling.
* [CSS] Scope at-rule keywords as keyword.control.directive
From reading specifications at-rules some of them may be compared to
preprocessor directives. Some are used to import rules, others are of
conditional meaning which cause certain selectors/rules to be applied
only if certain criteria are matched.
Hence this commit applies `keyword.control.directive` for all at-rule
related keywords. This scope is
a) used in Erlang
b) is the most recent agreed scope for such things
c) shouldn't impact highlighting too much as `keyword.control` is
maintained.
The keyword's name is removed from scope as it doesn't seem to make much
sense and it can be addressed by the meta scope if needed.
e.c.: meta.at-rule.charset keyword.control.directive
* [CSS] Common scope for all wildcards
Notes:
- Another `entity` might not be useful.
- `keyword.operator` is used by other syntaxes but might lead to
confusion with logical operators easily.
- An asterisk might be a `constant.other.placeholder`
- Decided for `variable.language` as we also use that for `_` (python)
to denote such special variables.
* [CSS] Fix custom element tags patterns
Custom elements consist of lowercase characters only according to the
specs. So revert those case-insensitive changes for them.
* [CSS] Add some reserved custom element tags
According to the specs several tags looking like custom-elements are
reserved for use by SVG and MathML. This commit adds those to the list
of svg tags.
* [CSS] Highlight counter-style-names in fallback property
May be used to drive goto reference/goto definition in future.
* [CSS] Consolidate counter/style idents
This commit...
1. reuses `counter-identifier` and `counter-style-identifier` contexts
in properties and function arguments to ensure consistent scope
names.
2. adds patterns to highlight predefined counter-styles in appropriate
contexts.
3. moves counter related functions into a separate context.
4. adds some tests to ensure incomplete counter function statements to
behave correctly.
* [CSS] Fix misc counter-style related tests
* [CSS] Rename other-property to builtin-property
`other` sounds like unspecified or user defined, but this context
actually matches builtin property names.
* [CSS] Add illegal-groups context
Fixes sublimehq#1100
This commit adds an `illegal-groups` context in order to keep bracket
balancing intact when detecting an unexpected parenthesis.
It follows the `illegal-blocks` context to fulfill the behavior
described at https://www.w3.org/TR/CSS22/syndata.html#parsing-errors
Parentheses are unexpected ...
1. outside of blocks / property-lists
2. within selectors
3. within other groups after `var()` was matched.
Note:
`var()` functions are to be supported in all property-value like
situations. Hence the context is already included in most/all
property value and function call argument contexts. Including
`illegal-groups` in `var-functions` therefore enables correct
matching to hopefully all appropriate contexts without further
efforts.
It is expected to match valid parentheses before `var-functions`
contexts are included.
* [CSS] Move var-functions contexts
Sort builtin-functions contexts alphabetically.
* [CSS] Move calc-functions contexts
Sort builtin-functions contexts alphabetically.
* [CSS] Move attr-functions contexts
Sort builtin-functions contexts alphabetically.
* [CSS] Move color-functions contexts
Sort builtin-functions contexts logical and alphabetically.
Seems to belong next to filters.
* [CSS] Move crossfade-functions contexts
Sort builtin-functions contexts logical and alphabetically.
* [CSS] Create context for repeat()
* [CSS] Rename minmax-functions
Rename it to make clear it belongs to grids.
* [CSS] Add tests for uppercase tags and properties
Fixes sublimehq#1017
Ensures fixes for the mentioned issue.
* [CSS] Add missing test rules
* [CSS] Remove special scope for media properties
This commit removes `.media` sub scope from `support.type.property-name`
within @media rules. It is the only use case of such a special case and
adding those for all kinds of properties to provide consistency is too
much of effort.
* [CSS] Highlight utf8 percent characters in urls
This commit follows commit e12b971
and adds ridge url highlighting by scoping utf-8 percent characters in
url strings.
specs: https://url.spec.whatwg.org/#userinfo-percent-encode-set
Notes:
1. The intention is to provide consistent url highlighting in HTML/CSS.
2. The implementation follows the one in HTML and adds `meta.path.url`
to the string content. It is injected before `meta.string` for the
same reasons: Ensure `meta.path` is maintained even if string scopes
are cleared for interpolation reasons by 3rd party syntaxes.
