Error.stackTraceLimit = Infinity
{Scope} = require './scope'
{isUnassignable, JS_FORBIDDEN} = require './lexer'
nodes.coffee
contains all of the node classes for the syntax tree. Most
nodes are created as the result of actions in the grammar,
but some are created by other nodes as a method of code generation. To convert
the syntax tree into a string of JavaScript code, call compile()
on the root.
Error.stackTraceLimit = Infinity
{Scope} = require './scope'
{isUnassignable, JS_FORBIDDEN} = require './lexer'
Import the helpers we plan to use.
{compact, flatten, extend, merge, del, starts, ends, some,
addDataToNode, attachCommentsToNode, locationDataToString,
throwSyntaxError, replaceUnicodeCodePointEscapes,
isFunction, isPlainObject, isNumber, parseNumber} = require './helpers'
Functions required by parser.
exports.extend = extend
exports.addDataToNode = addDataToNode
Constant functions for nodes that don’t need customization.
YES = -> yes
NO = -> no
THIS = -> this
NEGATE = -> @negated = not @negated; this
The various nodes defined below all compile to a collection of CodeFragment objects.
A CodeFragments is a block of generated code, and the location in the source file where the code
came from. CodeFragments can be assembled together into working code just by catting together
all the CodeFragments’ code
snippets, in order.
exports.CodeFragment = class CodeFragment
constructor: (parent, code) ->
@code = "#{code}"
@type = parent?.constructor?.name or 'unknown'
@locationData = parent?.locationData
@comments = parent?.comments
toString: ->
This is only intended for debugging.
"#{@code}#{if @locationData then ": " + locationDataToString(@locationData) else ''}"
Convert an array of CodeFragments into a string.
fragmentsToText = (fragments) ->
(fragment.code for fragment in fragments).join('')
The Base is the abstract base class for all nodes in the syntax tree.
Each subclass implements the compileNode
method, which performs the
code generation for that node. To compile a node to JavaScript,
call compile
on it, which wraps compileNode
in some generic extra smarts,
to know when the generated code needs to be wrapped up in a closure.
An options hash is passed and cloned throughout, containing information about
the environment from higher in the tree (such as if a returned value is
being requested by the surrounding function), information about the current
scope, and indentation level.
exports.Base = class Base
compile: (o, lvl) ->
fragmentsToText @compileToFragments o, lvl
Occasionally a node is compiled multiple times, for example to get the name
of a variable to add to scope tracking. When we know that a “premature”
compilation won’t result in comments being output, set those comments aside
so that they’re preserved for a later compile
call that will result in
the comments being included in the output.
compileWithoutComments: (o, lvl, method = 'compile') ->
if @comments
@ignoreTheseCommentsTemporarily = @comments
delete @comments
unwrapped = @unwrapAll()
if unwrapped.comments
unwrapped.ignoreTheseCommentsTemporarily = unwrapped.comments
delete unwrapped.comments
fragments = @[method] o, lvl
if @ignoreTheseCommentsTemporarily
@comments = @ignoreTheseCommentsTemporarily
delete @ignoreTheseCommentsTemporarily
if unwrapped.ignoreTheseCommentsTemporarily
unwrapped.comments = unwrapped.ignoreTheseCommentsTemporarily
delete unwrapped.ignoreTheseCommentsTemporarily
fragments
compileNodeWithoutComments: (o, lvl) ->
@compileWithoutComments o, lvl, 'compileNode'
Common logic for determining whether to wrap this node in a closure before compiling it, or to compile directly. We need to wrap if this node is a statement, and it’s not a pureStatement, and we’re not at the top level of a block (which would be unnecessary), and we haven’t already been asked to return the result (because statements know how to return results).
compileToFragments: (o, lvl) ->
o = extend {}, o
o.level = lvl if lvl
node = @unfoldSoak(o) or this
node.tab = o.indent
fragments = if o.level is LEVEL_TOP or not node.isStatement(o)
node.compileNode o
else
node.compileClosure o
@compileCommentFragments o, node, fragments
fragments
compileToFragmentsWithoutComments: (o, lvl) ->
@compileWithoutComments o, lvl, 'compileToFragments'
Statements converted into expressions via closure-wrapping share a scope object with their parent closure, to preserve the expected lexical scope.
