This is a reference manual for the CUE data constraint language. CUE, pronounced cue or Q, is a general-purpose and strongly typed constraint-based language. It can be used for data templating, data validation, code generation, scripting, and many other applications involving structured data. The CUE tooling, layered on top of CUE, provides a general purpose scripting language for creating scripts as well as simple servers, also expressed in CUE.
CUE was designed with cloud configuration, and related systems, in mind, but is not limited to this domain. It derives its formalism from relational programming languages. This formalism allows for managing and reasoning over large amounts of data in a straightforward manner.
The grammar is compact and regular, allowing for easy analysis by automatic tools such as integrated development environments.
This document is maintained by mpvl@golang.org. CUE has a lot of similarities with the Go language. This document draws heavily from the Go specification as a result.
CUE draws its influence from many languages. Its main influences were BCL/ GCL (internal to Google), LKB (LinGO), Go, and JSON. Others are Swift, Typescript, Javascript, Prolog, NCL (internal to Google), Jsonnet, HCL, Flabbergast, Nix, JSONPath, Haskell, Objective-C, and Python.
The syntax is specified using Extended Backus-Naur Form (EBNF):
Production = production_name "=" [ Expression ] "." .
Expression = Alternative { "|" Alternative } .
Alternative = Term { Term } .
Term = production_name | token [ "…" token ] | Group | Option | Repetition .
Group = "(" Expression ")" .
Option = "[" Expression "]" .
Repetition = "{" Expression "}" .
Productions are expressions constructed from terms and the following operators, in increasing precedence:
| alternation
() grouping
[] option (0 or 1 times)
{} repetition (0 to n times)
Lower-case production names are used to identify lexical tokens. Non-terminals are in CamelCase. Lexical tokens are enclosed in double quotes "" or back quotes ``.
The form a … b represents the set of characters from a through b as alternatives. The horizontal ellipsis … is also used elsewhere in the spec to informally denote various enumerations or code snippets that are not further specified. The character … (as opposed to the three characters ...) is not a token of the CUE language.
Source code is Unicode text encoded in UTF-8. Unless otherwise noted, the text is not canonicalized, so a single accented code point is distinct from the same character constructed from combining an accent and a letter; those are treated as two code points. For simplicity, this document will use the unqualified term character to refer to a Unicode code point in the source text.
Each code point is distinct; for instance, upper and lower case letters are different characters.
Implementation restriction: For compatibility with other tools, a compiler may disallow the NUL character (U+0000) in the source text.
Implementation restriction: For compatibility with other tools, a compiler may ignore a UTF-8-encoded byte order mark (U+FEFF) if it is the first Unicode code point in the source text. A byte order mark may be disallowed anywhere else in the source.
The following terms are used to denote specific Unicode character classes:
newline = /* the Unicode code point U+000A */ . unicode_char = /* an arbitrary Unicode code point except newline */ . unicode_letter = /* a Unicode code point classified as "Letter" */ . unicode_digit = /* a Unicode code point classified as "Number, decimal digit" */ .
In The Unicode Standard 8.0, Section 4.5 “General Category” defines a set of character categories. CUE treats all characters in any of the Letter categories Lu, Ll, Lt, Lm, or Lo as Unicode letters, and those in the Number category Nd as Unicode digits.
The underscore character _ (U+005F) is considered a letter.
letter = unicode_letter | "_" . decimal_digit = "0" … "9" . octal_digit = "0" … "7" . hex_digit = "0" … "9" | "A" … "F" | "a" … "f" .
Comments serve as program documentation. CUE supports line comments that start with the character sequence // and stop at the end of the line.
A comment cannot start inside a string literal or inside a comment. A comment acts like a newline.
Tokens form the vocabulary of the CUE language. There are four classes: identifiers, keywords, operators and punctuation, and literals. White space, formed from spaces (U+0020), horizontal tabs (U+0009), carriage returns (U+000D), and newlines (U+000A), is ignored except as it separates tokens that would otherwise combine into a single token. Also, a newline or end of file may trigger the insertion of a comma. While breaking the input into tokens, the next token is the longest sequence of characters that form a valid token.
The formal grammar uses commas “,” as terminators in a number of productions. CUE programs may omit most of these commas using the following two rules:
When the input is broken into tokens, a comma is automatically inserted into the token stream immediately after a line's final token if that token is
), ], }, or ?Although commas are automatically inserted, the parser will require explicit commas between two list elements.
To reflect idiomatic use, examples in this document elide commas using these rules.
Identifiers name entities such as fields and aliases. An identifier is a sequence of one or more letters (which includes _ and $) and digits, optionally preceded by # or _#. It may not be _ or $. The first character in an identifier, or after an # if it contains one, must be a letter. Identifiers starting with a # or _ are reserved for definitions and hidden fields.
identifier = [ "#" | "_#" ] letter { letter | unicode_digit } .
a _x9 fieldName αβ
Some identifiers are predeclared.
CUE has a limited set of keywords. In addition, CUE reserves all identifiers starting with __(double underscores) as keywords. These are typically targets of pre-declared identifiers.
All keywords may be used as labels (field names). Unless noted otherwise, they can also be used as identifiers to refer to the same name.
The following keywords are values.
null true false
These can never be used to refer to a field of the same name. This restriction is to ensure compatibility with JSON configuration files.
The following keywords are used at the preamble of a CUE file. After the preamble, they may be used as identifiers to refer to namesake fields.
package import
The following keywords are used in comprehensions.
for in if let
The following character sequences represent operators and punctuation:
+ && == < = ( )
- || != > : { }
* & =~ <= ? [ ] ,
/ | !~ >= ! _|_ ... .
There are several kinds of numeric literals.
int_lit = decimal_lit | si_lit | octal_lit | binary_lit | hex_lit .
decimal_lit = ( "1" … "9" ) { [ "_" ] decimal_digit } .
decimals = decimal_digit { [ "_" ] decimal_digit } .
si_it = decimals [ "." decimals ] multiplier |
"." decimals multiplier .
binary_lit = "0b" binary_digit { binary_digit } .
hex_lit = "0" ( "x" | "X" ) hex_digit { [ "_" ] hex_digit } .
octal_lit = "0o" octal_digit { [ "_" ] octal_digit } .
multiplier = ( "K" | "M" | "G" | "T" | "P" ) [ "i" ]
float_lit = decimals "." [ decimals ] [ exponent ] |
decimals exponent |
"." decimals [ exponent ].
exponent = ( "e" | "E" ) [ "+" | "-" ] decimals .
An integer literal is a sequence of digits representing an integer value. An optional prefix sets a non-decimal base: 0o for octal, 0x or 0X for hexadecimal, and 0b for binary. In hexadecimal literals, letters a-f and A-F represent values 10 through 15. All integers allow interstitial underscores “_”; these have no meaning and are solely for readability.
Integer literals may have an SI or IEC multiplier. Multipliers can be used with fractional numbers. When multiplying a fraction by a multiplier, the result is truncated towards zero if it is not an integer.