* [CSS] Add common function arguments context
This commit adds `function-arguments-common` context which takes common
arguments patterns and cleans up `group-end` to only contain the
primary `\)` pattern such as `block-end2`. This is to enable special
handling of secondary bailouts, if needed.
* [CSS] Tweak media query group bailout
Media queries are separated by `,` but groups may not contain them.
Hence `,` is added to the list of bailout patterns to maintain
highlighting in case of missing `)`.
* [CSS] Remove special scope for media operators
This commit follows the removal of special scopes for property names and
does so for logical operators in media queries as well.
* [CSS] Add import test to verify media queries
* [CSS] Keyframe names may be quoted strings
* [CSS] Limit supported keyframe seletors
This commit replaces `main` by only supported contexts in keyframe
declaration block.
* [CSS] Fix from,to keyword scopes
1st and 2nd level should only contain agreed scopes according to naming
guidelines.
* [CSS] Scope keyframe selectors meta.selector.css
* [CSS] Reorganize media-queries contexts
This commit...
1. renames the main context to `media-queries` as the `-content` suffix
indicates a required opening pattern, which does not exist.
2. renames `media-query-types` to `media-query-media-types`.
4. adds named contexts for property names and values to help inherited
syntaxes to override or extend them.
* [CSS] Rescope pseudo class argument constants
Those are normal predefined constants. No special keywords with certain
functionality. We might want to express that by using common scopes.
* [CSS] Scope dir() arguments as constants
Only `ltr` and `rtl` should be matched within :dir() pseudo classes.
* [CSS] Highlight invalid keyframes identifiers
Globally reserved words may not be used as custom-identifier.
See: https://www.w3.org/TR/css-values-4/#custom-idents
* [CSS] Use common invalid identifier scope everywhere
* [CSS] Add named at-counter-style-names context
May help inheritance.
* [CSS] Add named at-page-names context
May help inheritance.
* [CSS] Exclude prototypes in function calls
This commit prevents inherited syntax definitions from injecting
prototypes between function identifiers and arguments groups.
* [CSS] Add :nth-col and :nth-last-col pseudo classes
* [CSS] Reorganize pseudo-class contexts
Splits them into several named contexts per pattern to help extending
them by inherited syntaxes.
* [CSS] Reorganize common selector contexts
Don't fit into pseudo-class section, so moves them out.
* [CSS] Add dedicated lang-pseudo-class context
This commit handles `:lang()` pseudo-class special, as its arguments
may contain a comma-separated list of language-range identifiers, which
may be quoted or unquoted strings. The latter ones may contain wildcards.
see: https://drafts.csswg.org/selectors-4/#the-lang-pseudo
* [CSS] Add support for functional pseudo-elements
* [CSS] Group rational constants under scalars
A rational value (e.g. 21/9) without unit is a scalar value same as a
scalar float but with a different representation.
This commit...
1. renames and moves the context accordingly
2. scopes rational numbers as such. We know it from Ruby.
* [CSS] Tweak legacy pseudo-elements handling
This commit...
1. updates some comments about pseudo classes and elements.
2. leaves `pseudo-element-css2` context include after `:` only as those
patterns are already matched by `pseudo-element-regular` after `::`.
It doesn't change parsing behavior/result.
* [CSS] Remove :drop() pseudo class
This commit removes the special context for `:drop()` pseudo class, as
1. it can't be found at https://drafts.csswg.org
2. it is not listed at caniuse.com
3. the only reference found at stackoverflow seems dead
4. the only thing lost by removing it is special highlighting of
argument constants like `active`.
Seems to be a feature which never matured beyond draft state.
* [CSS] Create one token for :// in urls
Propably helps ligatures and reduces amount of tokens from 3 to 1.
* [CSS] Tweak unquoted url-token patterns
This commit implements unquoted url token parsing according to specs at
csswg.org
Matching single or double quotes or opening parentheses causes a
bad-url-token to be returned. It would normally mean to highlight
everything as illegal, but I don't want to start branching at this
point.
* [CSS] Stricter functional pseudo-class patterns
Only add meta.function-call scope if a pseudo-class looks like that.
* [CSS] Move language-ranges into the correct section
It's an unquoted string.
* [CSS] Enhance cross-fade(), image() and image-set() functions
This commit implements those functions according to the specification.
Thay are still highly experimental and rarely supported by browsers, but
have existed in this syntax before.
The `string-unquoted` is removed as those functions don't support them.