compileClosure: (o) ->
@checkForPureStatementInExpression()
o.sharedScope = yes
func = new Code [], Block.wrap [this]
args = []
if @contains ((node) -> node instanceof SuperCall)
func.bound = yes
else if (argumentsNode = @contains isLiteralArguments) or @contains isLiteralThis
args = [new ThisLiteral]
if argumentsNode
meth = 'apply'
args.push new IdentifierLiteral 'arguments'
else
meth = 'call'
func = new Value func, [new Access new PropertyName meth]
parts = (new Call func, args).compileNode o
switch
when func.isGenerator or func.base?.isGenerator
parts.unshift @makeCode "(yield* "
parts.push @makeCode ")"
when func.isAsync or func.base?.isAsync
parts.unshift @makeCode "(await "
parts.push @makeCode ")"
parts
compileCommentFragments: (o, node, fragments) ->
return fragments unless node.comments
This is where comments, that are attached to nodes as a comments
property, become CodeFragment
s. “Inline block comments,” e.g.
/* */
-delimited comments that are interspersed within code on a line,
are added to the current fragments
stream. All other fragments are
attached as properties to the nearest preceding or following fragment,
to remain stowaways until they get properly output in compileComments
later on.
unshiftCommentFragment = (commentFragment) ->
if commentFragment.unshift
Find the first non-comment fragment and insert commentFragment
before it.
unshiftAfterComments fragments, commentFragment
else
if fragments.length isnt 0
precedingFragment = fragments[fragments.length - 1]
if commentFragment.newLine and precedingFragment.code isnt '' and
not /\n\s*$/.test precedingFragment.code
commentFragment.code = "\n#{commentFragment.code}"
fragments.push commentFragment
for comment in node.comments when comment not in @compiledComments
@compiledComments.push comment # Don’t output this comment twice.
For block/here comments, denoted by ###
, that are inline comments
like 1 + ### comment ### 2
, create fragments and insert them into
the fragments array.
Otherwise attach comment fragments to their closest fragment for now,
so they can be inserted into the output later after all the newlines
have been added.
if comment.here # Block comment, delimited by `###`.
commentFragment = new HereComment(comment).compileNode o
else # Line comment, delimited by `#`.
commentFragment = new LineComment(comment).compileNode o
if (commentFragment.isHereComment and not commentFragment.newLine) or
node.includeCommentFragments()
Inline block comments, like 1 + /* comment */ 2
, or a node whose
compileToFragments
method has logic for outputting comments.
unshiftCommentFragment commentFragment
else
fragments.push @makeCode '' if fragments.length is 0
if commentFragment.unshift
fragments[0].precedingComments ?= []
fragments[0].precedingComments.push commentFragment
else
fragments[fragments.length - 1].followingComments ?= []
fragments[fragments.length - 1].followingComments.push commentFragment
fragments
If the code generation wishes to use the result of a complex expression in multiple places, ensure that the expression is only ever evaluated once, by assigning it to a temporary variable. Pass a level to precompile.
If level
is passed, then returns [val, ref]
, where val
is the compiled value, and ref
is the compiled reference. If level
is not passed, this returns [val, ref]
where
the two values are raw nodes which have not been compiled.
cache: (o, level, shouldCache) ->
complex = if shouldCache? then shouldCache this else @shouldCache()
if complex
ref = new IdentifierLiteral o.scope.freeVariable 'ref'
sub = new Assign ref, this
if level then [sub.compileToFragments(o, level), [@makeCode(ref.value)]] else [sub, ref]
else
ref = if level then @compileToFragments o, level else this
[ref, ref]
Occasionally it may be useful to make an expression behave as if it was ‘hoisted’, whereby the result of the expression is available before its location in the source, but the expression’s variable scope corresponds to the source position. This is used extensively to deal with executable class bodies in classes.
