oasis-root

lexer.lua (92478B)

-- Copyright 2006-2024 Mitchell. See LICENSE.
--- Lexes Scintilla documents and source code with Lua and LPeg.
--
-- ### Writing Lua Lexers
--
-- Lexers recognize and tag elements of source code for syntax highlighting. Scintilla (the
-- editing component behind [Textadept][] and [SciTE][]) traditionally uses static, compiled C++
-- lexers which are notoriously difficult to create and/or extend. On the other hand, Lua makes
-- it easy to to rapidly create new lexers, extend existing ones, and embed lexers within one
-- another. Lua lexers tend to be more readable than C++ lexers too.
--
-- While lexers can be written in plain Lua, Scintillua prefers using Parsing Expression
-- Grammars, or PEGs, composed with the Lua [LPeg library][]. As a result, this document is
-- devoted to writing LPeg lexers. The following table comes from the LPeg documentation and
-- summarizes all you need to know about constructing basic LPeg patterns. This module provides
-- convenience functions for creating and working with other more advanced patterns and concepts.
--
-- Operator | Description
-- -|-
-- `lpeg.P(string)` | Matches `string` literally.
-- `lpeg.P(`_`n`_`)` | Matches exactly _`n`_ number of characters.
-- `lpeg.S(string)` | Matches any character in set `string`.
-- `lpeg.R("`_`xy`_`")`| Matches any character between range `x` and `y`.
-- `patt^`_`n`_ | Matches at least _`n`_ repetitions of `patt`.
-- `patt^-`_`n`_ | Matches at most _`n`_ repetitions of `patt`.
-- `patt1 * patt2` | Matches `patt1` followed by `patt2`.
-- `patt1 + patt2` | Matches `patt1` or `patt2` (ordered choice).
-- `patt1 - patt2` | Matches `patt1` if `patt2` does not also match.
-- `-patt` | Matches if `patt` does not match, consuming no input.
-- `#patt` | Matches `patt` but consumes no input.
--
-- The first part of this document deals with rapidly constructing a simple lexer. The next part
-- deals with more advanced techniques, such as embedding lexers within one another. Following
-- that is a discussion about code folding, or being able to tell Scintilla which code blocks
-- are "foldable" (temporarily hideable from view). After that are instructions on how to use
-- Lua lexers with the aforementioned Textadept and SciTE editors. Finally there are comments
-- on lexer performance and limitations.
--
-- [LPeg library]: http://www.inf.puc-rio.br/~roberto/lpeg/lpeg.html
-- [Textadept]: https://orbitalquark.github.io/textadept
-- [SciTE]: https://scintilla.org/SciTE.html
--
-- ### Lexer Basics
--
-- The *lexers/* directory contains all of Scintillua's Lua lexers, including any new ones you
-- write. Before attempting to write one from scratch though, first determine if your programming
-- language is similar to any of the 100+ languages supported. If so, you may be able to copy
-- and modify, or inherit from that lexer, saving some time and effort. The filename of your
-- lexer should be the name of your programming language in lower case followed by a *.lua*
-- extension. For example, a new Lua lexer has the name *lua.lua*.
--
-- Note: Try to refrain from using one-character language names like "c", "d", or "r". For
-- example, Scintillua uses "ansi_c", "dmd", and "rstats", respectively.
--
-- #### New Lexer Template
--
-- There is a *lexers/template.txt* file that contains a simple template for a new lexer. Feel
-- free to use it, replacing the '?' with the name of your lexer. Consider this snippet from
-- the template:
--
--     -- ? LPeg lexer.
--
--     local lexer = lexer
--     local P, S = lpeg.P, lpeg.S
--
--     local lex = lexer.new(...)
--
--     [... lexer rules ...]
--
--     -- Identifier.
--     local identifier = lex:tag(lexer.IDENTIFIER, lexer.word)
--     lex:add_rule('identifier', identifier)
--
--     [... more lexer rules ...]
--
--     return lex
--
-- The first line of code is a Lua convention to store a global variable into a local variable
-- for quick access. The second line simply defines often used convenience variables. The third
-- and last lines [define](#lexer.new) and return the lexer object Scintillua uses; they are
-- very important and must be part of every lexer. Note the `...` passed to `lexer.new()` is
-- literal: the lexer will assume the name of its filename or an alternative name specified
-- by `lexer.load()` in embedded lexer applications. The fourth line uses something called a
-- "tag", an essential component of lexers. You will learn about tags shortly. The fifth line
-- defines a lexer grammar rule, which you will learn about later. (Be aware that it is common
-- practice to combine these two lines for short rules.)  Note, however, the `local` prefix in
-- front of variables, which is needed so-as not to affect Lua's global environment. All in all,
-- this is a minimal, working lexer that you can build on.
--
-- #### Tags
--
-- Take a moment to think about your programming language's structure. What kind of key elements
-- does it have? Most languages have elements like keywords, strings, and comments. The
-- lexer's job is to break down source code into these elements and "tag" them for syntax
-- highlighting. Therefore, tags are an essential component of lexers. It is up to you how
-- specific your lexer is when it comes to tagging elements. Perhaps only distinguishing between
-- keywords and identifiers is necessary, or maybe recognizing constants and built-in functions,
-- methods, or libraries is desirable. The Lua lexer, for example, tags the following elements:
-- keywords, functions, constants, identifiers, strings, comments, numbers, labels, attributes,
-- and operators. Even though functions and constants are subsets of identifiers, Lua programmers
-- find it helpful for the lexer to distinguish between them all. It is perfectly acceptable
-- to just recognize keywords and identifiers.
--
-- In a lexer, LPeg patterns that match particular sequences of characters are tagged with a
-- tag name using the the `lexer.tag()` function. Let us examine the "identifier" tag used in
-- the template shown earlier:
--
--     local identifier = lex:tag(lexer.IDENTIFIER, lexer.word)
--
-- At first glance, the first argument does not appear to be a string name and the second
-- argument does not appear to be an LPeg pattern. Perhaps you expected something like:
--
--     lex:tag('identifier', (lpeg.R('AZ', 'az')  + '_') * (lpeg.R('AZ', 'az', '09') + '_')^0)
--
-- The `lexer` module actually provides a convenient list of common tag names and common LPeg
-- patterns for you to use. Tag names for programming languages include (but are not limited
-- to) `lexer.DEFAULT`, `lexer.COMMENT`, `lexer.STRING`, `lexer.NUMBER`, `lexer.KEYWORD`,
-- `lexer.IDENTIFIER`, `lexer.OPERATOR`, `lexer.ERROR`, `lexer.PREPROCESSOR`, `lexer.CONSTANT`,
-- `lexer.CONSTANT_BUILTIN`, `lexer.VARIABLE`, `lexer.VARIABLE_BUILTIN`, `lexer.FUNCTION`,
-- `lexer.FUNCTION_BUILTIN`, `lexer.FUNCTION_METHOD`, `lexer.CLASS`, `lexer.TYPE`, `lexer.LABEL`,
-- `lexer.REGEX`, `lexer.EMBEDDED`, and `lexer.ANNOTATION`. Tag names for markup languages include
-- (but are not limited to) `lexer.TAG`, `lexer.ATTRIBUTE`, `lexer.HEADING`, `lexer.BOLD`,
-- `lexer.ITALIC`, `lexer.UNDERLINE`, `lexer.CODE`, `lexer.LINK`, `lexer.REFERENCE`, and
-- `lexer.LIST`. Patterns include `lexer.any`, `lexer.alpha`, `lexer.digit`, `lexer.alnum`,
-- `lexer.lower`, `lexer.upper`, `lexer.xdigit`, `lexer.graph`, `lexer.punct`, `lexer.space`,
-- `lexer.newline`, `lexer.nonnewline`, `lexer.dec_num`, `lexer.hex_num`, `lexer.oct_num`,
-- `lexer.bin_num`, `lexer.integer`, `lexer.float`, `lexer.number`, and `lexer.word`. You may
-- use your own tag names if none of the above fit your language, but an advantage to using
-- predefined tag names is that the language elements your lexer recognizes will inherit any
-- universal syntax highlighting color theme that your editor uses. You can also "subclass"
-- existing tag names by appending a '.*subclass*' string to them. For example, the HTML lexer
-- tags unknown tags as `lexer.TAG .. '.unknown'`. This gives editors the opportunity to style
-- those subclassed tags in a different way than normal tags, or fall back to styling them as
-- normal tags.
--
-- ##### Example Tags
--
-- So, how might you recognize and tag elements like keywords, comments, and strings?  Here are
-- some examples.
--
-- **Keywords**
--
-- Instead of matching _n_ keywords with _n_ `P('keyword_`_`n`_`')` ordered choices, use one
-- of of the following methods:
--
-- 1. Use the convenience function `lexer.word_match()` optionally coupled with
--   `lexer.set_word_list()`. It is much easier and more efficient to write word matches like:
--
--        local keyword = lex:tag(lexer.KEYWORD, lex:word_match(lexer.KEYWORD))
--        [...]
--        lex:set_word_list(lexer.KEYWORD, {
--          'keyword_1', 'keyword_2', ..., 'keyword_n'
--        })
--
--        local case_insensitive_word = lex:tag(lexer.KEYWORD, lex:word_match(lexer.KEYWORD, true))
--        [...]
--        lex:set_word_list(lexer.KEYWORD, {
--          'KEYWORD_1', 'keyword_2', ..., 'KEYword_n'
--        })
--
--        local hyphenated_keyword = lex:tag(lexer.KEYWORD, lex:word_match(lexer.KEYWORD))
--        [...]
--        lex:set_word_list(lexer.KEYWORD, {
--          'keyword-1', 'keyword-2', ..., 'keyword-n'
--        })
--
--    The benefit of using this method is that other lexers that inherit from, embed, or embed
--    themselves into your lexer can set, replace, or extend these word lists. For example,
--    the TypeScript lexer inherits from JavaScript, but extends JavaScript's keyword and type
--    lists with more options.
--
--    This method also allows applications that use your lexer to extend or replace your word
--    lists. For example, the Lua lexer includes keywords and functions for the latest version
--    of Lua (5.4 at the time of writing). However, editors using that lexer might want to use
--    keywords from Lua version 5.1, which is still quite popular.
--
--    Note that calling `lex:set_word_list()` is completely optional. Your lexer is allowed to
--    expect the editor using it to supply word lists. Scintilla-based editors can do so via
--    Scintilla's `ILexer5` interface.
--
-- 2. Use the lexer-agnostic form of `lexer.word_match()`:
--
--        local keyword = lex:tag(lexer.KEYWORD, lexer.word_match{
--          'keyword_1', 'keyword_2', ..., 'keyword_n'
--        })
--
--        local case_insensitive_keyword = lex:tag(lexer.KEYWORD, lexer.word_match({
--          'KEYWORD_1', 'keyword_2', ..., 'KEYword_n'
--        }, true))
--
--        local hyphened_keyword = lex:tag(lexer.KEYWORD, lexer.word_match{
--          'keyword-1', 'keyword-2', ..., 'keyword-n'
--        })
--
--    For short keyword lists, you can use a single string of words. For example:
--
--        local keyword = lex:tag(lexer.KEYWORD, lexer.word_match('key_1 key_2 ... key_n'))
--
--    You can use this method for static word lists that do not change, or where it does not
--    make sense to allow applications or other lexers to extend or replace a word list.
--
-- **Comments**
--
-- Line-style comments with a prefix character(s) are easy to express:
--
--     local shell_comment = lex:tag(lexer.COMMENT, lexer.to_eol('#'))
--     local c_line_comment = lex:tag(lexer.COMMENT, lexer.to_eol('//', true))
--
-- The comments above start with a '#' or "//" and go to the end of the line (EOL). The second
-- comment recognizes the next line also as a comment if the current line ends with a '\'
-- escape character.