42 1.5G // 1_000_000_000 1.3Ki // 1.3 * 1024 = trunc(1331.2) = 1331 170_141_183_460_469_231_731_687_303_715_884_105_727 0xBad_Face 0o755 0b0101_0001
A decimal floating-point literal is a representation of a decimal floating-point value (a float). It has an integer part, a decimal point, a fractional part, and an exponent part. The integer and fractional part comprise decimal digits; the exponent part is an e or E followed by an optionally signed decimal exponent. One of the integer part or the fractional part may be elided; one of the decimal point or the exponent may be elided.
0. 72.40 072.40 // == 72.40 2.71828 1.e+0 6.67428e-11 1E6 .25 .12345E+5
Neither a float_lit nor an si_lit may not appear after a token that is:
), ], }, ?, or ..A string literal represents a string constant obtained from concatenating a sequence of characters. Byte sequences are a sequence of bytes.
String and byte sequence literals are character sequences between, respectively, double and single quotes, as in "bar" and 'bar'. Within the quotes, any character may appear except newline and, respectively, unescaped double or single quote. String literals may only be valid UTF-8. Byte sequences may contain any sequence of bytes.
Several escape sequences allow arbitrary values to be encoded as ASCII text. An escape sequence starts with an escape delimiter, which is \ by default. The escape delimiter may be altered to be \ plus a fixed number of hash symbols # by padding the start and end of a string or byte sequence literal with this number of hash symbols.
There are four ways to represent the integer value as a numeric constant: \x followed by exactly two hexadecimal digits; \u followed by exactly four hexadecimal digits; \U followed by exactly eight hexadecimal digits, and a plain backslash \ followed by exactly three octal digits. In each case the value of the literal is the value represented by the digits in the corresponding base. Hexadecimal and octal escapes are only allowed within byte sequences (single quotes).
Although these representations all result in an integer, they have different valid ranges. Octal escapes must represent a value between 0 and 255 inclusive. Hexadecimal escapes satisfy this condition by construction. The escapes \u and \U represent Unicode code points so within them some values are illegal, in particular those above 0x10FFFF. Surrogate halves are allowed, but are translated into their non-surrogate equivalent internally.
The three-digit octal (\nnn) and two-digit hexadecimal (\xnn) escapes represent individual bytes of the resulting string; all other escapes represent the (possibly multi-byte) UTF-8 encoding of individual characters. Thus inside a string literal \377 and \xFF represent a single byte of value 0xFF=255, while ÿ, \u00FF, \U000000FF and \xc3\xbf represent the two bytes 0xc3 0xbf of the UTF-8 encoding of character U+00FF.
\a U+0007 alert or bell \b U+0008 backspace \f U+000C form feed \n U+000A line feed or newline \r U+000D carriage return \t U+0009 horizontal tab \v U+000b vertical tab \/ U+002f slash (solidus) \\ U+005c backslash \' U+0027 single quote (valid escape only within single quoted literals) \" U+0022 double quote (valid escape only within double quoted literals)
The escape \( is used as an escape for string interpolation. A \( must be followed by a valid CUE Expression, followed by a ).
All other sequences starting with a backslash are illegal inside literals.
escaped_char = `\` { `#` } ( "a" | "b" | "f" | "n" | "r" | "t" | "v" | "/" | `\` | "'" | `"` ) .
byte_value = octal_byte_value | hex_byte_value .
octal_byte_value = `\` octal_digit octal_digit octal_digit .
hex_byte_value = `\` "x" hex_digit hex_digit .
little_u_value = `\` "u" hex_digit hex_digit hex_digit hex_digit .
big_u_value = `\` "U" hex_digit hex_digit hex_digit hex_digit
hex_digit hex_digit hex_digit hex_digit .
unicode_value = unicode_char | little_u_value | big_u_value | escaped_char .
interpolation = "\(" Expression ")" .
string_lit = simple_string_lit |
multiline_string_lit |
simple_bytes_lit |
multiline_bytes_lit |
`#` string_lit `#` .
simple_string_lit = `"` { unicode_value | interpolation } `"` .
simple_bytes_lit = `'` { unicode_value | interpolation | byte_value } `'` .
multiline_string_lit = `"""` newline
{ unicode_value | interpolation | newline }
newline `"""` .
multiline_bytes_lit = "'''" newline
{ unicode_value | interpolation | byte_value | newline }
newline "'''" .
Carriage return characters (\r) inside string literals are discarded from the string value.
'a\000\xab' '\007' '\377' '\xa' // illegal: too few hexadecimal digits "\n" "\"" 'Hello, world!\n' "Hello, \( name )!" "日本語" "\u65e5本\U00008a9e" "\xff\u00FF" "\uD800" // illegal: surrogate half (TODO: probably should allow) "\U00110000" // illegal: invalid Unicode code point #"This is not an \(interpolation)"# #"This is an \#(interpolation)"# #"The sequence "\U0001F604" renders as \#U0001F604."#
These examples all represent the same string:
"日本語" // UTF-8 input text '日本語' // UTF-8 input text as byte sequence `日本語` // UTF-8 input text as a raw literal "\u65e5\u672c\u8a9e" // the explicit Unicode code points "\U000065e5\U0000672c\U00008a9e" // the explicit Unicode code points "\xe6\x97\xa5\xe6\x9c\xac\xe8\xaa\x9e" // the explicit UTF-8 bytes
If the source code represents a character as two code points, such as a combining form involving an accent and a letter, the result will appear as two code points if placed in a string literal.
Strings and byte sequences have a multiline equivalent. Multiline strings are like their single-line equivalent, but allow newline characters.
Multiline strings and byte sequences respectively start with a triple double quote (""") or triple single quote ('''), immediately followed by a newline, which is discarded from the string contents. The string is closed by a matching triple quote, which must be by itself on a newline, preceded by optional whitespace. The newline preceding the closing quote is discarded from the string contents. The whitespace before a closing triple quote must appear before any non-empty line after the opening quote and will be removed from each of these lines in the string literal. A closing triple quote may not appear in the string. To include it is suffices to escape one of the quotes.
"""
lily:
out of the water
out of itself
bass
picking bugs
off the moon
— Nick Virgilio, Selected Haiku, 1988
"""
This represents the same string as:
"lily:\nout of the water\nout of itself\n\n" + "bass\npicking bugs\noff the moon\n" + " — Nick Virgilio, Selected Haiku, 1988"
In addition to simple values like "hello" and 42.0, CUE has structs. A struct is a map from labels to values, like {a: 42.0, b: "hello"}. Structs are CUE's only way of building up complex values; lists, which we will see later, are defined in terms of structs.
All possible values are ordered in a lattice, a partial order where every two elements have a single greatest lower bound. A value a is an instance of a value b, denoted a ⊑ b, if b == a or b is more general than a, that is if a orders before b in the partial order (⊑ is not a CUE operator). We also say that b subsumes a in this case. In graphical terms, b is “above” a in the lattice.
At the top of the lattice is the single ancestor of all values, called top, denoted _ in CUE. Every value is an instance of top.
At the bottom of the lattice is the value called bottom, denoted _|_. A bottom value usually indicates an error. Bottom is an instance of every value.
An atom is any value whose only instances are itself and bottom. Examples of atoms are 42.0, "hello", true, null.
A value is concrete if it is either an atom, or a struct all of whose field values are themselves concrete, recursively.