Calling this method mutates the node, proxying the compileNode
and compileToFragments
methods to store their result for later replacing the target
node, which is returned by the
call.
hoist: ->
@hoisted = yes
target = new HoistTarget @
compileNode = @compileNode
compileToFragments = @compileToFragments
@compileNode = (o) ->
target.update compileNode, o
@compileToFragments = (o) ->
target.update compileToFragments, o
target
cacheToCodeFragments: (cacheValues) ->
[fragmentsToText(cacheValues[0]), fragmentsToText(cacheValues[1])]
Construct a node that returns the current node’s result.
Note that this is overridden for smarter behavior for
many statement nodes (e.g. If
, For
).
makeReturn: (results, mark) ->
if mark
Mark this node as implicitly returned, so that it can be part of the node metadata returned in the AST.
@canBeReturned = yes
return
node = @unwrapAll()
if results
new Call new Literal("#{results}.push"), [node]
else
new Return node
Does this node, or any of its children, contain a node of a certain kind?
Recursively traverses down the children nodes and returns the first one
that verifies pred
. Otherwise return undefined. contains
does not cross
scope boundaries.
contains: (pred) ->
node = undefined
@traverseChildren no, (n) ->
if pred n
node = n
return no
node
Pull out the last node of a node list.
lastNode: (list) ->
if list.length is 0 then null else list[list.length - 1]
Debugging representation of the node, for inspecting the parse tree.
This is what coffee --nodes
prints out.
toString: (idt = '', name = @constructor.name) ->
tree = '\n' + idt + name
tree += '?' if @soak
@eachChild (node) -> tree += node.toString idt + TAB
tree
checkForPureStatementInExpression: ->
if jumpNode = @jumps()
jumpNode.error 'cannot use a pure statement in an expression'
Plain JavaScript object representation of the node, that can be serialized
as JSON. This is what the ast
option in the Node API returns.
We try to follow the Babel AST spec
as closely as possible, for improved interoperability with other tools.
WARNING: DO NOT OVERRIDE THIS METHOD IN CHILD CLASSES.
Only override the component ast*
methods as needed.
ast: (o, level) ->
Merge level
into o
and perform other universal checks.
o = @astInitialize o, level
Create serializable representation of this node.
astNode = @astNode o
Mark AST nodes that correspond to expressions that (implicitly) return.
We can’t do this as part of astNode
because we need to assemble child
nodes first before marking the parent being returned.
if @astNode? and @canBeReturned
Object.assign astNode, {returns: yes}
astNode
astInitialize: (o, level) ->
o = Object.assign {}, o
o.level = level if level?
if o.level > LEVEL_TOP
@checkForPureStatementInExpression()
@makeReturn
must be called before astProperties
, because the latter may call
.ast()
for child nodes and those nodes would need the return logic from makeReturn
already executed by then.
@makeReturn null, yes if @isStatement(o) and o.level isnt LEVEL_TOP and o.scope?
o
astNode: (o) ->
Every abstract syntax tree node object has four categories of properties:
type
field and a string like NumberLiteral
.loc
, start
, end
and range
fields.parsedValue
.body
.
These fields are all intermixed in the Babel spec; type
and start
and
parsedValue
are all top level fields in the AST node object. We have
separate methods for returning each category, that we merge together here. Object.assign {}, {type: @astType(o)}, @astProperties(o), @astLocationData()
By default, a node class has no specific properties.
astProperties: -> {}
By default, a node class’s AST type
is its class name.
astType: -> @constructor.name
The AST location data is a rearranged version of our Jison location data, mutated into the structure that the Babel spec uses.
astLocationData: ->
jisonLocationDataToAstLocationData @locationData
Determines whether an AST node needs an ExpressionStatement
wrapper.