--
-- C-style "block" comments with a start and end delimiter are also easy to express:
--
--     local c_comment = lex:tag(lexer.COMMENT, lexer.range('/*', '*/'))
--
-- This comment starts with a "/\*" sequence and contains anything up to and including an ending
-- "\*/" sequence. The ending "\*/" is optional so the lexer can recognize unfinished comments
-- as comments and highlight them properly.
--
-- **Strings**
--
-- Most programming languages allow escape sequences in strings such that a sequence like
-- "\\"" in a double-quoted string indicates that the '"' is not the end of the
-- string. `lexer.range()` handles escapes inherently.
--
--     local dq_str = lexer.range('"')
--     local sq_str = lexer.range("'")
--     local string = lex:tag(lexer.STRING, dq_str + sq_str)
--
-- In this case, the lexer treats '\' as an escape character in a string sequence.
--
-- **Numbers**
--
-- Most programming languages have the same format for integers and floats, so it might be as
-- simple as using a predefined LPeg pattern:
--
--     local number = lex:tag(lexer.NUMBER, lexer.number)
--
-- However, some languages allow postfix characters on integers.
--
--     local integer = P('-')^-1 * (lexer.dec_num * S('lL')^-1)
--     local number = lex:tag(lexer.NUMBER, lexer.float + lexer.hex_num + integer)
--
-- Other languages allow separaters within numbers for better readability.
--
--     local number = lex:tag(lexer.NUMBER, lexer.number_('_')) -- recognize 1_000_000
--
-- Your language may need other tweaks, but it is up to you how fine-grained you want your
-- highlighting to be. After all, you are not writing a compiler or interpreter!
--
-- #### Rules
--
-- Programming languages have grammars, which specify valid syntactic structure. For example,
-- comments usually cannot appear within a string, and valid identifiers (like variable names)
-- cannot be keywords. In Lua lexers, grammars consist of LPeg pattern rules, many of which
-- are tagged.  Recall from the lexer template the `lexer.add_rule()` call, which adds a rule
-- to the lexer's grammar:
--
--     lex:add_rule('identifier', identifier)
--
-- Each rule has an associated name, but rule names are completely arbitrary and serve only to
-- identify and distinguish between different rules. Rule order is important: if text does not
-- match the first rule added to the grammar, the lexer tries to match the second rule added, and
-- so on. Right now this lexer simply matches identifiers under a rule named "identifier".
--
-- To illustrate the importance of rule order, here is an example of a simplified Lua lexer:
--
--     lex:add_rule('keyword', lex:tag(lexer.KEYWORD, ...))
--     lex:add_rule('identifier', lex:tag(lexer.IDENTIFIER, ...))
--     lex:add_rule('string', lex:tag(lexer.STRING, ...))
--     lex:add_rule('comment', lex:tag(lexer.COMMENT, ...))
--     lex:add_rule('number', lex:tag(lexer.NUMBER, ...))
--     lex:add_rule('label', lex:tag(lexer.LABEL, ...))
--     lex:add_rule('operator', lex:tag(lexer.OPERATOR, ...))
--
-- Notice how identifiers come _after_ keywords. In Lua, as with most programming languages,
-- the characters allowed in keywords and identifiers are in the same set (alphanumerics plus
-- underscores). If the lexer added the "identifier" rule before the "keyword" rule, all keywords
-- would match identifiers and thus would be incorrectly tagged (and likewise incorrectly
-- highlighted) as identifiers instead of keywords. The same idea applies to function names,
-- constants, etc. that you may want to distinguish between: their rules should come before
-- identifiers.
--
-- So what about text that does not match any rules? For example in Lua, the '!' character is
-- meaningless outside a string or comment. Normally the lexer skips over such text. If instead
-- you want to highlight these "syntax errors", add a final rule:
--
--     lex:add_rule('keyword', keyword)
--     ...
--     lex:add_rule('error', lex:tag(lexer.ERROR, lexer.any))
--
-- This identifies and tags any character not matched by an existing rule as a `lexer.ERROR`.
--
-- Even though the rules defined in the examples above contain a single tagged pattern, rules may
-- consist of multiple tagged patterns. For example, the rule for an HTML tag could consist of a
-- tagged tag followed by an arbitrary number of tagged attributes, separated by whitespace. This
-- allows the lexer to produce all tags separately, but in a single, convenient rule. That rule
-- might look something like this:
--
--     local ws = lex:get_rule('whitespace') -- predefined rule for all lexers
--     lex:add_rule('tag', tag_start * (ws * attributes)^0 * tag_end^-1)
--
-- Note however that lexers with complex rules like these are more prone to lose track of their
-- state, especially if they span multiple lines.
--
-- #### Summary
--
-- Lexers primarily consist of tagged patterns and grammar rules. These patterns match language
-- elements like keywords, comments, and strings, and rules dictate the order in which patterns
-- are matched. At your disposal are a number of convenience patterns and functions for rapidly
-- creating a lexer. If you choose to use predefined tag names (or perhaps even subclassed
-- names) for your patterns, you do not have to update your editor's theme to specify how to
-- syntax-highlight those patterns. Your language's elements will inherit the default syntax
-- highlighting color theme your editor uses.
--
-- ### Advanced Techniques
--
-- #### Line Lexers
--
-- By default, lexers match the arbitrary chunks of text passed to them by Scintilla. These
-- chunks may be a full document, only the visible part of a document, or even just portions
-- of lines. Some lexers need to match whole lines. For example, a lexer for the output of a
-- file "diff" needs to know if the line started with a '+' or '-' and then style the entire
-- line accordingly. To indicate that your lexer matches by line, create the lexer with an
-- extra parameter:
--
--     local lex = lexer.new(..., {lex_by_line = true})
--
-- Now the input text for the lexer is a single line at a time. Keep in mind that line lexers
-- do not have the ability to look ahead to subsequent lines.
--
-- #### Embedded Lexers
--
-- Scintillua lexers embed within one another very easily, requiring minimal effort. In the
-- following sections, the lexer being embedded is called the "child" lexer and the lexer a child
-- is being embedded in is called the "parent". For example, consider an HTML lexer and a CSS
-- lexer. Either lexer stands alone for styling their respective HTML and CSS files. However, CSS
-- can be embedded inside HTML. In this specific case, the CSS lexer is the "child" lexer with
-- the HTML lexer being the "parent". Now consider an HTML lexer and a PHP lexer. This sounds
-- a lot like the case with CSS, but there is a subtle difference: PHP _embeds itself into_
-- HTML while CSS is _embedded in_ HTML. This fundamental difference results in two types of
-- embedded lexers: a parent lexer that embeds other child lexers in it (like HTML embedding CSS),
-- and a child lexer that embeds itself into a parent lexer (like PHP embedding itself in HTML).
--
-- ##### Parent Lexer
--
-- Before embedding a child lexer into a parent lexer, the parent lexer needs to load the child
-- lexer. This is done with the `lexer.load()` function. For example, loading the CSS lexer
-- within the HTML lexer looks like:
--
--     local css = lexer.load('css')
--
-- The next part of the embedding process is telling the parent lexer when to switch over
-- to the child lexer and when to switch back. The lexer refers to these indications as the
-- "start rule" and "end rule", respectively, and are just LPeg patterns. Continuing with the
-- HTML/CSS example, the transition from HTML to CSS is when the lexer encounters a "style"
-- tag with a "type" attribute whose value is "text/css":
--
--     local css_tag = P('<style') * P(function(input, index)
--       if input:find('^[^>]+type="text/css"', index) then return true end
--     end)
--
-- This pattern looks for the beginning of a "style" tag and searches its attribute list for
-- the text "`type="text/css"`". (In this simplified example, the Lua pattern does not consider
-- whitespace between the '=' nor does it consider that using single quotes is valid.) If there
-- is a match, the functional pattern returns `true`. However, we ultimately want to style the
-- "style" tag as an HTML tag, so the actual start rule looks like this:
--
--     local css_start_rule = #css_tag * tag
--
-- Now that the parent knows when to switch to the child, it needs to know when to switch
-- back. In the case of HTML/CSS, the switch back occurs when the lexer encounters an ending
-- "style" tag, though the lexer should still style the tag as an HTML tag:
--
--     local css_end_rule = #P('</style>') * tag
--
-- Once the parent loads the child lexer and defines the child's start and end rules, it embeds
-- the child with the `lexer.embed()` function:
--
--     lex:embed(css, css_start_rule, css_end_rule)
--
-- ##### Child Lexer
--
-- The process for instructing a child lexer to embed itself into a parent is very similar to
-- embedding a child into a parent: first, load the parent lexer into the child lexer with the
-- `lexer.load()` function and then create start and end rules for the child lexer. However,
-- in this case, call `lexer.embed()` with switched arguments. For example, in the PHP lexer:
--
--     local html = lexer.load('html')
--     local php_start_rule = lex:tag('php_tag', '<?php' * lexer.space)
--     local php_end_rule = lex:tag('php_tag', '?>')
--     html:embed(lex, php_start_rule, php_end_rule)
--
-- Note that the use of a 'php_tag' tag will require the editor using the lexer to specify how
-- to highlight text with that tag. In order to avoid this, you could use the `lexer.PREPROCESSOR`
-- tag instead.
--
-- #### Lexers with Complex State
--
-- A vast majority of lexers are not stateful and can operate on any chunk of text in a
-- document. However, there may be rare cases where a lexer does need to keep track of some
-- sort of persistent state. Rather than using `lpeg.P` function patterns that set state
-- variables, it is recommended to make use of Scintilla's built-in, per-line state integers via
-- `lexer.line_state`. It was designed to accommodate up to 32 bit-flags for tracking state.
-- `lexer.line_from_position()` will return the line for any position given to an `lpeg.P`
-- function pattern. (Any positions derived from that position argument will also work.)
--
-- Writing stateful lexers is beyond the scope of this document.
--
-- ### Code Folding
--
-- When reading source code, it is occasionally helpful to temporarily hide blocks of code like
-- functions, classes, comments, etc. This is the concept of "folding". In the Textadept and
-- SciTE editors for example, little indicators in the editor margins appear next to code that
-- can be folded at places called "fold points". When the user clicks an indicator, the editor
-- hides the code associated with the indicator until the user clicks the indicator again. The
-- lexer specifies these fold points and what code exactly to fold.
--
-- The fold points for most languages occur on keywords or character sequences. Examples of
-- fold keywords are "if" and "end" in Lua and examples of fold character sequences are '{',
-- '}', "/\*", and "\*/" in C for code block and comment delimiters, respectively. However,
-- these fold points cannot occur just anywhere. For example, lexers should not recognize fold
-- keywords that appear within strings or comments. The `lexer.add_fold_point()` function allows
-- you to conveniently define fold points with such granularity. For example, consider C:
--
--     lex:add_fold_point(lexer.OPERATOR, '{', '}')
--     lex:add_fold_point(lexer.COMMENT, '/*', '*/')
--
-- The first assignment states that any '{' or '}' that the lexer tagged as an `lexer.OPERATOR`
-- is a fold point. Likewise, the second assignment states that any "/\*" or "\*/" that the
-- lexer tagged as part of a `lexer.COMMENT` is a fold point. The lexer does not consider any
-- occurrences of these characters outside their tagged elements (such as in a string) as fold
-- points. How do you specify fold keywords? Here is an example for Lua:
--
--     lex:add_fold_point(lexer.KEYWORD, 'if', 'end')
--     lex:add_fold_point(lexer.KEYWORD, 'do', 'end')
--     lex:add_fold_point(lexer.KEYWORD, 'function', 'end')
--     lex:add_fold_point(lexer.KEYWORD, 'repeat', 'until')
--
-- If your lexer has case-insensitive keywords as fold points, simply add a
-- `case_insensitive_fold_points = true` option to `lexer.new()`, and specify keywords in
-- lower case.