CUE‘s values also include what we normally think of as types, like string and float. But CUE does not distinguish between types and values; only the relationship of values in the lattice is important. Each CUE “type” subsumes the concrete values that one would normally think of as part of that type. For example, “hello” is an instance of string, and 42.0 is an instance of float. In addition to string and float, CUE has null, int, bool and bytes. We informally call these CUE’s “basic types”.
false ⊑ bool true ⊑ bool true ⊑ true 5.0 ⊑ float bool ⊑ _ _|_ ⊑ _ _|_ ⊑ _|_ _ ⋢ _|_ _ ⋢ bool int ⋢ bool bool ⋢ int false ⋢ true true ⋢ false float ⋢ 5.0 5 ⋢ 6
The unification of values a and b is defined as the greatest lower bound of a and b. (That is, the value u such that u ⊑ a and u ⊑ b, and for any other value v for which v ⊑ a and v ⊑ b it holds that v ⊑ u.) Since CUE values form a lattice, the unification of two CUE values is always unique.
These all follow from the definition of unification:
a with itself is always a.a and b where a ⊑ b is always a.Unification in CUE is a binary expression, written a & b. It is commutative and associative. As a consequence, order of evaluation is irrelevant, a property that is key to many of the constructs in the CUE language as well as the tooling layered on top of it.
The disjunction of values a and b is defined as the least upper bound of a and b. (That is, the value d such that a ⊑ d and b ⊑ d, and for any other value e for which a ⊑ e and b ⊑ e, it holds that d ⊑ e.) This style of disjunctions is sometimes also referred to as sum types. Since CUE values form a lattice, the disjunction of two CUE values is always unique.
These all follow from the definition of disjunction:
a with itself is always a.a and b where a ⊑ b is always b.a with bottom is always a.Disjunction in CUE is a binary expression, written a | b. It is commutative, associative, and idempotent.
The unification of a disjunction with another value is equal to the disjunction composed of the unification of this value with all of the original elements of the disjunction. In other words, unification distributes over disjunction.
(a_0 | ... |a_n) & b ==> a_0&b | ... | a_n&b.
Expression Result
({a:1} | {b:2}) & {c:3} {a:1, c:3} | {b:2, c:3}
(int | string) & "foo" "foo"
("a" | "b") & "c" _|_
A disjunction is normalized if there is no element a for which there is an element b such that a ⊑ b.
Any element of a disjunction can be marked as a default by prefixing it with an asterisk *. Intuitively, when an expression needs to be resolved for an operation other than unification or disjunctions, non-starred elements are dropped in favor of starred ones if the starred ones do not resolve to bottom.
More precisely, any value v may be associated with a default value d, denoted (v, d) (not CUE syntax), where d must be in instance of v (d ⊑ v). The rules for unifying and disjoining such values are as follows:
U1: (v1, d1) & v2 => (v1&v2, d1&v2) U2: (v1, d1) & (v2, d2) => (v1&v2, d1&d2) D1: (v1, d1) | v2 => (v1|v2, d1) D2: (v1, d1) | (v2, d2) => (v1|v2, d1|d2)
Default values may be introduced within disjunctions by marking terms of a disjunction with an asterisk * (a unary expression). The default value of a disjunction with marked terms is the disjunction of those marked terms, applying the following rules for marks:
M1: *v => (v, v) M2: *(v1, d1) => (v1, d1)
In general, any operation f in CUE involving default values proceeds along the following lines
O1: f((v1, d1), ..., (vn, dn)) => (f(v1, ..., vn), f(d1, ..., dn))
where, with the exception of disjunction, a value v without a default value is promoted to (v, v).
Expression Value-default pair Rules applied
*"tcp" | "udp" ("tcp"|"udp", "tcp") M1, D1
string | *"foo" (string, "foo") M1, D1
*1 | 2 | 3 (1|2|3, 1) M1, D1
(*1|2|3) | (1|*2|3) (1|2|3, 1|2) M1, D1, D2
(*1|2|3) | *(1|*2|3) (1|2|3, 1|2) M1, D1, M2, D2
(*1|2|3) | (1|*2|3)&2 (1|2|3, 1|2) M1, D1, U1, D2
(*1|2) & (1|*2) (1|2, _|_) M1, D1, U2
(*1|2) + (1|*2) ((1|2)+(1|2), 3) M1, D1, O1
The rules of subsumption for defaults can be derived from the above definitions and are as follows.
(v2, d2) ⊑ (v1, d1) if v2 ⊑ v1 and d2 ⊑ d1 (v1, d1) ⊑ v if v1 ⊑ v v ⊑ (v1, d1) if v ⊑ d1
Expression Resolves to
"tcp" | "udp" "tcp" | "udp"
*"tcp" | "udp" "tcp"
float | *1 1
*string | 1.0 string
(*1|2|3) | (1|*2|3) 1|2
(*1|2|3) & (1|*2|3) 1|2|3 // default is _|_
(* >=5 | int) & (* <=5 | int) 5
(*"tcp"|"udp") & ("udp"|*"tcp") "tcp"
(*"tcp"|"udp") & ("udp"|"tcp") "tcp"
(*"tcp"|"udp") & "tcp" "tcp"
(*"tcp"|"udp") & (*"udp"|"tcp") "tcp" | "udp" // default is _|_
(*true | false) & bool true
(*true | false) & (true | false) true
{a: 1} | {b: 1} {a: 1} | {b: 1}
{a: 1} | *{b: 1} {b:1}
*{a: 1} | *{b: 1} {a: 1} | {b: 1}
({a: 1} | {b: 1}) & {a:1} {a:1} // after eliminating {a:1,b:1} by normalization
({a:1}|*{b:1}) & ({a:1}|*{b:1}) {b:1} // after eliminating {a:1,b:1} by normalization
Any evaluation error in CUE results in a bottom value, respresented by the token _|_. Bottom is an instance of every other value. Any evaluation error is represented as bottom.
Implementations may associate error strings with different instances of bottom; logically they all remain the same value.
Top is represented by the underscore character _, lexically an identifier. Unifying any value v with top results v itself.
Expr Result _ & 5 5 _ & _ _ _ & _|_ _|_ _ | _|_ _
The null value is represented with the keyword null. It has only one parent, top, and one child, bottom. It is unordered with respect to any other value.
null_lit = "null"
null & 8 _|_ null & _ null null & _|_ _|_
A boolean type represents the set of Boolean truth values denoted by the keywords true and false. The predeclared boolean type is bool; it is a defined type and a separate element in the lattice.
boolean_lit = "true" | "false"
bool & true true true & true true true & false _|_ bool & (false|true) false | true bool & (true|false) true | false
The integer type represents the set of all integral numbers. The decimal floating-point type represents the set of all decimal floating-point numbers. They are two distinct types. Both are instances instances of a generic number type.
The predeclared number, integer, decimal floating-point types are number, int and float; they are defined types.
A decimal floating-point literal always has type float; it is not an instance of int even if it is an integral number.
Integer literals are always of type int and don't match type float.
Numeric literals are exact values of arbitrary precision. If the operation permits it, numbers should be kept in arbitrary precision.
Implementation restriction: although numeric values have arbitrary precision in the language, implementations may implement them using an internal representation with limited precision. That said, every implementation must:
The string type represents the set of UTF-8 strings, not allowing surrogates. The predeclared string type is string; it is a defined type.