Typically matches our isStatement()
logic but this allows overriding.
isStatementAst: (o) ->
@isStatement o
Passes each child to a function, breaking when the function returns false
.
eachChild: (func) ->
return this unless @children
for attr in @children when @[attr]
for child in flatten [@[attr]]
return this if func(child) is false
this
traverseChildren: (crossScope, func) ->
@eachChild (child) ->
recur = func(child)
child.traverseChildren(crossScope, func) unless recur is no
replaceInContext
will traverse children looking for a node for which match
returns
true. Once found, the matching node will be replaced by the result of calling replacement
.
replaceInContext: (match, replacement) ->
return false unless @children
for attr in @children when children = @[attr]
if Array.isArray children
for child, i in children
if match child
children[i..i] = replacement child, @
return true
else
return true if child.replaceInContext match, replacement
else if match children
@[attr] = replacement children, @
return true
else
return true if children.replaceInContext match, replacement
invert: ->
new Op '!', this
unwrapAll: ->
node = this
continue until node is node = node.unwrap()
node
Default implementations of the common node properties and methods. Nodes will override these with custom logic, if needed.
children
are the properties to recurse into when tree walking. The
children
list is the structure of the AST. The parent
pointer, and
the pointer to the children
are how you can traverse the tree.
children: []
isStatement
has to do with “everything is an expression”. A few things
can’t be expressions, such as break
. Things that isStatement
returns
true
for are things that can’t be used as expressions. There are some
error messages that come from nodes.coffee
due to statements ending up
in expression position.
isStatement: NO
Track comments that have been compiled into fragments, to avoid outputting them twice.
compiledComments: []
includeCommentFragments
lets compileCommentFragments
know whether this node
has special awareness of how to handle comments within its output.
includeCommentFragments: NO
jumps
tells you if an expression, or an internal part of an expression,
has a flow control construct (like break
, continue
, or return
)
that jumps out of the normal flow of control and can’t be used as a value.
(Note that throw
is not considered a flow control construct.)
This is important because flow control in the middle of an expression
makes no sense; we have to disallow it.
jumps: NO
If node.shouldCache() is false
, it is safe to use node
more than once.
Otherwise you need to store the value of node
in a variable and output
that variable several times instead. Kind of like this: 5
need not be
cached. returnFive()
, however, could have side effects as a result of
evaluating it more than once, and therefore we need to cache it. The
parameter is named shouldCache
rather than mustCache
because there are
also cases where we might not need to cache but where we want to, for
example a long expression that may well be idempotent but we want to cache
for brevity.
shouldCache: YES
isChainable: NO
isAssignable: NO
isNumber: NO
unwrap: THIS
unfoldSoak: NO
Is this node used to assign a certain variable?
assigns: NO
For this node and all descendents, set the location data to locationData
if the location data is not already set.
updateLocationDataIfMissing: (locationData, force) ->
@forceUpdateLocation = yes if force
return this if @locationData and not @forceUpdateLocation
delete @forceUpdateLocation
@locationData = locationData
@eachChild (child) ->
child.updateLocationDataIfMissing locationData
Add location data from another node
withLocationDataFrom: ({locationData}) ->
@updateLocationDataIfMissing locationData
Add location data and comments from another node
withLocationDataAndCommentsFrom: (node) ->
@withLocationDataFrom node
{comments} = node
@comments = comments if comments?.length
this
Throw a SyntaxError associated with this node’s location.
error: (message) ->
throwSyntaxError message, @locationData
makeCode: (code) ->
new CodeFragment this, code
wrapInParentheses: (fragments) ->
[@makeCode('('), fragments..., @makeCode(')')]
wrapInBraces: (fragments) ->
[@makeCode('{'), fragments..., @makeCode('}')]
fragmentsList
is an array of arrays of fragments. Each array in fragmentsList will be
concatenated together, with joinStr
added in between each, to produce a final flat array
of fragments.
joinFragmentArrays: (fragmentsList, joinStr) ->
answer = []
for fragments, i in fragmentsList
if i then answer.push @makeCode joinStr
answer = answer.concat fragments
answer
A HoistTargetNode represents the output location in the node tree for a hoisted node. See Base#hoist.
exports.HoistTarget = class HoistTarget extends Base
Expands hoisted fragments in the given array
@expand = (fragments) ->
for fragment, i in fragments by -1 when fragment.fragments
fragments[i..i] = @expand fragment.fragments
fragments
constructor: (@source) ->
super()
Holds presentational options to apply when the source node is compiled.