--
-- If your lexer needs to do some additional processing in order to determine if a tagged element
-- is a fold point, pass a function to `lex:add_fold_point()` that returns an integer. A return
-- value of `1` indicates the element is a beginning fold point and a return value of `-1`
-- indicates the element is an ending fold point. A return value of `0` indicates the element
-- is not a fold point. For example:
--
--     local function fold_strange_element(text, pos, line, s, symbol)
--       if ... then
--         return 1 -- beginning fold point
--       elseif ... then
--         return -1 -- ending fold point
--       end
--       return 0
--     end
--
--     lex:add_fold_point('strange_element', '|', fold_strange_element)
--
-- Any time the lexer encounters a '|' that is tagged as a "strange_element", it calls the
-- `fold_strange_element` function to determine if '|' is a fold point. The lexer calls these
-- functions with the following arguments: the text to identify fold points in, the beginning
-- position of the current line in the text to fold, the current line's text, the position in
-- the current line the fold point text starts at, and the fold point text itself.
--
-- #### Fold by Indentation
--
-- Some languages have significant whitespace and/or no delimiters that indicate fold points. If
-- your lexer falls into this category and you would like to mark fold points based on changes
-- in indentation, create the lexer with a `fold_by_indentation = true` option:
--
--     local lex = lexer.new(..., {fold_by_indentation = true})
--
-- ### Using Lexers
--
-- **Textadept**
--
-- Place your lexer in your *~/.textadept/lexers/* directory so you do not overwrite it when
-- upgrading Textadept. Also, lexers in this directory override default lexers. Thus, Textadept
-- loads a user *lua* lexer instead of the default *lua* lexer. This is convenient for tweaking
-- a default lexer to your liking. Then add a [file extension](#lexer.detect_extensions) for
-- your lexer if necessary.
--
-- **SciTE**
--
-- Create a *.properties* file for your lexer and `import` it in either your *SciTEUser.properties*
-- or *SciTEGlobal.properties*. The contents of the *.properties* file should contain:
--
--     file.patterns.[lexer_name]=[file_patterns]
--     lexer.$(file.patterns.[lexer_name])=scintillua.[lexer_name]
--     keywords.$(file.patterns.[lexer_name])=scintillua
--     keywords2.$(file.patterns.[lexer_name])=scintillua
--     ...
--     keywords9.$(file.patterns.[lexer_name])=scintillua
--
-- where `[lexer_name]` is the name of your lexer (minus the *.lua* extension) and
-- `[file_patterns]` is a set of file extensions to use your lexer for. The `keyword` settings are
-- only needed if another SciTE properties file has defined keyword sets for `[file_patterns]`.
-- The `scintillua` keyword setting instructs Scintillua to use the keyword sets defined within
-- the lexer. You can override a lexer's keyword set(s) by specifying your own in the same order
-- that the lexer calls `lex:set_word_list()`. For example, the Lua lexer's first set of keywords
-- is for reserved words, the second is for built-in global functions, the third is for library
-- functions, the fourth is for built-in global constants, and the fifth is for library constants.
--
-- SciTE assigns styles to tag names in order to perform syntax highlighting. Since the set of
-- tag names used for a given language changes, your *.properties* file should specify styles
-- for tag names instead of style numbers. For example:
--
--     scintillua.styles.my_tag=$(scintillua.styles.keyword),bold
--
-- ### Migrating Legacy Lexers
--
-- Legacy lexers are of the form:
--
--     local lexer = require('lexer')
--     local token, word_match = lexer.token, lexer.word_match
--     local P, S = lpeg.P, lpeg.S
--
--     local lex = lexer.new('?')
--
--     -- Whitespace.
--     lex:add_rule('whitespace', token(lexer.WHITESPACE, lexer.space^1))
--
--     -- Keywords.
--     lex:add_rule('keyword', token(lexer.KEYWORD, word_match{
--       [...]
--     }))
--
--     [... other rule definitions ...]
--
--     -- Custom.
--     lex:add_rule('custom_rule', token('custom_token', ...))
--     lex:add_style('custom_token', lexer.styles.keyword .. {bold = true})
--
--     -- Fold points.
--     lex:add_fold_point(lexer.OPERATOR, '{', '}')
--
--     return lex
--
-- While Scintillua will mostly handle such legacy lexers just fine without any changes, it is
-- recommended that you migrate yours. The migration process is fairly straightforward:
--
-- 1. `lexer` exists in the default lexer environment, so `require('lexer')` should be replaced
--    by simply `lexer`. (Keep in mind `local lexer = lexer` is a Lua idiom.)
-- 2. Every lexer created using `lexer.new()` should no longer specify a lexer name by string,
--    but should instead use `...` (three dots), which evaluates to the lexer's filename or
--    alternative name in embedded lexer applications.
-- 3. Every lexer created using `lexer.new()` now includes a rule to match whitespace. Unless
--    your lexer has significant whitespace, you can remove your legacy lexer's whitespace
--    token and rule. Otherwise, your defined whitespace rule will replace the default one.
-- 4. The concept of tokens has been replaced with tags. Instead of calling a `token()` function,
--    call [`lex:tag()`](#lexer.tag) instead.
-- 5. Lexers now support replaceable word lists. Instead of calling `lexer.word_match()` with
--    large word lists, call it as an instance method with an identifier string (typically
--    something like `lexer.KEYWORD`). Then at the end of the lexer (before `return lex`), call
--    [`lex:set_word_list()`](#lexer.set_word_list) with the same identifier and the usual
--    list of words to match. This allows users of your lexer to call `lex:set_word_list()`
--    with their own set of words should they wish to.
-- 6. Lexers no longer specify styling information. Remove any calls to `lex:add_style()`. You
--    may need to add styling information for custom tags to your editor's theme.
-- 7. `lexer.last_char_includes()` has been deprecated in favor of the new `lexer.after_set()`.
--    Use the character set and pattern as arguments to that new function.
--
-- As an example, consider the following sample legacy lexer:
--
--     local lexer = require('lexer')
--     local token, word_match = lexer.token, lexer.word_match
--     local P, S = lpeg.P, lpeg.S
--
--     local lex = lexer.new('legacy')
--
--     lex:add_rule('whitespace', token(lexer.WHITESPACE, lexer.space^1))
--     lex:add_rule('keyword', token(lexer.KEYWORD, word_match('foo bar baz')))
--     lex:add_rule('custom', token('custom', 'quux'))
--     lex:add_style('custom', lexer.styles.keyword .. {bold = true})
--     lex:add_rule('identifier', token(lexer.IDENTIFIER, lexer.word))
--     lex:add_rule('string', token(lexer.STRING, lexer.range('"')))
--     lex:add_rule('comment', token(lexer.COMMENT, lexer.to_eol('#')))
--     lex:add_rule('number', token(lexer.NUMBER, lexer.number))
--     lex:add_rule('operator', token(lexer.OPERATOR, S('+-*/%^=<>,.()[]{}')))
--
--     lex:add_fold_point(lexer.OPERATOR, '{', '}')
--
--     return lex
--
-- Following the migration steps would yield:
--
--     local lexer = lexer
--     local P, S = lpeg.P, lpeg.S
--
--     local lex = lexer.new(...)
--
--     lex:add_rule('keyword', lex:tag(lexer.KEYWORD, lex:word_match(lexer.KEYWORD)))
--     lex:add_rule('custom', lex:tag('custom', 'quux'))
--     lex:add_rule('identifier', lex:tag(lexer.IDENTIFIER, lexer.word))
--     lex:add_rule('string', lex:tag(lexer.STRING, lexer.range('"')))
--     lex:add_rule('comment', lex:tag(lexer.COMMENT, lexer.to_eol('#')))
--     lex:add_rule('number', lex:tag(lexer.NUMBER, lexer.number))
--     lex:add_rule('operator', lex:tag(lexer.OPERATOR, S('+-*/%^=<>,.()[]{}')))
--
--     lex:add_fold_point(lexer.OPERATOR, '{', '}')
--
--     lex:set_word_list(lexer.KEYWORD, {'foo', 'bar', 'baz'})
--
--     return lex
--
-- Any editors using this lexer would have to add a style for the 'custom' tag.
--
-- ### Considerations
--
-- #### Performance
--
-- There might be some slight overhead when initializing a lexer, but loading a file from disk
-- into Scintilla is usually more expensive. Actually painting the syntax highlighted text to
-- the screen is often more expensive than the lexing operation. On modern computer systems,
-- I see no difference in speed between Lua lexers and Scintilla's C++ ones. Optimize lexers for
-- speed by re-arranging `lexer.add_rule()` calls so that the most common rules match first. Do
-- keep in mind that order matters for similar rules.
--
-- In some cases, folding may be far more expensive than lexing, particularly in lexers with a
-- lot of potential fold points. If your lexer is exhibiting signs of slowness, try disabling
-- folding in your text editor first. If that speeds things up, you can try reducing the number
-- of fold points you added, overriding `lexer.fold()` with your own implementation, or simply
-- eliminating folding support from your lexer.
--
-- #### Limitations
--
-- Embedded preprocessor languages like PHP cannot completely embed themselves into their parent
-- languages because the parent's tagged patterns do not support start and end rules. This
-- mostly goes unnoticed, but code like
--
--     <div id="<?php echo $id; ?>">
--
-- will not style correctly. Also, these types of languages cannot currently embed themselves
-- into their parent's child languages either.
--
-- A language cannot embed itself into something like an interpolated string because it is
-- possible that if lexing starts within the embedded entity, it will not be detected as such,
-- so a child to parent transition cannot happen. For example, the following Ruby code will
-- not style correctly:
--
--     sum = "1 + 2 = #{1 + 2}"
--
-- Also, there is the potential for recursion for languages embedding themselves within themselves.
--
-- #### Troubleshooting
--
-- Errors in lexers can be tricky to debug. Lexers print Lua errors to `io.stderr` and `_G.print()`
-- statements to `io.stdout`. Running your editor from a terminal is the easiest way to see
-- errors as they occur.
--
-- #### Risks
--
-- Poorly written lexers have the ability to crash Scintilla (and thus its containing application),
-- so unsaved data might be lost. However, I have only observed these crashes in early lexer
-- development, when syntax errors or pattern errors are present. Once the lexer actually
-- starts processing and tagging text (either correctly or incorrectly, it does not matter),
-- I have not observed any crashes.
--
-- #### Acknowledgements
--
-- Thanks to Peter Odding for his [lexer post][] on the Lua mailing list that provided inspiration,
-- and thanks to Roberto Ierusalimschy for LPeg.
--
-- [lexer post]: http://lua-users.org/lists/lua-l/2007-04/msg00116.html
-- @module lexer
local M = {}
--- The tag name for default elements.
-- @field DEFAULT
--- The tag name for comment elements.
-- @field COMMENT
--- The tag name for string elements.
-- @field STRING
--- The tag name for number elements.
-- @field NUMBER
--- The tag name for keyword elements.
-- @field KEYWORD
--- The tag name for identifier elements.
-- @field IDENTIFIER
--- The tag name for operator elements.
-- @field OPERATOR
--- The tag name for error elements.
-- @field ERROR
--- The tag name for preprocessor elements.
-- @field PREPROCESSOR
--- The tag name for constant elements.
-- @field CONSTANT
--- The tag name for variable elements.
-- @field VARIABLE
--- The tag name for function elements.
-- @field FUNCTION
--- The tag name for class elements.
-- @field CLASS
--- The tag name for type elements.
-- @field TYPE
--- The tag name for label elements.
-- @field LABEL
--- The tag name for regex elements.
-- @field REGEX
--- The tag name for embedded elements.
-- @field EMBEDDED
--- The tag name for builtin function elements.
-- @field FUNCTION_BUILTIN
--- The tag name for builtin constant elements.