The length of a string s (its size in bytes) can be discovered using the built-in function len.
The bytes type represents the set of byte sequences. A byte sequence value is a (possibly empty) sequence of bytes. The number of bytes is called the length of the byte sequence and is never negative. The predeclared byte sequence type is bytes; it is a defined type.
A bound, syntactically a unary expression, defines an infinite disjunction of concrete values than can be represented as a single comparison.
For any comparison operator op except ==, op a is the disjunction of every x such that x op a.
2 & >=2 & <=5 // 2, where 2 is either an int or float. 2.5 & >=1 & <=5 // 2.5 2 & >=1.0 & <3.0 // 2.0 2 & >1 & <3.0 // 2.0 2.5 & int & >1 & <5 // _|_ 2.5 & float & >1 & <5 // 2.5 int & 2 & >1.0 & <3.0 // _|_ 2.5 & >=(int & 1) & <5 // _|_ >=0 & <=7 & >=3 & <=10 // >=3 & <=7 !=null & 1 // 1 >=5 & <=5 // 5
A struct is a set of elements called fields, each of which has a name, called a label, and value.
We say a label is defined for a struct if the struct has a field with the corresponding label. The value for a label f of struct a is denoted a.f. A struct a is an instance of b, or a ⊑ b, if for any label f defined for b, label f is also defined for a and a.f ⊑ b.f. Note that if a is an instance of b it may have fields with labels that are not defined for b.
The (unique) struct with no fields, written {}, has every struct as an instance. It can be considered the type of all structs.
{a: 1} ⊑ {}
{a: 1, b: 1} ⊑ {a: 1}
{a: 1} ⊑ {a: int}
{a: 1, b: 1} ⊑ {a: int, b: float}
{} ⋢ {a: 1}
{a: 2} ⋢ {a: 1}
{a: 1} ⋢ {b: 1}
A field may be required or optional. The successful unification of structs a and b is a new struct c which has all fields of both a and b, where the value of a field f in c is a.f & b.f if f is in both a and b, or just a.f or b.f if f is in just a or b, respectively. If a field f is in both a and b, c.f is optional only if both a.f and b.f are optional. Any references to a or b in their respective field values need to be replaced with references to c. The result of a unification is bottom (_|_) if any of its non-optional fields evaluates to bottom, recursively.
Syntactically, a field is marked as optional by following its label with a ?. The question mark is not part of the field name. A struct literal may contain multiple fields with the same label, the result of which is a single field with the same properties as defined as the unification of two fields resulting from unifying two structs.
These examples illustrate required fields only. Examples with optional fields follow below.
Expression Result (without optional fields)
{a: int, a: 1} {a: 1}
{a: int} & {a: 1} {a: 1}
{a: >=1 & <=7} & {a: >=5 & <=9} {a: >=5 & <=7}
{a: >=1 & <=7, a: >=5 & <=9} {a: >=5 & <=7}
{a: 1} & {b: 2} {a: 1, b: 2}
{a: 1, b: int} & {b: 2} {a: 1, b: 2}
{a: 1} & {a: 2} _|_
A struct may define constraints that apply to fields that are added when unified with another struct using pattern or default constraints.
A pattern constraint, denoted [pattern]: value, defines a pattern, which is a value of type string, and a value to unify with fields whose label match that pattern. When unifying structs a and b, a pattern constraint [p]: v declared in a defines that the value v should unify with any field in the resulting struct c whose label unifies with pattern p and for which there exists no field in a with the same label.
Additionally, a default constraint, denoted ...value, defines a value to unify with any field for which there is no other declaration in a struct. When unifying structs a and b, a default constraint ...v declared in a defines that the value v should unify with any field in the resulting struct c whose label does not unify with any of the patterns of the pattern constraints defined for a and for which there exists no field in a with that label. The token ... is a shorthand for ..._.
a: {
foo: string // foo is a string
["^i"]: int // all other fields starting with i are integers
["^b"]: bool // all other fields starting with b are booleans
...string // all other fields must be a string
}
b: a & {
i3: 3
bar: true
other: "a string"
}
Concrete field labels may be an identifier or string, the latter of which may be interpolated. Fields with identifier labels can be referred to within the scope they are defined, string labels cannot. References within such interpolated strings are resolved within the scope of the struct in which the label sequence is defined and can reference concrete labels lexically preceding the label within a label sequence.
StructLit = "{" { Declaration "," } "}" .
Declaration = Field | Ellipsis | Embedding | LetClause | attribute .
Ellipsis = "..." [ Expression ] .
Embedding = Comprehension | AliasExpr .
Field = Label ":" { Label ":" } Expression { attribute } .
Label = [ identifier "=" ] LabelExpr .
LabelExpr = LabelName [ "?" ] | "[" AliasExpr "]" .
LabelName = identifier | simple_string_lit .
attribute = "@" identifier "(" attr_tokens ")" .
attr_tokens = { attr_token |
"(" attr_tokens ")" |
"[" attr_tokens "]" |
"{" attr_tokens "}" } .
attr_token = /* any token except '(', ')', '[', ']', '{', or '}' */
Expression Result (without optional fields)
a: { foo?: string } {}
b: { foo: "bar" } { foo: "bar" }
c: { foo?: *"bar" | string } {}
d: a & b { foo: "bar" }
e: b & c { foo: "bar" }
f: a & c {}
g: a & { foo?: number } {}
h: b & { foo?: number } _|_
i: c & { foo: string } { foo: "bar" }
intMap: [string]: int
intMap: {
t1: 43
t2: 2.4 // error: 2.4 is not an integer
}
nameMap: [string]: {
firstName: string
nickName: *firstName | string
}
nameMap: hank: { firstName: "Hank" }
The optional field set defined by nameMap matches every field, in this case just hank, and unifies the associated constraint with the matched field, resulting in:
nameMap: hank: {
firstName: "Hank"
nickName: "Hank"
}
By default, structs are open to adding fields. Instances of an open struct p may contain fields not defined in p. This is makes it easy to add fields, but can lead to bugs:
S: {
field1: string
}
S1: S & { field2: "foo" }
// S1 is { field1: string, field2: "foo" }
A: {
field1: string
field2: string
}
A1: A & {
feild1: "foo" // "field1" was accidentally misspelled
}
// A1 is
// { field1: string, field2: string, feild1: "foo" }
// not the intended
// { field1: "foo", field2: string }
A closed struct c is a struct whose instances may not declare any field with a name that does not match the name of a regular or optional field, or the pattern of a pattern constraint defined in c. Hidden fields are excluded from this limitation. A struct that is the result of unifying any struct with a ... declaration is defined for all fields. Closing a struct is equivalent to adding ..._|_ to it.
Syntactically, structs are closed explicitly with the close builtin or implicitly and recursively by definitions.