@options = {}
Placeholder fragments to be replaced by the source node’s compilation.
@targetFragments = { fragments: [] }
isStatement: (o) ->
@source.isStatement o
Update the target fragments with the result of compiling the source. Calls the given compile function with the node and options (overriden with the target presentational options).
update: (compile, o) ->
@targetFragments.fragments = compile.call @source, merge o, @options
Copies the target indent and level, and returns the placeholder fragments
compileToFragments: (o, level) ->
@options.indent = o.indent
@options.level = level ? o.level
[ @targetFragments ]
compileNode: (o) ->
@compileToFragments o
compileClosure: (o) ->
@compileToFragments o
The root node of the node tree
exports.Root = class Root extends Base
constructor: (@body) ->
super()
@isAsync = (new Code [], @body).isAsync
children: ['body']
Wrap everything in a safety closure, unless requested not to. It would be better not to generate them in the first place, but for now, clean up obvious double-parentheses.
compileNode: (o) ->
o.indent = if o.bare then '' else TAB
o.level = LEVEL_TOP
o.compiling = yes
@initializeScope o
fragments = @body.compileRoot o
return fragments if o.bare
functionKeyword = "#{if @isAsync then 'async ' else ''}function"
[].concat @makeCode("(#{functionKeyword}() {\n"), fragments, @makeCode("\n}).call(this);\n")
initializeScope: (o) ->
o.scope = new Scope null, @body, null, o.referencedVars ? []
Mark given local variables in the root scope as parameters so they don’t end up being declared on the root block.
o.scope.parameter name for name in o.locals or []
commentsAst: ->
@allComments ?=
for commentToken in (@allCommentTokens ? []) when not commentToken.heregex
if commentToken.here
new HereComment commentToken
else
new LineComment commentToken
comment.ast() for comment in @allComments
astNode: (o) ->
o.level = LEVEL_TOP
@initializeScope o
super o
astType: -> 'File'
astProperties: (o) ->
@body.isRootBlock = yes
return
program: Object.assign @body.ast(o), @astLocationData()
comments: @commentsAst()
The block is the list of expressions that forms the body of an
indented block of code – the implementation of a function, a clause in an
if
, switch
, or try
, and so on…
exports.Block = class Block extends Base
constructor: (nodes) ->
super()
@expressions = compact flatten nodes or []
children: ['expressions']
Tack an expression on to the end of this expression list.
push: (node) ->
@expressions.push node
this
Remove and return the last expression of this expression list.
pop: ->
@expressions.pop()
Add an expression at the beginning of this expression list.
unshift: (node) ->
@expressions.unshift node
this
If this Block consists of just a single node, unwrap it by pulling it back out.
unwrap: ->
if @expressions.length is 1 then @expressions[0] else this
Is this an empty block of code?
isEmpty: ->
not @expressions.length
isStatement: (o) ->
for exp in @expressions when exp.isStatement o
return yes
no
jumps: (o) ->
for exp in @expressions
return jumpNode if jumpNode = exp.jumps o
A Block node does not return its entire body, rather it ensures that the final expression is returned.
makeReturn: (results, mark) ->
len = @expressions.length
[..., lastExp] = @expressions
lastExp = lastExp?.unwrap() or no
We also need to check that we’re not returning a JSX tag if there’s an adjacent one at the same level; JSX doesn’t allow that.
