Adapter adapts Caddyfile to Caddy JSON.

Dispenser is a type that dispenses tokens, similarly to a lexer, except that it can do so with some notion of structure. An empty Dispenser is invalid; call NewDispenser to make a proper instance.

Segment is a list of tokens which begins with a directive and ends at the end of the directive (either at the end of the line, or at the end of a block it opens).

ServerBlock associates any number of keys from the head of the server block with tokens, which are grouped by segments.

ServerType is a type that can evaluate a Caddyfile and set up a caddy config.

Unmarshaler is a type that can unmarshal Caddyfile tokens to set itself up for a JSON encoding. The goal of an unmarshaler is not to set itself up for actual use, but to set itself up for being marshaled into JSON. Caddyfile-unmarshaled values will not be used directly; they will be encoded as JSON and then used from that. Implementations may be able to support multiple segments (instances of their directive or batch of tokens); typically this means wrapping parsing logic in a loop: for d.Next() { ... }. More commonly, only a single segment is supported, so a simple d.Next() at the start should be used to consume the module identifier token (directive name, etc).

This type doesn't have documentation.

This type doesn't have documentation.

This type doesn't have documentation.

This type doesn't have documentation.

Format formats the input Caddyfile to a standard, nice-looking appearance. It works by reading each rune of the input and taking control over all the bracing and whitespace that is written; otherwise, words, comments, placeholders, and escaped characters are all treated literally and written as they appear in the input.

FormattingDifference returns a warning and true if the formatted version is any different from the input; empty warning and false otherwise. TODO: also perform this check on imported files

NewDispenser returns a Dispenser filled with the given tokens.

NewTestDispenser parses input into tokens and creates a new Dispenser for test purposes only; any errors are fatal.

Parse parses the input just enough to group tokens, in order, by server block. No further parsing is performed. Server blocks are returned in the order in which they appear. Directives that do not appear in validDirectives will cause an error. If you do not want to check for valid directives, pass in nil instead.

Environment variables in {$ENVIRONMENT_VARIABLE} notation will be replaced before parsing begins.

Tokenize takes bytes as input and lexes it into a list of tokens that can be parsed as a Caddyfile. Also takes a filename to fill the token's File as the source of the tokens, which is important to determine relative paths for import directives.

UnmarshalModule instantiates a module with the given ID and invokes UnmarshalCaddyfile on the new value using the immediate next segment of d as input. In other words, d's next token should be the first token of the module's Caddyfile input.

This function is used when the next segment of Caddyfile tokens belongs to another Caddy module. The returned value is often type-asserted to the module's associated type for practical use when setting up a config.

AllArgs is like Args, but if there are more argument tokens available than there are targets, false is returned. The number of available argument tokens must match the number of targets exactly to return true.

ArgErr returns an argument error, meaning that another argument was expected but not found. In other words, a line break or open curly brace was encountered instead of an argument.

Args is a convenience function that loads the next arguments (tokens on the same line) into an arbitrary number of strings pointed to in targets. If there are not enough argument tokens available to fill targets, false is returned and the remaining targets are left unchanged. If all the targets are filled, then true is returned.

CountRemainingArgs counts the amount of remaining arguments (tokens on the same line) without consuming the tokens.

Delete deletes the current token and returns the updated slice of tokens. The cursor is not advanced to the next token. Because deletion modifies the underlying slice, this method should only be called if you have access to the original slice of tokens and/or are using the slice of tokens outside this Dispenser instance. If you do not re-assign the slice with the return value of this method, inconsistencies in the token array will become apparent (or worse, hide from you like they did me for 3 and a half freaking hours late one night).

DeleteN is the same as Delete, but can delete many tokens at once. If there aren't N tokens available to delete, none are deleted.

EOFErr returns an error indicating that the dispenser reached the end of the input when searching for the next token.

Err generates a custom parse-time error with a message of msg.

Errf is like Err, but for formatted error messages

File gets the filename where the current token originated.

GetContext gets the value of a key in the context map.

GetContextString gets the value of a key in the context map as a string, or an empty string if the key does not exist.

Line gets the line number of the current token. If there is no token loaded, it returns 0.

Nesting returns the current nesting level. Necessary if using NextBlock()

NewFromNextSegment returns a new dispenser with a copy of the tokens from the current token until the end of the "directive" whether that be to the end of the line or the end of a block that starts at the end of the line; in other words, until the end of the segment.

Next loads the next token. Returns true if a token was loaded; false otherwise. If false, all tokens have been consumed.

NextArg loads the next token if it is on the same line and if it is not a block opening (open curly brace). Returns true if an argument token was loaded; false otherwise. If false, all tokens on the line have been consumed except for potentially a block opening. It handles imported tokens correctly.

NextBlock can be used as the condition of a for loop to load the next token as long as it opens a block or is already in a block nested more than initialNestingLevel. In other words, a loop over NextBlock() will iterate all tokens in the block assuming the next token is an open curly brace, until the matching closing brace. The open and closing brace tokens for the outer-most block will be consumed internally and omitted from the iteration.

Proper use of this method looks like this:

for nesting := d.Nesting(); d.NextBlock(nesting); {
}

However, in simple cases where it is known that the Dispenser is new and has not already traversed state by a loop over NextBlock(), this will do:

for d.NextBlock(0) {
}

As with other token parsing logic, a loop over NextBlock() should be contained within a loop over Next(), as it is usually prudent to skip the initial token.

NextLine loads the next token only if it is not on the same line as the current token, and returns true if a token was loaded; false otherwise. If false, there is not another token or it is on the same line. It handles imported tokens correctly.

NextSegment returns a copy of the tokens from the current token until the end of the line or block that starts at the end of the line.

Prev moves to the previous token. It does the inverse of Next(), except this function may decrement the cursor to -1 so that the next call to Next() points to the first token; this allows dispensing to "start over". This method returns true if the cursor ends up pointing to a valid token.

RemainingArgs loads any more arguments (tokens on the same line) into a slice and returns them. Open curly brace tokens also indicate the end of arguments, and the curly brace is not included in the return value nor is it loaded.

RemainingArgsRaw loads any more arguments (tokens on the same line, retaining quotes) into a slice and returns them. Open curly brace tokens also indicate the end of arguments, and the curly brace is not included in the return value nor is it loaded.

Reset sets d's cursor to the beginning, as if this was a new and unused dispenser.

ScalarVal gets value of the current token, converted to the closest scalar type. If there is no token loaded, it returns nil.

SetContext sets a key-value pair in the context map.

SyntaxErr creates a generic syntax error which explains what was found and what was expected.

Token returns the current token.

Val gets the text of the current token. If there is no token loaded, it returns empty string.

ValRaw gets the raw text of the current token (including quotes). If the token was a heredoc, then the delimiter is not included, because that is not relevant to any unmarshaling logic at this time. If there is no token loaded, it returns empty string.

WrapErr takes an existing error and adds the Caddyfile file and line number.

Adapt converts the Caddyfile config in body to Caddy JSON.

Directive returns the directive name for the segment. The directive name is the text of the first token.

DispenseDirective returns a dispenser that contains all the tokens in the server block.

Clone returns a deep copy of the token.

NumLineBreaks counts how many line breaks are in the token text.

Quoted returns true if the token was enclosed in quotes (i.e. double quotes, backticks, or heredoc).