A: close({
field1: string
field2: string
})
A1: A & {
feild1: string
} // _|_ feild1 not defined for A
A2: A & {
for k,v in { feild1: string } {
k: v
}
} // _|_ feild1 not defined for A
C: close({
[_]: _
})
C2: C & {
for k,v in { thisIsFine: string } {
"\(k)": v
}
}
D: close({
// Values generated by comprehensions are treated as embeddings.
for k,v in { x: string } {
"\(k)": v
}
})
A struct may contain an embedded value, an operand used as a declaration. An embedded value of type struct is unified with the struct in which it is embedded, but disregarding the restrictions imposed by closed structs. So if an embedding contains a closed struct, the corresponding resulting struct will also be closed, but may have fields that are not allowed if normal rules for closed structs were observed.
If an embedded value is not of type struct, the struct may only have definitions or hidden fields. Regular fields are not allowed in such case.
The result of { A } is A for any A (including definitions).
Syntactically, embeddings may be any expression.
S1: {
a: 1
b: 2
{
c: 3
}
}
// S1 is { a: 1, b: 2, c: 3 }
S2: close({
a: 1
b: 2
{
c: 3
}
})
// same as close(S1)
S3: {
a: 1
b: 2
close({
c: 3
})
}
// same as S2
A field is a definition if its identifier starts with # or _#. A field is hidden if its starts with a _. All other fields are regular.
Definitions and hidden fields are not emitted when converting a CUE program to data and are never required to be concrete.
Referencing a definition will recursively close it. That is, a referenced definition will not unify with a struct that would add a field anywhere within the definition that it does not already define or explicitly allow with a pattern constraint or .... Embeddings allow bypassing this check.
If referencing a definition would always result in an error, implementations may report this inconsistency at the point of its declaration.
#MyStruct: {
sub: field: string
}
#MyStruct: {
sub: enabled?: bool
}
myValue: #MyStruct & {
sub: feild: 2 // error, feild not defined in #MyStruct
sub: enabled: true // okay
}
#D: {
#OneOf
c: int // adds this field.
}
#OneOf: { a: int } | { b: int }
D1: #D & { a: 12, c: 22 } // { a: 12, c: 22 }
D2: #D & { a: 12, b: 33 } // _|_ // cannot define both `a` and `b`
#A: {a: int}
B: {
#A
b: c: int
}
x: B
x: d: 3 // not allowed, as closed by embedded #A
y: B.b
y: d: 3 // allowed as nothing closes b
#B: {
#A
b: c: int
}
z: #B.b
z: d: 3 // not allowed, as referencing #B closes b
Attributes allow associating meta information with values. Their primary purpose is to define mappings between CUE and other representations. Attributes do not influence the evaluation of CUE.
An attribute associates an identifier with a value, a balanced token sequence, which is a sequence of CUE tokens with balanced brackets ((), [], and {}). The sequence may not contain interpolations.
Fields, structs and packages can be associated with a set of attributes. Attributes accumulate during unification, but implementations may remove duplicates that have the same source string representation. The interpretation of an attribute, including the handling of multiple attributes for a given identifier, is up to the consumer of the attribute.
Field attributes define additional information about a field, such as a mapping to a protocol buffer tag or alternative name of the field when mapping to a different language.
// Package attribute
@protobuf(proto3)
myStruct1: {
// Struct attribute:
@jsonschema(id="https://example.org/mystruct1.json")
// Field attributes
field: string @go(Field)
attr: int @xml(,attr) @go(Attr)
}
myStruct2: {
field: string @go(Field)
attr: int @xml(a1,attr) @go(Attr)
}
Combined: myStruct1 & myStruct2
// field: string @go(Field)
// attr: int @xml(,attr) @xml(a1,attr) @go(Attr)
Aliases name values that can be referred to within the scope in which they are declared. The name of an alias must be unique within its scope.
AliasExpr = Expression | identifier "=" Expression .
Aliases can appear in several positions:
In front of a Label (X=label: value):
label would be bound to if it were a valid identifier.foo?: bar and [foo]: bar), the bound identifier is only visible within the field value (bar).Inside a bracketed label ([X=expr]: value):
expr within the instances of the field value (value).Before a list element ([ X=value, X+1 ]) (Not yet implemented)
// An alias declaration
Alias = 3
a: Alias // 3
// A field alias
foo: X // 4
X="not an identifier": 4
// A label alias
[Y=string]: { name: Y }
foo: { value: 1 } // outputs: foo: { name: "foo", value: 1 }
Let declarations bind an identifier to an expression. The identifier is visible within the scope in which it is declared. The identifier must be unique within its scope.
let x = expr a: x + 1 b: x + 2
A field whose value is a struct with a single field may be written as a colon-separated sequence of the two field names, followed by a colon and the value of that single field.
job: myTask: replicas: 2
expands to
job: {
myTask: {
replicas: 2
}
}
A list literal defines a new value of type list. A list may be open or closed. An open list is indicated with a ... at the end of an element list, optionally followed by a value for the remaining elements.
The length of a closed list is the number of elements it contains. The length of an open list is the its number of elements as a lower bound and an unlimited number of elements as its upper bound.
ListLit = "[" [ ElementList [ "," [ Ellipsis ] ] [ "," ] "]" .
ElementList = Embedding { "," Embedding } .
Lists can be thought of as structs:
List: *null | {
Elem: _
Tail: List
}
For closed lists, Tail is null for the last element, for open lists it is *null | List, defaulting to the shortest variant. For instance, the open list [ 1, 2, ... ] can be represented as:
open: List & { Elem: 1, Tail: { Elem: 2 } }
and the closed version of this list, [ 1, 2 ], as
closed: List & { Elem: 1, Tail: { Elem: 2, Tail: null } }
Using this representation, the subsumption rule for lists can be derived from those of structs. Implementations are not required to implement lists as structs. The Elem and Tail fields are not special and len will not work as expected in these cases.
A block is a possibly empty sequence of declarations. The braces of a struct literal { ... } form a block, but there are others as well:
for and let clause in a comprehension is considered to be its own implicit block.Blocks nest and influence scoping.
A declaration may bind an identifier to a field, alias, or package. Every identifier in a program must be declared. Other than for fields, no identifier may be declared twice within the same block. For fields an identifier may be declared more than once within the same block, resulting in a field with a value that is the result of unifying the values of all fields with the same identifier. String labels do not bind an identifier to the respective field.
The scope of a declared identifier is the extent of source text in which the identifier denotes the specified field, alias, or package.
CUE is lexically scoped using blocks:
An identifier declared in a block may be redeclared in an inner block. While the identifier of the inner declaration is in scope, it denotes the entity declared by the inner declaration.
The package clause is not a declaration; the package name does not appear in any scope. Its purpose is to identify the files belonging to the same package and to specify the default name for import declarations.
CUE predefines a set of types and builtin functions. For each of these there is a corresponding keyword which is the name of the predefined identifier, prefixed with __.