-- @field CONSTANT_BUILTIN
--- The tag name for function method elements.
-- @field FUNCTION_METHOD
--- The tag name for function tag elements, typically in markup.
-- @field TAG
--- The tag name for function attribute elements, typically in markup.
-- @field ATTRIBUTE
--- The tag name for builtin variable elements.
-- @field VARIABLE_BUILTIN
--- The tag name for heading elements, typically in markup.
-- @field HEADING
--- The tag name for bold elements, typically in markup.
-- @field BOLD
--- The tag name for builtin italic elements, typically in markup.
-- @field ITALIC
--- The tag name for underlined elements, typically in markup.
-- @field UNDERLINE
--- The tag name for code elements, typically in markup.
-- @field CODE
--- The tag name for link elements, typically in markup.
-- @field LINK
--- The tag name for reference elements, typically in markup.
-- @field REFERENCE
--- The tag name for annotation elements.
-- @field ANNOTATION
--- The tag name for list item elements, typically in markup.
-- @field LIST
--- The initial (root) fold level.
-- @field FOLD_BASE
--- Flag indicating that the line is blank.
-- @field FOLD_BLANK
--- Flag indicating the line is fold point.
-- @field FOLD_HEADER
-- This comment is needed for LDoc to process the previous field.
if not lpeg then lpeg = require('lpeg') end -- Scintillua's Lua environment defines _G.lpeg
local lpeg = lpeg
local P, R, S, V, B = lpeg.P, lpeg.R, lpeg.S, lpeg.V, lpeg.B
local Ct, Cc, Cp, Cmt, C = lpeg.Ct, lpeg.Cc, lpeg.Cp, lpeg.Cmt, lpeg.C
--- Default tags.
local default = {
  'whitespace', 'comment', 'string', 'number', 'keyword', 'identifier', 'operator', 'error',
  'preprocessor', 'constant', 'variable', 'function', 'class', 'type', 'label', 'regex', 'embedded',
  'function.builtin', 'constant.builtin', 'function.method', 'tag', 'attribute', 'variable.builtin',
  'heading', 'bold', 'italic', 'underline', 'code', 'link', 'reference', 'annotation', 'list'
}
for _, name in ipairs(default) do M[name:upper():gsub('%.', '_')] = name end
--- Names for predefined Scintilla styles.
-- Having these here simplifies style number handling between Scintillua and Scintilla.
local predefined = {
  'default', 'line.number', 'brace.light', 'brace.bad', 'control.char', 'indent.guide', 'call.tip',
  'fold.display.text'
}
for _, name in ipairs(predefined) do M[name:upper():gsub('%.', '_')] = name end
--- Creates and returns a pattern that tags pattern *patt* with name *name* in lexer *lexer*.
-- If *name* is not a predefined tag name, its Scintilla style will likely need to be defined
-- by the editor or theme using this lexer.
-- @param lexer The lexer to tag the given pattern in.
-- @param name The name to use.
-- @param patt The LPeg pattern to tag.
-- @return pattern
-- @usage local number = lex:tag(lexer.NUMBER, lexer.number)
-- @usage local addition = lex:tag('addition', '+' * lexer.word)
function M.tag(lexer, name, patt)
  if not lexer._TAGS then
    -- Create the initial maps for tag names to style numbers and styles.
    local tags = {}
    for i, name in ipairs(default) do tags[name], tags[i] = i, name end
    for i, name in ipairs(predefined) do tags[name], tags[i + 32] = i + 32, name end
    lexer._TAGS, lexer._num_styles = tags, #default + 1
    lexer._extra_tags = {}
  end
  if not assert(lexer._TAGS, 'not a lexer instance')[name] then
    local num_styles = lexer._num_styles
    if num_styles == 33 then num_styles = num_styles + 8 end -- skip predefined
    assert(num_styles <= 256, 'too many styles defined (256 MAX)')
    lexer._TAGS[name], lexer._TAGS[num_styles], lexer._num_styles = num_styles, name, num_styles + 1
    lexer._extra_tags[name] = true
    -- If the lexer is a proxy or a child that embedded itself, make this tag name known to
    -- the parent lexer.
    if lexer._lexer then lexer._lexer:tag(name, false) end
  end
  return Cc(name) * (P(patt) / 0) * Cp()
end
--- Returns a unique grammar rule name for the given lexer's i-th word list.
local function word_list_id(lexer, i) return lexer._name .. '_wordlist' .. i end
--- Either returns a pattern for lexer *lexer* (if given) that matches one word in the word list
-- identified by string *word_list*, ignoring case if *case_sensitive* is `true`, or, if *lexer*
-- is not given, creates and returns a pattern that matches any single word in list or string
-- *word_list*, ignoring case if *case_insensitive* is `true`.
-- This is a convenience function for simplifying a set of ordered choice word patterns and
-- potentially allowing downstream users to configure word lists.
-- If there is ultimately no word list set via `set_word_list()`, no error will be raised,
-- but the returned pattern will not match anything.
-- @param[opt] lexer Optional lexer to match a word in a wordlist for. This parameter may be
--   omitted for lexer-agnostic matching.
-- @param word_list Either a string name of the word list to match from if *lexer* is given,
--   or, if *lexer* is omitted, a list of words or a string list of words separated by spaces.
-- @param[opt] case_insensitive Optional boolean flag indicating whether or not the word match
--   is case-insensitive. The default value is `false`.
-- @return pattern
-- @usage lex:add_rule('keyword', lex:tag(lexer.KEYWORD, lex:word_match(lexer.KEYWORD)))
-- @usage local keyword = lex:tag(lexer.KEYWORD, lexer.word_match{'foo', 'bar', 'baz'})
-- @usage local keyword = lex:tag(lexer.KEYWORD, lexer.word_match({'foo-bar', 'foo-baz',
--   'bar-foo', 'bar-baz', 'baz-foo', 'baz-bar'}, true))
-- @usage local keyword = lex:tag(lexer.KEYWORD, lexer.word_match('foo bar baz'))
function M.word_match(lexer, word_list, case_insensitive)
  if type(lexer) == 'table' and getmetatable(lexer) then
    if lexer._lexer then
      -- If this lexer is a proxy (e.g. rails), get the true parent (ruby) in order to get the
      -- parent's word list. If this lexer is a child embedding itself (e.g. php), continue
      -- getting its word list, not the parent's (html).
      local parent = lexer._lexer
      if not parent._CHILDREN or not parent._CHILDREN[lexer] then lexer = parent end
    end
    if not lexer._WORDLISTS then lexer._WORDLISTS = {case_insensitive = {}} end
    local i = lexer._WORDLISTS[word_list] or #lexer._WORDLISTS + 1
    lexer._WORDLISTS[word_list], lexer._WORDLISTS[i] = i, '' -- empty placeholder word list
    lexer._WORDLISTS.case_insensitive[i] = case_insensitive
    return V(word_list_id(lexer, i))
  end
  -- Lexer-agnostic word match.
  word_list, case_insensitive = lexer, word_list
  if type(word_list) == 'string' then
    local words = word_list -- space-separated list of words
    word_list = {}
    for word in words:gmatch('%S+') do word_list[#word_list + 1] = word end
  end
  local word_chars = M.alnum + '_'
  local extra_chars = ''
  for _, word in ipairs(word_list) do
    word_list[case_insensitive and word:lower() or word] = true
    for char in word:gmatch('[^%w_%s]') do
      if not extra_chars:find(char, 1, true) then extra_chars = extra_chars .. char end
    end
  end
  if extra_chars ~= '' then word_chars = word_chars + S(extra_chars) end
  -- Optimize small word sets as ordered choice. "Small" is arbitrary.
  if #word_list <= 6 and not case_insensitive then
    local choice = P(false)
    for _, word in ipairs(word_list) do choice = choice + word:match('%S+') end
    return choice * -word_chars
  end
  return Cmt(word_chars^1, function(input, index, word)
    if case_insensitive then word = word:lower() end
    return word_list[word]
  end)
end
--- Sets in lexer *lexer* the word list identified by string or number *name* to string or
-- list *word_list*, appending to any existing word list if *append* is `true`.
-- This only has an effect if *lexer* uses `word_match()` to reference the given list.
-- Case-insensitivity is specified by `word_match()`.
-- @param lexer The lexer to add the given word list to.
-- @param name The string name or number of the word list to set.
-- @param word_list A list of words or a string list of words separated by spaces.
-- @param append Whether or not to append *word_list* to the existing word list (if any). The
--   default value is `false`.
function M.set_word_list(lexer, name, word_list, append)
  if word_list == 'scintillua' then return end -- for SciTE
  if lexer._lexer then
    -- If this lexer is a proxy (e.g. rails), get the true parent (ruby) in order to set the
    -- parent's word list. If this lexer is a child embedding itself (e.g. php), continue
    -- setting its word list, not the parent's (html).
    local parent = lexer._lexer
    if not parent._CHILDREN or not parent._CHILDREN[lexer] then lexer = parent end
  end
  assert(lexer._WORDLISTS, 'lexer has no word lists')
  local i = tonumber(lexer._WORDLISTS[name]) or name -- lexer._WORDLISTS[name] --> i
  if type(i) ~= 'number' or i > #lexer._WORDLISTS then return end -- silently return
  if type(word_list) == 'string' then
    local list = {}
    for word in word_list:gmatch('%S+') do list[#list + 1] = word end
    word_list = list
  end
  if not append or lexer._WORDLISTS[i] == '' then
    lexer._WORDLISTS[i] = word_list
  else
    local list = lexer._WORDLISTS[i]
    for _, word in ipairs(word_list) do list[#list + 1] = word end
  end
  lexer._grammar_table = nil -- invalidate
end
--- Adds pattern *rule* identified by string *id* to the ordered list of rules for lexer *lexer*.
-- @param lexer The lexer to add the given rule to.
-- @param id The id associated with this rule. It does not have to be the same as the name
--   passed to `tag()`.
-- @param rule The LPeg pattern of the rule.
-- @see modify_rule
function M.add_rule(lexer, id, rule)
  if lexer._lexer then lexer = lexer._lexer end -- proxy; get true parent
  if not lexer._rules then lexer._rules = {} end
  if id == 'whitespace' and lexer._rules[id] then -- legacy
    lexer:modify_rule(id, rule)
    return
  end
  lexer._rules[#lexer._rules + 1], lexer._rules[id] = id, rule
  lexer._grammar_table = nil -- invalidate
end
--- Replaces in lexer *lexer* the existing rule identified by string *id* with pattern *rule*.
-- @param lexer The lexer to modify.
-- @param id The id associated with this rule.
-- @param rule The LPeg pattern of the rule.
function M.modify_rule(lexer, id, rule)
  if lexer._lexer then lexer = lexer._lexer end -- proxy; get true parent
  assert(lexer._rules[id], 'rule does not exist')
  lexer._rules[id] = rule
  lexer._grammar_table = nil -- invalidate
end
--- Returns a unique grammar rule name for the given lexer's rule name.
local function rule_id(lexer, name) return lexer._name .. '.' .. name end
--- Returns the rule identified by string *id*.
-- @param lexer The lexer to fetch a rule from.
-- @param id The id of the rule to fetch.
-- @return pattern
function M.get_rule(lexer, id)
  if lexer._lexer then lexer = lexer._lexer end -- proxy; get true parent
  if id == 'whitespace' then return V(rule_id(lexer, id)) end -- special case
  return assert(lexer._rules[id], 'rule does not exist')
end
--- Embeds child lexer *child* in parent lexer *lexer* using patterns *start_rule* and *end_rule*,
-- which signal the beginning and end of the embedded lexer, respectively.
-- @param lexer The parent lexer.
-- @param child The child lexer.
-- @param start_rule The pattern that signals the beginning of the embedded lexer.
-- @param end_rule The pattern that signals the end of the embedded lexer.