The `quoted-strings` context is replaced by `quoted-urls` as all quoted
strings contain urls in those functions.
* [CSS] Reorganize page-margin at-rule names in a variable.
May help inherited syntaxes
* [CSS] Revert punctuation.accessor to punctuation.definition
CSS denotes the different kinds of selectors by the opening punctuation.
The concept of accessor would either need to be applied to all those
elements or to nothing. With the usage of stand-alone pseudo classes as
identifier of a page at-rule, it feels more appropriate to keep going
with `punctuation.definition` in all situations.
* [CSS] Update generic-font-families variable
Adds more level 4 constants.
* [CSS] Remove scalars from keyframe selectors
We prefer to use the syntax for validation purposes, instead of
maintaining highlighting of incomplete expressions.
* [CSS] Consolidate font property tests
* [CSS] Handle vendor-prefixes in pseudo-classes/elements
This commit ensures to handle vendor prefixed pseudo classes and
elements the same way as other properties and constants and to scope
those in any kind of token.
* [CSS] Fix property-name word boundaries
* [CSS] Consolidate identifiers
Any custom CSS identifier may look like `--1dent`, `-name1` or `name2`.
In some situations only the first one is appropriate, in others we might
only match the latter two as the first one is handled separately. See
custom-property for instance.
This commit splits `ident` variable into those two major groups to be
used where needed. `punctuation.definition` is removed from custom
properties for consistency reasons. To keep it there would mean to add
it to any other custom property, which includes quite a number of
patterns to customize, which felt not worth it at this point.
The class #id must consist of a valid custom identifier so `#1` is not
valid.
* [CSS] Start grouping builtin property value constants
* [CSS] Fix meda-queries property values
The `hover` property-value was highlighted as property-name before.
* [CSS] Fix property-value word boundaries
Consume unknown identifiers to prevent tokens to be partly highlighted
as something known in the middle of a word.
* [CSS] Add justify-items and justify-self property
* [CSS] Rename at-block to at-rule-block
The name `at-block` looks like a context name of an `@block`.
To avoid that this commit follows `at-rule-end` and renames `at-block`
to `at-rule-block`.
* [CSS] Remove obsolete bailout
Possible due to 93bffe9
* [CSS] Fix local() function arguments
This commit adds support to highlight all kinds of font-family names in
`local()` arguments groups.
* [CSS] Refactor var() fallback values
This commit makes use of multi-set statement to...
1. match the first identifier as custom property name reference, only.
(This is to comply with other use-cases this strategy is already
applied to.)
2. add support for unquoted strings in default values.
The `font-property-value-content` context is used which includes the
`font-family-names` context. This causes a space separated list of
identifiers to be scoped `string.unquoted`.
* [CSS] Revert grid-...-functions context names
The `grid-` prefix doesn't match the naming scheme somehow.
* [CSS] Second approach to reorganize builtin functions
This commit organizes builtin function contexts in sections to group
them by "library" or logical/functional context.
It doesn't change a pattern.
* [CSS] Add other-functions context
This commit attempts to avoid breaking syntax highlighting due to
unknown functions being called. Using `variable.function` should enable
color schemes to distinguish them from valid builtins.
* [CSS] Scope unknown constants as constant.other
This commit attempts to avoid breaking syntax highlighting due to
unknown constants being used. Using `constant.other` should enable
color schemes to distinguish them from valid builtins.
* [CSS] Offer property completions in property-lists only
This commit limits property name completions to property lists and
incomplete property names.
1. `meta.property-list`
Immediately suggest properties via ctrl+space in { } blocks.
2. `meta.property-name`
Suggest in inline html style tags or whenever the syntax started
detecting a property name.
We might not want properties suggested in at-rules or outside of blocks.
* [CSS] Don't append semicolon in groups
If property value completions are provided within groups, a trailing
semicolon is to be omitted as it breaks syntax.
Examples:
@media screen and (min-with: |) {}
must not result in
@media screen and (min-with: 10px;) {}
but
@media screen and (min-with: 10px) {}
The caret is represented by `|`
* [CSS] Organize grid constants in named context
Avoids duplicate definitions.
* [CSS] Organize direction constants in named context
Avoids duplicate definitions.
* [CSS] Add some spec references
* [CSS] Consolidate keywords under keyword.other
Replaces `support.keyword` and names scopes according to usage.
* [CSS] Replace remaining unquoted string
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
No description provided.