Functions
len close and or
Types
null The null type and value
bool All boolean values
int All integral numbers
float All decimal floating-point numbers
string Any valid UTF-8 sequence
bytes Any valid byte sequence
Derived Value
number int | float
uint >=0
uint8 >=0 & <=255
int8 >=-128 & <=127
uint16 >=0 & <=65536
int16 >=-32_768 & <=32_767
rune >=0 & <=0x10FFFF
uint32 >=0 & <=4_294_967_296
int32 >=-2_147_483_648 & <=2_147_483_647
uint64 >=0 & <=18_446_744_073_709_551_615
int64 >=-9_223_372_036_854_775_808 & <=9_223_372_036_854_775_807
uint128 >=0 & <=340_282_366_920_938_463_463_374_607_431_768_211_455
int128 >=-170_141_183_460_469_231_731_687_303_715_884_105_728 &
<=170_141_183_460_469_231_731_687_303_715_884_105_727
float32 >=-3.40282346638528859811704183484516925440e+38 &
<=3.40282346638528859811704183484516925440e+38
float64 >=-1.797693134862315708145274237317043567981e+308 &
<=1.797693134862315708145274237317043567981e+308
An identifier of a package may be exported to permit access to it from another package. All identifiers not starting with _ (so all regular fields and definitions starting with #) are exported. Any identifier starting with _ is not visible outside the package and resides in a separate namespace than namesake identifiers of other packages.
package mypackage
foo: string // visible outside mypackage
"bar": string // visible outside mypackage
#Foo: { // visible outside mypackage
a: 1 // visible outside mypackage
_b: 2 // not visible outside mypackage
#C: { // visible outside mypackage
d: 4 // visible outside mypackage
}
_#E: foo // not visible outside mypackage
}
Given a set of identifiers, an identifier is called unique if it is different from every other in the set, after applying normalization following Unicode Annex #31. Two identifiers are different if they are spelled differently or if they appear in different packages and are not exported. Otherwise, they are the same.
A field associates the value of an expression to a label within a struct. If this label is an identifier, it binds the field to that identifier, so the field's value can be referenced by writing the identifier. String labels are not bound to fields.
a: {
b: 2
"s": 3
c: b // 2
d: s // _|_ unresolved identifier "s"
e: a.s // 3
}
If an expression may result in a value associated with a default value as described in default values, the field binds to this value-default pair.
Within a struct, a let clause binds an identifier to the given expression.
Within the scope of the identifier, the identifier refers to the locally declared expression. The expression is evaluated in the scope it was declared.
An expression specifies the computation of a value by applying operators and built-in functions to operands.
Expressions that require concrete values are called incomplete if any of their operands are not concrete, but define a value that would be legal for that expression. Incomplete expressions may be left unevaluated until a concrete value is requested at the application level.
Operands denote the elementary values in an expression. An operand may be a literal, a (possibly qualified) identifier denoting field, alias, or let declaration, or a parenthesized expression.
Operand = Literal | OperandName | "(" Expression ")" .
Literal = BasicLit | ListLit | StructLit .
BasicLit = int_lit | float_lit | string_lit |
null_lit | bool_lit | bottom_lit | top_lit .
OperandName = identifier | QualifiedIdent .
A qualified identifier is an identifier qualified with a package name prefix.
QualifiedIdent = PackageName "." identifier .
A qualified identifier accesses an identifier in a different package, which must be [imported]. The identifier must be declared in the [package block] of that package.
math.Sin // denotes the Sin function in package math
An identifier operand refers to a field and is called a reference. The value of a reference is a copy of the expression associated with the field that it is bound to, with any references within that expression bound to the respective copies of the fields they were originally bound to. Implementations may use a different mechanism to evaluate as long as these semantics are maintained.
a: {
place: string
greeting: "Hello, \(place)!"
}
b: a & { place: "world" }
c: a & { place: "you" }
d: b.greeting // "Hello, world!"
e: c.greeting // "Hello, you!"
Primary expressions are the operands for unary and binary expressions.
Slice: indices must be complete ([0, 1, 2, 3] | [2, 3])[0:2] => [0, 1] | [2, 3] ([0, 1, 2, 3] | *[2, 3])[0:2] => [0, 1] | [2, 3] ([0,1,2,3]|[2,3], [2,3])[0:2] => ([0,1]|[2,3], [2,3]) Index a: (1|2, 1) b: ([0,1,2,3]|[2,3], [2,3])[a] => ([0,1,2,3]|[2,3][a], 3) Binary operation A binary is only evaluated if its operands are complete. Input Maximum allowed evaluation a: string string b: 2 2 c: a * b a * 2 An error in a struct is if the evaluation of any expression results in bottom, where an incomplete expression is not considered bottom.
PrimaryExpr =
Operand |
PrimaryExpr Selector |
PrimaryExpr Index |
PrimaryExpr Slice |
PrimaryExpr Arguments .
Selector = "." (identifier | simple_string_lit) .
Index = "[" Expression "]" .
Argument = Expression .
Arguments = "(" [ ( Argument { "," Argument } ) [ "," ] ] ")" .
x 2 (s + ".txt") f(3.1415, true) m["foo"] s[i : j + 1] obj.color f.p[i].x
For a primary expression x that is not a package name, the selector expression
x.f
denotes the element of a struct x identified by f.
f must be an identifier or a string literal identifying any definition or regular non-optional field. The identifier f is called the field selector.
If x is a package name, see the section on qualified identifiers.
Otherwise, if x is not a struct, or if f does not exist in x, the result of the expression is bottom (an error). In the latter case the expression is incomplete. The operand of a selector may be associated with a default.
T: {
x: int
y: 3
"x-y": 4
}
a: T.x // int
b: T.y // 3
c: T.z // _|_ // field 'z' not found in T
d: T."x-y" // 4
e: {a: 1|*2} | *{a: 3|*4}
f: e.a // 4 (default value)
A primary expression of the form
a[x]
denotes the element of a list or struct a indexed by x. The value x is called the index or field name, respectively. The following rules apply:
If a is not a struct:
a is a list (which need not be complete)x unified with int must be concrete.x is in range if 0 <= x < len(a), where only the explicitly defined values of an open-ended list are considered, otherwise it is out of rangeThe result of a[x] is
for a of list type:
x, if x is within rangefor a of struct type:
x unified with string must be concrete.x of struct a, if this field exists[ 1, 2 ][1] // 2 [ 1, 2 ][2] // _|_ [ 1, 2, ...][2] // _|_
Both the operand and index value may be a value-default pair.
va[vi] => va[vi] va[(vi, di)] => (va[vi], va[di]) (va, da)[vi] => (va[vi], da[vi]) (va, da)[(vi, di)] => (va[vi], da[di])
Fields Result x: [1, 2] | *[3, 4] ([1,2]|[3,4], [3,4]) i: int | *1 (int, 1) v: x[i] (x[i], 4)
Operators combine operands into expressions.
Expression = UnaryExpr | Expression binary_op Expression . UnaryExpr = PrimaryExpr | unary_op UnaryExpr . binary_op = "|" | "&" | "||" | "&&" | "==" | rel_op | add_op | mul_op . rel_op = "!=" | "<" | "<=" | ">" | ">=" | "=~" | "!~" . add_op = "+" | "-" . mul_op = "*" | "/" . unary_op = "+" | "-" | "!" | "*" | rel_op .
Comparisons are discussed elsewhere. For any binary operators, the operand types must unify.
Operands of unary and binary expressions may be associated with a default using the following
Field Resulting Value-Default pair a: *1|2 (1|2, 1) b: -a (-a, -1) c: a + 2 (a+2, 3) d: a + a (a+a, 2)
Unary operators have the highest precedence.