-- @usage html:embed(css, css_start_rule, css_end_rule)
-- @usage html:embed(lex, php_start_rule, php_end_rule) -- from php lexer
function M.embed(lexer, child, start_rule, end_rule)
  if lexer._lexer then lexer = lexer._lexer end -- proxy; get true parent
  -- Add child rules.
  assert(child._rules, 'cannot embed lexer with no rules')
  if not child._start_rules then child._start_rules = {} end
  if not child._end_rules then child._end_rules = {} end
  child._start_rules[lexer], child._end_rules[lexer] = start_rule, end_rule
  if not lexer._CHILDREN then lexer._CHILDREN = {} end
  lexer._CHILDREN[#lexer._CHILDREN + 1], lexer._CHILDREN[child] = child, true
  -- Add child tags.
  for name in pairs(child._extra_tags) do lexer:tag(name, true) end
  -- Add child fold symbols.
  if child._fold_points then
    for tag_name, symbols in pairs(child._fold_points) do
      if tag_name ~= '_symbols' then
        for symbol, v in pairs(symbols) do lexer:add_fold_point(tag_name, symbol, v) end
      end
    end
  end
  -- Add child word lists.
  if child._WORDLISTS then
    for name, i in pairs(child._WORDLISTS) do
      if type(name) == 'string' and type(i) == 'number' then
        name = child._name .. '.' .. name
        lexer:word_match(name) -- for side effects
        lexer:set_word_list(name, child._WORDLISTS[i])
      end
    end
  end
  child._lexer = lexer -- use parent's rules if child is embedding itself
end
--- Adds to lexer *lexer* a fold point whose beginning and end points are tagged with string
-- *tag_name* tags and have string content *start_symbol* and *end_symbol*, respectively.
-- In the event that *start_symbol* may or may not be a fold point depending on context, and that
-- additional processing is required, *end_symbol* may be a function that ultimately returns
-- `1` (indicating a beginning fold point), `-1` (indicating an ending fold point), or `0`
-- (indicating no fold point). That function is passed the following arguments:
--
--   - `text`: The text being processed for fold points.
--   - `pos`: The position in *text* of the beginning of the line currently being processed.
--   - `line`: The text of the line currently being processed.
--   - `s`: The position of *start_symbol* in *line*.
--   - `symbol`: *start_symbol* itself.
-- @param lexer The lexer to add a fold point to.
-- @param tag_name The tag name for text that indicates a fold point.
-- @param start_symbol The text that indicates the beginning of a fold point.
-- @param end_symbol Either the text that indicates the end of a fold point, or a function that
--   returns whether or not *start_symbol* is a beginning fold point (1), an ending fold point
--   (-1), or not a fold point at all (0).
-- @usage lex:add_fold_point(lexer.OPERATOR, '{', '}')
-- @usage lex:add_fold_point(lexer.KEYWORD, 'if', 'end')
-- @usage lex:add_fold_point('custom', function(text, pos, line, s, symbol) ... end)
function M.add_fold_point(lexer, tag_name, start_symbol, end_symbol)
  if not start_symbol and not end_symbol then return end -- from legacy fold_consecutive_lines()
  if not lexer._fold_points then lexer._fold_points = {_symbols = {}} end
  local symbols = lexer._fold_points._symbols
  if not lexer._fold_points[tag_name] then lexer._fold_points[tag_name] = {} end
  if lexer._case_insensitive_fold_points then
    start_symbol = start_symbol:lower()
    if type(end_symbol) == 'string' then end_symbol = end_symbol:lower() end
  end
  if type(end_symbol) == 'string' then
    if not symbols[end_symbol] then symbols[#symbols + 1], symbols[end_symbol] = end_symbol, true end
    lexer._fold_points[tag_name][start_symbol] = 1
    lexer._fold_points[tag_name][end_symbol] = -1
  else
    lexer._fold_points[tag_name][start_symbol] = end_symbol -- function or int
  end
  if not symbols[start_symbol] then
    symbols[#symbols + 1], symbols[start_symbol] = start_symbol, true
  end
  -- If the lexer is a proxy or a child that embedded itself, copy this fold point to the
  -- parent lexer.
  if lexer._lexer then lexer._lexer:add_fold_point(tag_name, start_symbol, end_symbol) end
end
--- Recursively adds the rules for the given lexer and its children to the given grammar.
-- @param g The grammar to add rules to.
-- @param lexer The lexer whose rules to add.
local function add_lexer(g, lexer)
  local rule = P(false)
  -- Add this lexer's rules.
  for _, name in ipairs(lexer._rules) do
    local id = rule_id(lexer, name)
    g[id] = lexer._rules[name] -- ['lua.keyword'] = keyword_patt
    rule = rule + V(id) -- V('lua.keyword') + V('lua.function') + V('lua.constant') + ...
  end
  local any_id = lexer._name .. '_fallback'
  g[any_id] = lexer:tag(M.DEFAULT, M.any) -- ['lua_fallback'] = any_char
  rule = rule + V(any_id) -- ... + V('lua.operator') + V('lua_fallback')
  -- Add this lexer's word lists.
  if lexer._WORDLISTS then
    for i = 1, #lexer._WORDLISTS do
      local id = word_list_id(lexer, i)
      local list, case_insensitive = lexer._WORDLISTS[i], lexer._WORDLISTS.case_insensitive[i]
      local patt = list ~= '' and M.word_match(list, case_insensitive) or P(false)
      g[id] = patt -- ['lua_wordlist.1'] = word_match_patt or P(false)
    end
  end
  -- Add this child lexer's end rules.
  if lexer._end_rules then
    for parent, end_rule in pairs(lexer._end_rules) do
      local back_id = lexer._name .. '_to_' .. parent._name
      g[back_id] = end_rule -- ['css_to_html'] = css_end_rule
      rule = rule - V(back_id) + -- (V('css.property') + ... + V('css_fallback')) - V('css_to_html')
      V(back_id) * V(parent._name) -- V('css_to_html') * V('html')
    end
  end
  -- Add this child lexer's start rules.
  if lexer._start_rules then
    for parent, start_rule in pairs(lexer._start_rules) do
      local to_id = parent._name .. '_to_' .. lexer._name
      g[to_id] = start_rule * V(lexer._name) -- ['html_to_css'] = css_start_rule * V('css')
    end
  end
  -- Finish adding this lexer's rules.
  local rule_id = lexer._name .. '_rule'
  g[rule_id] = rule -- ['lua_rule'] = V('lua.keyword') + ... + V('lua_fallback')
  g[lexer._name] = V(rule_id)^0 -- ['lua'] = V('lua_rule')^0
  -- Add this lexer's children's rules.
  -- TODO: preprocessor languages like PHP should also embed themselves into their parent's
  -- children like HTML's CSS and Javascript.
  if not lexer._CHILDREN then return end
  for _, child in ipairs(lexer._CHILDREN) do
    add_lexer(g, child)
    local to_id = lexer._name .. '_to_' .. child._name
    g[rule_id] = V(to_id) + g[rule_id] -- ['html_rule'] = V('html_to_css') + V('html.comment') + ...
    -- Add a child's inherited parent's rules (e.g. rhtml parent with rails child inheriting ruby).
    if child._parent_name then
      local name = child._name
      child._name = child._parent_name -- ensure parent and transition rule names are correct
      add_lexer(g, child)
      child._name = name -- restore
      local to_id = lexer._name .. '_to_' .. child._parent_name
      g[rule_id] = V(to_id) + g[rule_id] -- ['html_rule'] = V('html_to_ruby') + V('html.comment') + ...
    end
  end
end
--- Returns a grammar for the given lexer and initial rule, (re)constructing it if necessary.
-- @param lexer The lexer to build a grammar for.
-- @param init_style The current style. Multiple-language lexers use this to determine which
--   language to start lexing in.
local function build_grammar(lexer, init_style)
  if not lexer._rules then return end
  if not lexer._initial_rule then lexer._initial_rule = lexer._parent_name or lexer._name end
  if not lexer._grammar_table then
    local grammar = {lexer._initial_rule}
    if not lexer._parent_name then
      add_lexer(grammar, lexer)
      -- {'lua',
      --   ['lua.keyword'] = patt, ['lua.function'] = patt, ...,
      --   ['lua_wordlist.1'] = patt, ['lua_wordlist.2'] = patt, ...,
      --   ['lua_rule'] = V('lua.keyword') + ... + V('lua_fallback'),
      --   ['lua'] = V('lua_rule')^0
      -- }
      -- {'html'
      --   ['html.comment'] = patt, ['html.doctype'] = patt, ...,
      --   ['html_wordlist.1'] = patt, ['html_wordlist.2'] = patt, ...,
      --   ['html_rule'] = V('html_to_css') * V('css') + V('html.comment') + ... + V('html_fallback'),
      --   ['html'] = V('html')^0,
      --   ['css.property'] = patt, ['css.value'] = patt, ...,
      --   ['css_wordlist.1'] = patt, ['css_wordlist.2'] = patt, ...,
      --   ['css_to_html'] = patt,
      --   ['css_rule'] = ((V('css.property') + ... + V('css_fallback')) - V('css_to_html')) +
      --     V('css_to_html') * V('html'),
      --   ['html_to_css'] = patt,
      --   ['css'] = V('css_rule')^0
      -- }
    else
      local name = lexer._name
      lexer._name = lexer._parent_name -- ensure parent and transition rule names are correct
      add_lexer(grammar, lexer)
      lexer._name = name -- restore
      -- {'html',
      --   ...
      --   ['html_rule'] = V('html_to_php') * V('php') + V('html_to_css') * V('css') +
      --     V('html.comment') + ... + V('html_fallback'),
      --   ...
      --   ['php.keyword'] = patt, ['php.type'] = patt, ...,
      --   ['php_wordlist.1'] = patt, ['php_wordlist.2'] = patt, ...,
      --   ['php_to_html'] = patt,
      --   ['php_rule'] = ((V('php.keyword') + ... + V('php_fallback')) - V('php_to_html')) +
      --     V('php_to_html') * V('html')
      --   ['html_to_php'] = patt,
      --   ['php'] = V('php_rule')^0
      -- }
    end
    lexer._grammar, lexer._grammar_table = Ct(P(grammar)), grammar
  end
  -- For multilang lexers, build a new grammar whose initial rule is the current language
  -- if necessary. LPeg does not allow a variable initial rule.
  if lexer._CHILDREN then
    for style_num, tag in ipairs(lexer._TAGS) do
      if style_num == init_style then
        local lexer_name = tag:match('^whitespace%.(.+)$') or lexer._parent_name or lexer._name
        if lexer._initial_rule == lexer_name then break end
        if not lexer._grammar_table[lexer_name] then
          -- For proxy lexers like RHTML, the 'whitespace.rhtml' tag would produce the 'rhtml'
          -- lexer name, but there is no 'rhtml' rule. It should be the 'html' rule (parent)
          -- instead.
          lexer_name = lexer._parent_name or lexer._name
        end
        lexer._initial_rule = lexer_name
        lexer._grammar_table[1] = lexer._initial_rule
        lexer._grammar = Ct(P(lexer._grammar_table))
        return lexer._grammar
      end
    end
  end
  return lexer._grammar
end
--- Lexes a chunk of text *text* (that has an initial style number of *init_style*) using lexer
-- *lexer*, returning a list of tag names and positions.
-- @param lexer The lexer to lex text with.
-- @param text The text in the buffer to lex.
-- @param init_style The current style. Multiple-language lexers use this to determine which
--   language to start lexing in.