There are eight precedence levels for binary operators. Multiplication operators binds strongest, followed by addition operators, comparison operators, && (logical AND), || (logical OR), & (unification), and finally | (disjunction):
Precedence Operator
7 * /
6 + -
5 == != < <= > >= =~ !~
4 &&
3 ||
2 &
1 |
Binary operators of the same precedence associate from left to right. For instance, x / y * z is the same as (x / y) * z.
+x
23 + 3*x[i]
x <= f()
f() || g()
x == y+1 && y == z-1
2 | int
{ a: 1 } & { b: 2 }
Arithmetic operators apply to numeric values and yield a result of the same type as the first operand. The four standard arithmetic operators (+, -, *, /) apply to integer and decimal floating-point types; + and * also apply to lists and strings.
+ sum integers, floats, lists, strings, bytes - difference integers, floats * product integers, floats, lists, strings, bytes / quotient integers, floats
For any operator that accepts operands of type float, any operand may be of type int or float, in which case the result will be float if it cannot be represented as an int or if any of the operands are float, or int otherwise. So the result of 1 / 2 is 0.5 and is of type float.
The result of division by zero is bottom (an error).
Integer division is implemented through the builtin functions quo, rem, div, and mod.
The unary operators + and - are defined for numeric values as follows:
+x is 0 + x -x negation is 0 - x
Lists can be concatenated using the + operator. Opens list are closed to their default value beforehand.
[ 1, 2 ] + [ 3, 4 ] // [ 1, 2, 3, 4 ] [ 1, 2, ... ] + [ 3, 4 ] // [ 1, 2, 3, 4 ] [ 1, 2 ] + [ 3, 4, ... ] // [ 1, 2, 3, 4 ]
Lists can be multiplied with a non-negativeint using the * operator to create a repeated the list by the indicated number.
3*[1,2] // [1, 2, 1, 2, 1, 2] 3*[1, 2, ...] // [1, 2, 1, 2, 1 ,2] [byte]*4 // [byte, byte, byte, byte] 0*[1,2] // []
Strings can be concatenated using the + operator:
s: "hi " + name + " and good bye"
String addition creates a new string by concatenating the operands.
A string can be repeated by multiplying it:
s: "etc. "*3 // "etc. etc. etc. "
Comparison operators compare two operands and yield an untyped boolean value.
== equal != not equal < less <= less or equal > greater >= greater or equal =~ matches regular expression !~ does not match regular expression
In any comparison, the types of the two operands must unify or one of the operands must be null.
The equality operators == and != apply to operands that are comparable. The ordering operators <, <=, >, and >= apply to operands that are ordered. The matching operators =~ and !~ apply to a string and regular expression operand. These terms and the result of the comparisons are defined as follows:
\C.s =~ r is true if s matches the regular expression r.s !~ r is true if s does not match regular expression r.3 < 4 // true
3 < 4.0 // true
null == 2 // false
null != {} // true
{} == {} // _|_: structs are not comparable against structs
"Wild cats" =~ "cat" // true
"Wild cats" !~ "dog" // true
"foo" =~ "^[a-z]{3}$" // true
"foo" =~ "^[a-z]{4}$" // false
Logical operators apply to boolean values and yield a result of the same type as the operands. The right operand is evaluated conditionally.
&& conditional AND p && q is "if p then q else false" || conditional OR p || q is "if p then true else q" ! NOT !p is "not p"
Calls can be made to core library functions, called builtins. Given an expression f of function type F,
f(a1, a2, … an)
calls f with arguments a1, a2, … an. Arguments must be expressions of which the values are an instance of the parameter types of F and are evaluated before the function is called.
a: math.Atan2(x, y)
In a function call, the function value and arguments are evaluated in the usual order. After they are evaluated, the parameters of the call are passed by value to the function and the called function begins execution. The return parameters of the function are passed by value back to the calling function when the function returns.
Lists and fields can be constructed using comprehensions.
Comprehensions define a clause sequence that consists of a sequence of for, if, and let clauses, nesting from left to right. The sequence must start with a for or if clause. The for and let clauses each define a new scope in which new values are bound to be available for the next clause.
The for clause binds the defined identifiers, on each iteration, to the next value of some iterable value in a new scope. A for clause may bind one or two identifiers. If there is one identifier, it binds it to the value of a list element or struct field value. If there are two identifiers, the first value will be the key or index, if available, and the second will be the value.
For lists, for iterates over all elements in the list after closing it. For structs, for iterates over all non-optional regular fields.
An if clause, or guard, specifies an expression that terminates the current iteration if it evaluates to false.
The let clause binds the result of an expression to the defined identifier in a new scope.
A current iteration is said to complete if the innermost block of the clause sequence is reached. Syntactically, the comprehension value is a struct. A comprehension can generate non-struct values by embedding such values within this struct.
Within lists, the values yielded by a comprehension are inserted in the list at the position of the comprehension. Within structs, the values yielded by a comprehension are embedded within the struct. Both structs and lists may contain multiple comprehensions.
Comprehension = Clauses StructLit .
Clauses = StartClause { [ "," ] Clause } .
StartClause = ForClause | GuardClause .
Clause = StartClause | LetClause .
ForClause = "for" identifier [ "," identifier ] "in" Expression .
GuardClause = "if" Expression .
LetClause = "let" identifier "=" Expression .
a: [1, 2, 3, 4]
b: [ for x in a if x > 1 { x+1 } ] // [3, 4, 5]
c: {
for x in a
if x < 4
let y = 1 {
"\(x)": x + y
}
}
d: { "1": 2, "2": 3, "3": 4 }
String interpolation allows constructing strings by replacing placeholder expressions with their string representation. String interpolation may be used in single- and double-quoted strings, as well as their multiline equivalent.
A placeholder consists of “(” followed by an expression and a “)”. The expression is evaluated in the scope within which the string is defined.
The result of the expression is substituted as follows:
a: "World" b: "Hello \( a )!" // Hello World!
Built-in functions are predeclared. They are called like any other function.
lenThe built-in function len takes arguments of various types and return a result of type int.
Argument type Result string string length in bytes bytes length of byte sequence list list length, smallest length for an open list struct number of distinct data fields, excluding optional
Expression Result
len("Hellø") 6
len([1, 2, 3]) 3
len([1, 2, ...]) >=2
closeThe builtin function close converts a partially defined, or open, struct to a fully defined, or closed, struct.
andThe built-in function and takes a list and returns the result of applying the & operator to all elements in the list. It returns top for the empty list.
Expression: Result and([a, b]) a & b and([a]) a and([]) _
orThe built-in function or takes a list and returns the result of applying the | operator to all elements in the list. It returns bottom for the empty list.
Expression: Result or([a, b]) a | b or([a]) a or([]) _|_
div, mod, quo and remFor two integer values x and y, the integer quotient q = div(x, y) and remainder r = mod(x, y) implement Euclidean division and satisfy the following relationship:
r = x - y*q with 0 <= r < |y|
where |y| denotes the absolute value of y.
x y div(x, y) mod(x, y) 5 3 1 2 -5 3 -2 1 5 -3 -1 2 -5 -3 2 1
For two integer values x and y, the integer quotient q = quo(x, y) and remainder r = rem(x, y) implement truncated division and satisfy the following relationship:
x = q*y + r and |r| < |y|
with quo(x, y) truncated towards zero.
x y quo(x, y) rem(x, y) 5 3 1 2 -5 3 -1 -2 5 -3 -1 2 -5 -3 1 -2
A zero divisor in either case results in bottom (an error).