-- @return list of tag names and positions.
function M.lex(lexer, text, init_style)
  local grammar = build_grammar(lexer, init_style)
  if not grammar then return {M.DEFAULT, #text + 1} end
  if M._standalone then M._text, M.line_state = text, {} end
  if lexer._lex_by_line then
    local line_from_position = M.line_from_position
    local function append(tags, line_tags, offset)
      for i = 1, #line_tags, 2 do
        tags[#tags + 1], tags[#tags + 2] = line_tags[i], line_tags[i + 1] + offset
      end
    end
    local tags = {}
    local offset = 0
    rawset(M, 'line_from_position', function(pos) return line_from_position(pos + offset) end)
    for line in text:gmatch('[^\r\n]*\r?\n?') do
      local line_tags = grammar:match(line)
      if line_tags then append(tags, line_tags, offset) end
      offset = offset + #line
      -- Use the default tag to the end of the line if none was specified.
      if tags[#tags] ~= offset + 1 then
        tags[#tags + 1], tags[#tags + 2] = 'default', offset + 1
      end
    end
    rawset(M, 'line_from_position', line_from_position)
    return tags
  end
  return grammar:match(text)
end
--- Determines fold points in a chunk of text *text* using lexer *lexer*, returning a table of
-- fold levels associated with line numbers.
-- *text* starts on line number *start_line* with a beginning fold level of *start_level*
-- in the buffer.
-- @param lexer The lexer to fold text with.
-- @param text The text in the buffer to fold.
-- @param start_line The line number *text* starts on, counting from 1.
-- @param start_level The fold level *text* starts on.
-- @return table of fold levels associated with line numbers.
function M.fold(lexer, text, start_line, start_level)
  local folds = {}
  if text == '' then return folds end
  local fold = M.property_int['fold'] > 0
  local FOLD_BASE = M.FOLD_BASE or 0x400
  local FOLD_HEADER, FOLD_BLANK = M.FOLD_HEADER or 0x2000, M.FOLD_BLANK or 0x1000
  if M._standalone then M._text, M.line_state = text, {} end
  if fold and lexer._fold_points then
    local lines = {}
    for p, l in (text .. '\n'):gmatch('()(.-)\r?\n') do lines[#lines + 1] = {p, l} end
    local fold_zero_sum_lines = M.property_int['fold.scintillua.on.zero.sum.lines'] > 0
    local fold_compact = M.property_int['fold.scintillua.compact'] > 0
    local fold_points = lexer._fold_points
    local fold_point_symbols = fold_points._symbols
    local style_at, fold_level = M.style_at, M.fold_level
    local line_num, prev_level = start_line, start_level
    local current_level = prev_level
    for _, captures in ipairs(lines) do
      local pos, line = captures[1], captures[2]
      if line ~= '' then
        if lexer._case_insensitive_fold_points then line = line:lower() end
        local ranges = {}
        local function is_valid_range(s, e)
          if not s or not e then return false end
          for i = 1, #ranges - 1, 2 do
            local range_s, range_e = ranges[i], ranges[i + 1]
            if s >= range_s and s <= range_e or e >= range_s and e <= range_e then
              return false
            end
          end
          ranges[#ranges + 1] = s
          ranges[#ranges + 1] = e
          return true
        end
        local level_decreased = false
        for _, symbol in ipairs(fold_point_symbols) do
          local word = not symbol:find('[^%w_]')
          local s, e = line:find(symbol, 1, true)
          while is_valid_range(s, e) do
            -- if not word or line:find('^%f[%w_]' .. symbol .. '%f[^%w_]', s) then
            local word_before = s > 1 and line:find('^[%w_]', s - 1)
            local word_after = line:find('^[%w_]', e + 1)
            if not word or not (word_before or word_after) then
              local style_name = style_at[pos + s - 1]
              local symbols = fold_points[style_name]
              if not symbols and style_name:find('%.') then
                symbols = fold_points[style_name:match('^[^.]+')]
              end
              local level = symbols and symbols[symbol]
              if type(level) == 'function' then
                level = level(text, pos, line, s, symbol)
              end
              if type(level) == 'number' then
                current_level = current_level + level
                if level < 0 and current_level < prev_level then
                  -- Potential zero-sum line. If the level were to go back up on the same line,
                  -- the line may be marked as a fold header.
                  level_decreased = true
                end
              end
            end
            s, e = line:find(symbol, s + 1, true)
          end
        end
        folds[line_num] = prev_level
        if current_level > prev_level then
          folds[line_num] = prev_level + FOLD_HEADER
        elseif level_decreased and current_level == prev_level and fold_zero_sum_lines then
          if line_num > start_line then
            folds[line_num] = prev_level - 1 + FOLD_HEADER
          else
            -- Typing within a zero-sum line.
            local level = fold_level[line_num] - 1
            if level > FOLD_HEADER then level = level - FOLD_HEADER end
            if level > FOLD_BLANK then level = level - FOLD_BLANK end
            folds[line_num] = level + FOLD_HEADER
            current_level = current_level + 1
          end
        end
        if current_level < FOLD_BASE then current_level = FOLD_BASE end
        prev_level = current_level
      else
        folds[line_num] = prev_level + (fold_compact and FOLD_BLANK or 0)
      end
      line_num = line_num + 1
    end
  elseif fold and
    (lexer._fold_by_indentation or M.property_int['fold.scintillua.by.indentation'] > 0) then
    -- Indentation based folding.
    -- Calculate indentation per line.
    local indentation = {}
    for indent, line in (text .. '\n'):gmatch('([\t ]*)([^\r\n]*)\r?\n') do
      indentation[#indentation + 1] = line ~= '' and #indent
    end
    -- Find the first non-blank line before start_line. If the current line is indented, make
    -- that previous line a header and update the levels of any blank lines inbetween. If the
    -- current line is blank, match the level of the previous non-blank line.
    local current_level = start_level
    for i = start_line, 1, -1 do
      local level = M.fold_level[i]
      if level >= FOLD_HEADER then level = level - FOLD_HEADER end
      if level < FOLD_BLANK then
        local indent = M.indent_amount[i]
        if indentation[1] and indentation[1] > indent then
          folds[i] = FOLD_BASE + indent + FOLD_HEADER
          for j = i + 1, start_line - 1 do folds[j] = start_level + FOLD_BLANK end
        elseif not indentation[1] then
          current_level = FOLD_BASE + indent
        end
        break
      end
    end
    -- Iterate over lines, setting fold numbers and fold flags.
    for i = 1, #indentation do
      if indentation[i] then
        current_level = FOLD_BASE + indentation[i]
        folds[start_line + i - 1] = current_level
        for j = i + 1, #indentation do
          if indentation[j] then
            if FOLD_BASE + indentation[j] > current_level then
              folds[start_line + i - 1] = current_level + FOLD_HEADER
              current_level = FOLD_BASE + indentation[j] -- for any blanks below
            end
            break
          end
        end
      else
        folds[start_line + i - 1] = current_level + FOLD_BLANK
      end
    end
  else
    -- No folding, reset fold levels if necessary.
    local current_line = start_line
    for _ in text:gmatch('\r?\n') do
      folds[current_line] = start_level
      current_line = current_line + 1
    end
  end
  return folds
end
--- Creates a returns a new lexer with the given name.
-- @param name The lexer's name.
-- @param opts Table of lexer options. Options currently supported:
--   - `lex_by_line`: Whether or not the lexer only processes whole lines of text (instead of
--     arbitrary chunks of text) at a time. Line lexers cannot look ahead to subsequent lines.
--     The default value is `false`.
--   - `fold_by_indentation`: Whether or not the lexer does not define any fold points and that
--     fold points should be calculated based on changes in line indentation. The default value
--     is `false`.
--   - `case_insensitive_fold_points`: Whether or not fold points added via
--     `lexer.add_fold_point()` ignore case. The default value is `false`.
--   - `no_user_word_lists`: Does not automatically allocate word lists that can be set by
--     users. This should really only be set by non-programming languages like markup languages.
--   - `inherit`: Lexer to inherit from. The default value is `nil`.
-- @usage lexer.new('rhtml', {inherit = lexer.load('html')})
function M.new(name, opts)
  local lexer = setmetatable({
    _name = assert(name, 'lexer name expected'), _lex_by_line = opts and opts['lex_by_line'],
    _fold_by_indentation = opts and opts['fold_by_indentation'],
    _case_insensitive_fold_points = opts and opts['case_insensitive_fold_points'],
    _no_user_word_lists = opts and opts['no_user_word_lists'], _lexer = opts and opts['inherit']
  }, {
    __index = {
      tag = M.tag, word_match = M.word_match, set_word_list = M.set_word_list,
      add_rule = M.add_rule, modify_rule = M.modify_rule, get_rule = M.get_rule,
      add_fold_point = M.add_fold_point, embed = M.embed, lex = M.lex, fold = M.fold, --
      add_style = function() end -- legacy
    }
  })
  -- Add initial whitespace rule.
  -- Use a unique whitespace tag name since embedded lexing relies on these unique names.
  lexer:add_rule('whitespace', lexer:tag('whitespace.' .. name, M.space^1))
  return lexer
end
--- Creates a substitute for some Scintilla tables and functions that Scintillua depends on
-- when using it as a standalone module.
local function initialize_standalone_library()
  M.property = setmetatable({['scintillua.lexers'] = package.path:gsub('/%?%.lua', '/lexers')}, {
    __index = function() return '' end, __newindex = function(t, k, v) rawset(t, k, tostring(v)) end
  })
  M.line_from_position = function(pos)
    local line = 1
    for s in M._text:gmatch('[^\n]*()') do
      if pos <= s then return line end
      line = line + 1
    end
    return line - 1 -- should not get to here
  end
  M.indent_amount = setmetatable({}, {
    __index = function(_, line)
      local current_line = 1
      for s in M._text:gmatch('()[^\n]*') do
        if current_line == line then
          return #M._text:match('^[ \t]*', s):gsub('\t', string.rep(' ', 8))
        end
        current_line = current_line + 1
      end
    end
  })
  M._standalone = true
end
--- Searches for the given *name* in the given *path*.
-- This is a safe implementation of Lua 5.2's `package.searchpath()` function that does not
-- require the package module to be loaded.
local function searchpath(name, path)
  local tried = {}
  for part in path:gmatch('[^;]+') do
    local filename = part:gsub('%?', name)
    local ok, errmsg = loadfile(filename)
    if ok or not errmsg:find('cannot open') then return filename end
    tried[#tried + 1] = string.format("no file '%s'", filename)
  end
  return nil, table.concat(tried, '\n')
end
--- Initializes or loads and then returns the lexer of string name *name*.
-- Scintilla calls this function in order to load a lexer. Parent lexers also call this function
-- in order to load child lexers and vice-versa. The user calls this function in order to load
-- a lexer when using Scintillua as a Lua library.
-- @param name The name of the lexing language.
-- @param[opt] alt_name Optional alternate name of the lexing language. This is useful for
--   embedding the same child lexer with multiple sets of start and end tags.
-- @return lexer object
function M.load(name, alt_name)
  assert(name, 'no lexer given')
  if not M.property then initialize_standalone_library() end
  if not M.property_int then
    -- Separate from initialize_standalone_library() so applications that choose to define
    -- M.property do not also have to define this.
    M.property_int = setmetatable({}, {
      __index = function(t, k) return tonumber(M.property[k]) or 0 end,
      __newindex = function() error('read-only property') end
    })
  end
  -- Load the language lexer with its rules, tags, etc.
  local path = M.property['scintillua.lexers']:gsub(';', '/?.lua;') .. '/?.lua'
  local ro_lexer = setmetatable({
    WHITESPACE = 'whitespace.' .. (alt_name or name) -- legacy
  }, {__index = M})
  local env = {
    'assert', 'error', 'ipairs', 'math', 'next', 'pairs', 'print', 'select', 'string', 'table',
    'tonumber', 'tostring', 'type', 'utf8', '_VERSION', lexer = ro_lexer, lpeg = lpeg, --
    require = function() return ro_lexer end -- legacy
  }
  for _, name in ipairs(env) do env[name] = _G[name] end
  local lexer = assert(loadfile(assert(searchpath(name, path)), 't', env))(alt_name or name)
  assert(lexer, string.format("'%s.lua' did not return a lexer", name))
  -- If the lexer is a proxy or a child that embedded itself, set the parent to be the main
  -- lexer. Keep a reference to the old parent name since embedded child start and end rules
  -- reference and use that name.
  if lexer._lexer then
    lexer = lexer._lexer
    lexer._parent_name, lexer._name = lexer._name, alt_name or name
  end
  M.property['scintillua.comment.' .. (alt_name or name)] = M.property['scintillua.comment']
  return lexer
end
--- Returns a list of all known lexer names.