Implementations are required to interpret or reject cycles encountered during evaluation according to the rules in this section.
A reference cycle occurs if a field references itself, either directly or indirectly.
// x references itself x: x // indirect cycles b: c c: d d: b
Implementations should treat these as _. Two particular cases are discussed below.
An expression of the form a & e, where a is an atom and e is an expression, always evaluates to a or bottom. As it does not matter how we fail, we can assume the result to be a and postpone validating a == e until after all referenecs in e have been resolved.
// Config Evaluates to (requiring concrete values)
x: { x: {
a: b + 100 a: _|_ // cycle detected
b: a - 100 b: _|_ // cycle detected
} }
y: x & { y: {
a: 200 a: 200 // asserted that 200 == b + 100
b: 100
} }
A field value of the form r & v, where r evaluates to a reference cycle and v is a concrete value, evaluates to v. Unification is idempotent and unifying a value with itself ad infinitum, which is what the cycle represents, results in this value. Implementations should detect cycles of this kind, ignore r, and take v as the result of unification.
Configuration Evaluated
// c Cycles in nodes of type struct evaluate
// ↙︎ ↖ to the fixed point of unifying their
// a → b values ad infinitum.
a: b & { x: 1 } // a: { x: 1, y: 2, z: 3 }
b: c & { y: 2 } // b: { x: 1, y: 2, z: 3 }
c: a & { z: 3 } // c: { x: 1, y: 2, z: 3 }
// resolve a b & {x:1}
// substitute b c & {y:2} & {x:1}
// substitute c a & {z:3} & {y:2} & {x:1}
// eliminate a (cycle) {z:3} & {y:2} & {x:1}
// simplify {x:1,y:2,z:3}
This rule also applies to field values that are disjunctions of unification operations of the above form.
a: b&{x:1} | {y:1} // {x:1,y:3,z:2} | {y:1}
b: {x:2} | c&{z:2} // {x:2} | {x:1,y:3,z:2}
c: a&{y:3} | {z:3} // {x:1,y:3,z:2} | {z:3}
// resolving a b&{x:1} | {y:1}
// substitute b ({x:2} | c&{z:2})&{x:1} | {y:1}
// simplify c&{z:2}&{x:1} | {y:1}
// substitute c (a&{y:3} | {z:3})&{z:2}&{x:1} | {y:1}
// simplify a&{y:3}&{z:2}&{x:1} | {y:1}
// eliminate a (cycle) {y:3}&{z:2}&{x:1} | {y:1}
// expand {x:1,y:3,z:2} | {y:1}
Note that all nodes that form a reference cycle to form a struct will evaluate to the same value. If a field value is a disjunction, any element that is part of a cycle will evaluate to this value.
A structural cycle is when a node references one of its ancestor nodes. It is possible to construct a structural cycle by unifying two acyclic values:
// acyclic
y: {
f: h: g
g: _
}
// acyclic
x: {
f: _
g: f
}
// introduces structural cycle
z: x & y
Implementations should be able to detect such structural cycles dynamically.
A structural cycle can result in infinite structure or evaluation loops.
// infinite structure
a: b: a
// infinite evaluation
f: {
n: int
out: n + (f & {n: 1}).out
}
CUE must allow or disallow structural cycles under certain circumstances.
If a node a references an ancestor node, we call it and any of its field values a.f cyclic. So if a is cyclic, all of its descendants are also regarded as cyclic. A given node x, whose value is composed of the conjuncts c1 & ... & cn, is valid if any of its conjuncts is not cyclic.
// Disallowed: a list of infinite length with all elements being 1.
#List: {
head: 1
tail: #List
}
// Disallowed: another infinite structure (a:{b:{d:{b:{d:{...}}}}}, ...).
a: {
b: c
}
c: {
d: a
}
// #List defines a list of arbitrary length. Because the recursive reference
// is part of a disjunction, this does not result in a structural cycle.
#List: {
head: _
tail: null | #List
}
// Usage of #List. The value of tail in the most deeply nested element will
// be `null`: as the value of the disjunct referring to list is the only
// conjunct, all conjuncts are cyclic and the value is invalid and so
// eliminated from the disjunction.
MyList: #List & { head: 1, tail: { head: 2 }}
CUE configurations are constructed combining instances. An instance, in turn, is constructed from one or more source files belonging to the same package that together declare the data representation. Elements of this data representation may be exported and used in other instances.
Each source file consists of an optional package clause defining collection of files to which it belongs, followed by a possibly empty set of import declarations that declare packages whose contents it wishes to use, followed by a possibly empty set of declarations.
Like with a struct, a source file may contain embeddings. Unlike with a struct, the embedded expressions may be any value. If the result of the unification of all embedded values is not a struct, it will be output instead of its enclosing file when exporting CUE to a data format
SourceFile = { attribute "," } [ PackageClause "," ] { ImportDecl "," } { Declaration "," } .
"Hello \(place)!" place: "world" // Outputs "Hello world!"
A package clause is an optional clause that defines the package to which a source file the file belongs.
PackageClause = "package" PackageName . PackageName = identifier .
The PackageName must not be the blank identifier or a definition identifier.
package math
A module defines a tree of directories, rooted at the module root.
All source files within a module with the same package belong to the same package.
A module may define multiple packages.
An instance of a package is any subset of files belonging to the same package.
It is interpreted as the concatenation of these files.
An implementation may impose conventions on the layout of package files to determine which files of a package belongs to an instance. For example, an instance may be defined as the subset of package files belonging to a directory and all its ancestors.
An import declaration states that the source file containing the declaration depends on definitions of the imported package (§Program initialization and execution) and enables access to exported identifiers of that package. The import names an identifier (PackageName) to be used for access and an ImportPath that specifies the package to be imported.
ImportDecl = "import" ( ImportSpec | "(" { ImportSpec "," } ")" ) .
ImportSpec = [ PackageName ] ImportPath .
ImportLocation = { unicode_value } .
ImportPath = `"` ImportLocation [ ":" identifier ] `"` .
The PackageName is used in qualified identifiers to access exported identifiers of the package within the importing source file. It is declared in the file block. It defaults to the identifier specified in the package clause of the imported package, which must match either the last path component of ImportLocation or the identifier following it.
The interpretation of the ImportPath is implementation-dependent but it is typically either the path of a builtin package or a fully qualifying location of a package within a source code repository.
An ImportLocation must be a non-empty strings using only characters belonging Unicode‘s L, M, N, P, and S general categories (the Graphic characters without spaces) and may not include the characters !"#$%&’()*,:;<=>?[]^`{|} or the Unicode replacement character U+FFFD.
Assume we have package containing the package clause “package math”, which exports function Sin at the path identified by “lib/math”. This table illustrates how Sin is accessed in files that import the package after the various types of import declaration.
Import declaration Local name of Sin import "lib/math" math.Sin import "lib/math:math" math.Sin import m "lib/math" m.Sin
An import declaration declares a dependency relation between the importing and imported package. It is illegal for a package to import itself, directly or indirectly, or to directly import a package without referring to any of its exported identifiers.
TODO