-- This function is not available to lexers and requires the LuaFileSystem (`lfs`) module to
-- be available.
-- @param[opt] path Optional ';'-delimited list of directories to search for lexers in. The
--   default value is Scintillua's configured lexer path.
-- @return lexer name list
function M.names(path)
  local lfs = require('lfs')
  if not path then path = M.property and M.property['scintillua.lexers'] end
  if not path or path == '' then
    for part in package.path:gmatch('[^;]+') do
      local dir = part:match('^(.-[/\\]?lexers)[/\\]%?%.lua$')
      if dir then
        path = dir
        break
      end
    end
  end
  local lexers = {}
  for dir in assert(path, 'lexer path not configured or found'):gmatch('[^;]+') do
    if lfs.attributes(dir, 'mode') == 'directory' then
      for file in lfs.dir(dir) do
        local name = file:match('^(.+)%.lua$')
        if name and name ~= 'lexer' and not lexers[name] then
          lexers[#lexers + 1], lexers[name] = name, true
        end
      end
    end
  end
  table.sort(lexers)
  return lexers
end
--- Map of file extensions, without the '.' prefix, to their associated lexer names.
-- This map has precedence over Scintillua's built-in map.
-- @see detect
M.detect_extensions = {}
--- Map of line patterns to their associated lexer names.
-- These are Lua string patterns, not LPeg patterns.
-- This map has precedence over Scintillua's built-in map.
-- @see detect
M.detect_patterns = {}
--- Returns the name of the lexer often associated with filename *filename* and/or content
-- line *line*.
-- @param[opt] filename Optional string filename. The default value is read from the
--   'lexer.scintillua.filename' property.
-- @param[opt] line Optional string first content line, such as a shebang line. The default
--   value is read from the 'lexer.scintillua.line' property.
-- @return string lexer name to pass to `load()`, or `nil` if none was detected
-- @see detect_extensions
-- @see detect_patterns
function M.detect(filename, line)
  if not filename then filename = M.property and M.property['lexer.scintillua.filename'] or '' end
  if not line then line = M.property and M.property['lexer.scintillua.line'] or '' end
  -- Locally scoped in order to avoid persistence in memory.
  local extensions = {
    as = 'actionscript', asc = 'actionscript', --
    adb = 'ada', ads = 'ada', --
    g = 'antlr', g4 = 'antlr', --
    ans = 'apdl', inp = 'apdl', mac = 'apdl', --
    apl = 'apl', --
    applescript = 'applescript', --
    asm = 'asm', ASM = 'asm', s = 'asm', S = 'asm', --
    asa = 'asp', asp = 'asp', hta = 'asp', --
    ahk = 'autohotkey', --
    au3 = 'autoit', a3x = 'autoit', --
    awk = 'awk', --
    bat = 'batch', cmd = 'batch', --
    bib = 'bibtex', --
    boo = 'boo', --
    cs = 'csharp', --
    c = 'ansi_c', C = 'ansi_c', cc = 'cpp', cpp = 'cpp', cxx = 'cpp', ['c++'] = 'cpp', h = 'cpp',
    hh = 'cpp', hpp = 'cpp', hxx = 'cpp', ['h++'] = 'cpp', --
    ck = 'chuck', --
    clj = 'clojure', cljs = 'clojure', cljc = 'clojure', edn = 'clojure', --
    ['CMakeLists.txt'] = 'cmake', cmake = 'cmake', ['cmake.in'] = 'cmake', ctest = 'cmake',
    ['ctest.in'] = 'cmake', --
    coffee = 'coffeescript', --
    cr = 'crystal', --
    css = 'css', --
    cu = 'cuda', cuh = 'cuda', --
    d = 'dmd', di = 'dmd', --
    dart = 'dart', --
    desktop = 'desktop', --
    diff = 'diff', patch = 'diff', --
    Dockerfile = 'dockerfile', --
    dot = 'dot', --
    e = 'eiffel', eif = 'eiffel', --
    ex = 'elixir', exs = 'elixir', --
    elm = 'elm', --
    erl = 'erlang', hrl = 'erlang', --
    fs = 'fsharp', --
    fan = 'fantom', --
    dsp = 'faust', --
    fnl = 'fennel', --
    fish = 'fish', --
    forth = 'forth', frt = 'forth', --
    f = 'fortran', ['for'] = 'fortran', ftn = 'fortran', fpp = 'fortran', f77 = 'fortran',
    f90 = 'fortran', f95 = 'fortran', f03 = 'fortran', f08 = 'fortran', --
    fstab = 'fstab', --
    gd = 'gap', gi = 'gap', gap = 'gap', --
    gmi = 'gemini', --
    po = 'gettext', pot = 'gettext', --
    feature = 'gherkin', --
    gleam = 'gleam', --
    glslf = 'glsl', glslv = 'glsl', --
    dem = 'gnuplot', plt = 'gnuplot', --
    go = 'go', --
    groovy = 'groovy', gvy = 'groovy', --
    gtkrc = 'gtkrc', --
    ha = 'hare', --
    hs = 'haskell', --
    htm = 'html', html = 'html', shtm = 'html', shtml = 'html', xhtml = 'html', vue = 'html', --
    icn = 'icon', --
    idl = 'idl', odl = 'idl', --
    ni = 'inform', --
    cfg = 'ini', cnf = 'ini', inf = 'ini', ini = 'ini', reg = 'ini', --
    io = 'io_lang', --
    bsh = 'java', java = 'java', --
    js = 'javascript', jsfl = 'javascript', --
    jq = 'jq', --
    json = 'json', --
    jsp = 'jsp', --
    jl = 'julia', --
    bbl = 'latex', dtx = 'latex', ins = 'latex', ltx = 'latex', tex = 'latex', sty = 'latex', --
    ledger = 'ledger', journal = 'ledger', --
    less = 'less', --
    lily = 'lilypond', ly = 'lilypond', --
    cl = 'lisp', el = 'lisp', lisp = 'lisp', lsp = 'lisp', --
    litcoffee = 'litcoffee', --
    lgt = 'logtalk', --
    lua = 'lua', --
    GNUmakefile = 'makefile', iface = 'makefile', mak = 'makefile', makefile = 'makefile',
    Makefile = 'makefile', --
    md = 'markdown', --
    ['meson.build'] = 'meson', --
    moon = 'moonscript', --
    myr = 'myrddin', --
    n = 'nemerle', --
    link = 'networkd', network = 'networkd', netdev = 'networkd', --
    nim = 'nim', --
    nsh = 'nsis', nsi = 'nsis', nsis = 'nsis', --
    obs = 'objeck', --
    m = 'objective_c', mm = 'objective_c', objc = 'objective_c', --
    caml = 'caml', ml = 'caml', mli = 'caml', mll = 'caml', mly = 'caml', --
    dpk = 'pascal', dpr = 'pascal', p = 'pascal', pas = 'pascal', --
    al = 'perl', perl = 'perl', pl = 'perl', pm = 'perl', pod = 'perl', --
    inc = 'php', php = 'php', php3 = 'php', php4 = 'php', phtml = 'php', --
    p8 = 'pico8', --
    pike = 'pike', pmod = 'pike', --
    PKGBUILD = 'pkgbuild', --
    pony = 'pony', --
    eps = 'ps', ps = 'ps', --
    ps1 = 'powershell', --
    prolog = 'prolog', --
    props = 'props', properties = 'props', --
    proto = 'protobuf', --
    pure = 'pure', --
    sc = 'python', py = 'python', pyw = 'python', --
    R = 'rstats', Rout = 'rstats', Rhistory = 'rstats', Rt = 'rstats', ['Rout.save'] = 'rstats',
    ['Rout.fail'] = 'rstats', --
    re = 'reason', --
    r = 'rebol', reb = 'rebol', --
    rst = 'rest', --
    orx = 'rexx', rex = 'rexx', --
    erb = 'rhtml', rhtml = 'rhtml', --
    rsc = 'routeros', --
    spec = 'rpmspec', --
    Rakefile = 'ruby', rake = 'ruby', rb = 'ruby', rbw = 'ruby', --
    rs = 'rust', --
    sass = 'sass', scss = 'sass', --
    scala = 'scala', --
    sch = 'scheme', scm = 'scheme', --
    bash = 'bash', bashrc = 'bash', bash_profile = 'bash', configure = 'bash', csh = 'bash',
    ksh = 'bash', mksh = 'bash', sh = 'bash', zsh = 'bash', --
    changes = 'smalltalk', st = 'smalltalk', sources = 'smalltalk', --
    sml = 'sml', fun = 'sml', sig = 'sml', --
    sno = 'snobol4', SNO = 'snobol4', --
    spin = 'spin', --
    ddl = 'sql', sql = 'sql', --
    automount = 'systemd', device = 'systemd', mount = 'systemd', path = 'systemd',
    scope = 'systemd', service = 'systemd', slice = 'systemd', socket = 'systemd', swap = 'systemd',
    target = 'systemd', timer = 'systemd', --
    taskpaper = 'taskpaper', --
    tcl = 'tcl', tk = 'tcl', --
    texi = 'texinfo', --
    toml = 'toml', --
    ['1'] = 'troff', ['2'] = 'troff', ['3'] = 'troff', ['4'] = 'troff', ['5'] = 'troff',
    ['6'] = 'troff', ['7'] = 'troff', ['8'] = 'troff', ['9'] = 'troff', ['1x'] = 'troff',
    ['2x'] = 'troff', ['3x'] = 'troff', ['4x'] = 'troff', ['5x'] = 'troff', ['6x'] = 'troff',
    ['7x'] = 'troff', ['8x'] = 'troff', ['9x'] = 'troff', --
    t2t = 'txt2tags', --
    ts = 'typescript', --
    vala = 'vala', --
    vcf = 'vcard', vcard = 'vcard', --
    v = 'verilog', ver = 'verilog', --
    vh = 'vhdl', vhd = 'vhdl', vhdl = 'vhdl', --
    bas = 'vb', cls = 'vb', ctl = 'vb', dob = 'vb', dsm = 'vb', dsr = 'vb', frm = 'vb', pag = 'vb',
    vb = 'vb', vba = 'vb', vbs = 'vb', --
    wsf = 'wsf', --
    dtd = 'xml', svg = 'xml', xml = 'xml', xsd = 'xml', xsl = 'xml', xslt = 'xml', xul = 'xml', --
    xs = 'xs', xsin = 'xs', xsrc = 'xs', --
    xtend = 'xtend', --
    yaml = 'yaml', yml = 'yaml', --
    zig = 'zig'
  }
  local patterns = {
    ['^#!.+[/ ][gm]?awk'] = 'awk', ['^#!.+[/ ]lua'] = 'lua', ['^#!.+[/ ]octave'] = 'matlab',
    ['^#!.+[/ ]perl'] = 'perl', ['^#!.+[/ ]php'] = 'php', ['^#!.+[/ ]python'] = 'python',
    ['^#!.+[/ ]ruby'] = 'ruby', ['^#!.+[/ ]bash'] = 'bash', ['^#!.+/m?ksh'] = 'bash',
    ['^#!.+/sh'] = 'bash', ['^%s*class%s+%S+%s*<%s*ApplicationController'] = 'rails',
    ['^%s*class%s+%S+%s*<%s*ActionController::Base'] = 'rails',
    ['^%s*class%s+%S+%s*<%s*ActiveRecord::Base'] = 'rails',
    ['^%s*class%s+%S+%s*<%s*ActiveRecord::Migration'] = 'rails', ['^%s*<%?xml%s'] = 'xml',
    ['^#cloud%-config'] = 'yaml'
  }
  for patt, name in pairs(M.detect_patterns) do if line:find(patt) then return name end end
  for patt, name in pairs(patterns) do if line:find(patt) then return name end end
  local name, ext = filename:match('[^/\\]+$'), filename:match('[^.]*$')
  return M.detect_extensions[name] or extensions[name] or M.detect_extensions[ext] or
    extensions[ext]
end
-- The following are utility functions lexers will have access to.
-- Common patterns.
--- A pattern that matches any single character.
M.any = P(1)
--- A pattern that matches any alphabetic character ('A'-'Z', 'a'-'z').
M.alpha = R('AZ', 'az')
--- A pattern that matches any digit ('0'-'9').
M.digit = R('09')
--- A pattern that matches any alphanumeric character ('A'-'Z', 'a'-'z', '0'-'9').
M.alnum = R('AZ', 'az', '09')
--- A pattern that matches any lower case character ('a'-'z').
M.lower = R('az')
--- A pattern that matches any upper case character ('A'-'Z').
M.upper = R('AZ')
--- A pattern that matches any hexadecimal digit ('0'-'9', 'A'-'F', 'a'-'f').
M.xdigit = R('09', 'AF', 'af')
--- A pattern that matches any graphical character ('!' to '~').
M.graph = R('!~')
--- A pattern that matches any punctuation character ('!' to '/', ':' to '@', '[' to ''', '{'
-- to '~').
M.punct = R('!/', ':@', '[\'', '{~')
--- A pattern that matches any whitespace character ('\t', '\v', '\f', '\n', '\r', space).
M.space = S('\t\v\f\n\r ')
--- A pattern that matches a sequence of end of line characters.
M.newline = P('\r')^-1 * '\n'
--- A pattern that matches any single, non-newline character.
M.nonnewline = 1 - M.newline
--- Returns a pattern that matches a decimal number, whose digits may be separated by character
-- *c*.
function M.dec_num_(c) return M.digit * (P(c)^-1 * M.digit)^0 end
--- Returns a pattern that matches a hexadecimal number, whose digits may be separated by
-- character *c*.
function M.hex_num_(c) return '0' * S('xX') * (P(c)^-1 * M.xdigit)^1 end
--- Returns a pattern that matches an octal number, whose digits may be separated by character *c*.
function M.oct_num_(c) return '0' * (P(c)^-1 * R('07'))^1 * -M.xdigit end
--- Returns a pattern that matches a binary number, whose digits may be separated by character *c*.
function M.bin_num_(c) return '0' * S('bB') * (P(c)^-1 * S('01'))^1 * -M.xdigit end
--- Returns a pattern that matches either a decimal, hexadecimal, octal, or binary number,
-- whose digits may be separated by character *c*.
function M.integer_(c)
  return S('+-')^-1 * (M.hex_num_(c) + M.bin_num_(c) + M.oct_num_(c) + M.dec_num_(c))
end
local function exp_(c) return S('eE') * S('+-')^-1 * M.digit * (P(c)^-1 * M.digit)^0 end
--- Returns a pattern that matches a floating point number, whose digits may be separated by
-- character *c*.
function M.float_(c)
  return S('+-')^-1 *
    ((M.dec_num_(c)^-1 * '.' * M.dec_num_(c) + M.dec_num_(c) * '.' * M.dec_num_(c)^-1 * -P('.')) *
      exp_(c)^-1 + (M.dec_num_(c) * exp_(c)))
end
--- Returns a pattern that matches a typical number, either a floating point, decimal, hexadecimal,
-- octal, or binary number, and whose digits may be separated by character *c*.
function M.number_(c) return M.float_(c) + M.integer_(c) end
--- A pattern that matches a decimal number.
M.dec_num = M.dec_num_(false)
--- A pattern that matches a hexadecimal number.
M.hex_num = M.hex_num_(false)
--- A pattern that matches an octal number.
M.oct_num = M.oct_num_(false)
--- A pattern that matches a binary number.
M.bin_num = M.bin_num_(false)
--- A pattern that matches either a decimal, hexadecimal, octal, or binary number.
M.integer = M.integer_(false)
--- A pattern that matches a floating point number.
M.float = M.float_(false)
--- A pattern that matches a typical number, either a floating point, decimal, hexadecimal,
-- octal, or binary number.
M.number = M.number_(false)
--- A pattern that matches a typical word. Words begin with a letter or underscore and consist
-- of alphanumeric and underscore characters.
M.word = (M.alpha + '_') * (M.alnum + '_')^0
--- Creates and returns a pattern that matches from string or pattern *prefix* until the end of
-- the line.
-- *escape* indicates whether the end of the line can be escaped with a '\' character.
-- @param[opt] prefix Optional string or pattern prefix to start matching at. The default value
--   is any non-newline character.
-- @param[opt] escape Optional flag indicating whether or not newlines can be escaped by a '\'
--  character. The default value is `false`.
-- @return pattern
-- @usage local line_comment = lexer.to_eol('//')
-- @usage local line_comment = lexer.to_eol(S('#;'))
function M.to_eol(prefix, escape)
  return (prefix or M.nonnewline) *
    (not escape and M.nonnewline or 1 - (M.newline + '\\') + '\\' * M.any)^0
end
--- Creates and returns a pattern that matches a range of text bounded by strings or patterns *s*
-- and *e*.
-- This is a convenience function for matching more complicated ranges like strings with escape
-- characters, balanced parentheses, and block comments (nested or not). *e* is optional and
-- defaults to *s*. *single_line* indicates whether or not the range must be on a single line;
-- *escapes* indicates whether or not to allow '\' as an escape character; and *balanced*
-- indicates whether or not to handle balanced ranges like parentheses, and requires *s* and *e*
-- to be different.
-- @param s String or pattern start of a range.
-- @param[opt] e Optional string or pattern end of a range. The default value is *s*.
-- @param[opt] single_line Optional flag indicating whether or not the range must be on a single
--   line. The default value is `false`.
-- @param[opt] escapes Optional flag indicating whether or not the range end may be escaped
--   by a '\' character. The default value is `false` unless *s* and *e* are identical,
--   single-character strings. In that case, the default value is `true`.
-- @param[opt] balanced Optional flag indicating whether or not to match a balanced range,
--   like the "%b" Lua pattern. This flag only applies if *s* and *e* are different.
-- @return pattern
-- @usage local dq_str_escapes = lexer.range('"')
-- @usage local dq_str_noescapes = lexer.range('"', false, false)
-- @usage local unbalanced_parens = lexer.range('(', ')')
-- @usage local balanced_parens = lexer.range('(', ')', false, false, true)
function M.range(s, e, single_line, escapes, balanced)
  if type(e) ~= 'string' and type(e) ~= 'userdata' then
    e, single_line, escapes, balanced = s, e, single_line, escapes
  end
  local any = M.any - e
  if single_line then any = any - '\n' end
  if balanced then any = any - s end
  -- Only allow escapes by default for ranges with identical, single-character string delimiters.
  if escapes == nil then escapes = type(s) == 'string' and #s == 1 and s == e end
  if escapes then any = any - '\\' + '\\' * M.any end
  if balanced and s ~= e then return P{s * (any + V(1))^0 * P(e)^-1} end
  return s * any^0 * P(e)^-1
end
--- Creates and returns a pattern that matches pattern *patt* only when it comes after one of
-- the characters in string *set* (or when there are no characters behind *patt*), skipping
-- over any characters in string *skip*, which is whitespace by default.
-- @param set String character set like one passed to `lpeg.S()`.
-- @param patt The LPeg pattern to match after a set character.
-- @param skip String character set to skip over. The default value is ' \t\r\n\v\f' (whitespace).
-- @usage local regex = lexer.after_set('+-*!%^&|=,([{', lexer.range('/'))
function M.after_set(set, patt, skip)
  if not skip then skip = ' \t\r\n\v\f' end
  local set_chars, skip_chars = {}, {}
  -- Note: cannot use utf8.codes() because Lua 5.1 is still supported.
  for char in set:gmatch('.') do set_chars[string.byte(char)] = true end
  for char in skip:gmatch('.') do skip_chars[string.byte(char)] = true end
  return (B(S(set)) + -B(1)) * patt + Cmt(C(patt), function(input, index, match, ...)
    local pos = index - #match
    if #skip > 0 then while pos > 1 and skip_chars[input:byte(pos - 1)] do pos = pos - 1 end end
    if pos == 1 or set_chars[input:byte(pos - 1)] then return index, ... end
    return nil
  end)
end
--- Creates and returns a pattern that matches pattern *patt* only at the beginning of a line,
-- or after any line indentation if *allow_indent* is `true`.
-- @param patt The LPeg pattern to match on the beginning of a line.
-- @param allow_indent Whether or not to consider line indentation as the start of a line. The
--   default value is `false`.
-- @return pattern
-- @usage local preproc = lex:tag(lexer.PREPROCESSOR, lexer.starts_line(lexer.to_eol('#')))
function M.starts_line(patt, allow_indent)
  return M.after_set('\r\n\v\f', patt, allow_indent and ' \t' or '')
end
M.colors = {} -- legacy
M.styles = setmetatable({}, { -- legacy
  __index = function() return setmetatable({}, {__concat = function() return nil end}) end,
  __newindex = function() end
})
M.property_expanded = setmetatable({}, {__index = function() return '' end}) -- legacy
-- Legacy function for creates and returns a token pattern with token name *name* and pattern
-- *patt*.
-- Use `tag()` instead.
-- @param name The name of token.
-- @param patt The LPeg pattern associated with the token.
-- @return pattern
-- @usage local number = token(lexer.NUMBER, lexer.number)
-- @usage local addition = token('addition', '+' * lexer.word)
function M.token(name, patt) return Cc(name) * (P(patt) / 0) * Cp() end
-- Legacy function that creates and returns a pattern that verifies the first non-whitespace
-- character behind the current match position is in string set *s*.
-- @param s String character set like one passed to `lpeg.S()`.
-- @return pattern
-- @usage local regex = #P('/') * lexer.last_char_includes('+-*!%^&|=,([{') * lexer.range('/')
function M.last_char_includes(s) return M.after_set(s, true) end
function M.fold_consecutive_lines() end -- legacy
-- The functions and fields below were defined in C.
--- Table of fold level bit-masks for line numbers starting from 1. (Read-only)
-- Fold level masks are composed of an integer level combined with any of the following bits:
--
--   - `lexer.FOLD_BASE`
--     The initial fold level.
--   - `lexer.FOLD_BLANK`
--     The line is blank.
--   - `lexer.FOLD_HEADER`
--     The line is a header, or fold point.
-- @table fold_level
--- Table of indentation amounts in character columns, for line numbers starting from
-- 1. (Read-only)
-- @table indent_amount
--- Table of integer line states for line numbers starting from 1.
-- Line states can be used by lexers for keeping track of persistent states. For example,
-- the output lexer uses this to mark lines that have warnings or errors.
-- @table line_state
--- Map of key-value string pairs.
-- @table property
--- Map of key-value pairs with values interpreted as numbers, or `0` if not found. (Read-only)
-- @table property_int
--- Table of style names at positions in the buffer starting from 1. (Read-only)
-- @table style_at
--- Returns the line number (starting from 1) of the line that contains position *pos*, which
-- starts from 1.
-- @param pos The position to get the line number of.
-- @return number
-- @function line_from_position
return M