if lastExp and lastExp instanceof Parens and lastExp.body.expressions.length > 1
{body:{expressions}} = lastExp
[..., penult, last] = expressions
penult = penult.unwrap()
last = last.unwrap()
if penult instanceof JSXElement and last instanceof JSXElement
expressions[expressions.length - 1].error 'Adjacent JSX elements must be wrapped in an enclosing tag'
if mark
@expressions[len - 1]?.makeReturn results, mark
return
while len--
expr = @expressions[len]
@expressions[len] = expr.makeReturn results
@expressions.splice(len, 1) if expr instanceof Return and not expr.expression
break
this
compile: (o, lvl) ->
return new Root(this).withLocationDataFrom(this).compile o, lvl unless o.scope
super o, lvl
Compile all expressions within the Block body. If we need to return the result, and it’s an expression, simply return it. If it’s a statement, ask the statement to do so.
compileNode: (o) ->
@tab = o.indent
top = o.level is LEVEL_TOP
compiledNodes = []
for node, index in @expressions
if node.hoisted
This is a hoisted expression. We want to compile this and ignore the result.
node.compileToFragments o
continue
node = (node.unfoldSoak(o) or node)
if node instanceof Block
This is a nested block. We don’t do anything special here like enclose it in a new scope; we just compile the statements in this block along with our own.
compiledNodes.push node.compileNode o
else if top
node.front = yes
fragments = node.compileToFragments o
unless node.isStatement o
fragments = indentInitial fragments, @
[..., lastFragment] = fragments
unless lastFragment.code is '' or lastFragment.isComment
fragments.push @makeCode ';'
compiledNodes.push fragments
else
compiledNodes.push node.compileToFragments o, LEVEL_LIST
if top
if @spaced
return [].concat @joinFragmentArrays(compiledNodes, '\n\n'), @makeCode('\n')
else
return @joinFragmentArrays(compiledNodes, '\n')
if compiledNodes.length
answer = @joinFragmentArrays(compiledNodes, ', ')
else
answer = [@makeCode 'void 0']
if compiledNodes.length > 1 and o.level >= LEVEL_LIST then @wrapInParentheses answer else answer
compileRoot: (o) ->
@spaced = yes
fragments = @compileWithDeclarations o
HoistTarget.expand fragments
@compileComments fragments
Compile the expressions body for the contents of a function, with declarations of all inner variables pushed up to the top.
compileWithDeclarations: (o) ->
fragments = []
post = []
for exp, i in @expressions
exp = exp.unwrap()
break unless exp instanceof Literal
o = merge(o, level: LEVEL_TOP)
if i
rest = @expressions.splice i, 9e9
[spaced, @spaced] = [@spaced, no]
[fragments, @spaced] = [@compileNode(o), spaced]
@expressions = rest
post = @compileNode o
{scope} = o
if scope.expressions is this
declars = o.scope.hasDeclarations()
assigns = scope.hasAssignments
if declars or assigns
fragments.push @makeCode '\n' if i
fragments.push @makeCode "#{@tab}var "
if declars
declaredVariables = scope.declaredVariables()
for declaredVariable, declaredVariablesIndex in declaredVariables
fragments.push @makeCode declaredVariable
if Object::hasOwnProperty.call o.scope.comments, declaredVariable
fragments.push o.scope.comments[declaredVariable]...
if declaredVariablesIndex isnt declaredVariables.length - 1
fragments.push @makeCode ', '
if assigns
fragments.push @makeCode ",\n#{@tab + TAB}" if declars
fragments.push @makeCode scope.assignedVariables().join(",\n#{@tab + TAB}")
fragments.push @makeCode ";\n#{if @spaced then '\n' else ''}"
else if fragments.length and post.length
fragments.push @makeCode "\n"
fragments.concat post
compileComments: (fragments) ->
for fragment, fragmentIndex in fragments
Insert comments into the output at the next or previous newline. If there are no newlines at which to place comments, create them.
if fragment.precedingComments
Determine the indentation level of the fragment that we are about
to insert comments before, and use that indentation level for our
inserted comments. At this point, the fragments’ code
property
is the generated output JavaScript, and CoffeeScript always
generates output indented by two spaces; so all we need to do is
search for a code
property that begins with at least two spaces.
fragmentIndent = ''
for pastFragment in fragments[0...(fragmentIndex + 1)] by -1
indent = /^ {2,}/m.exec pastFragment.code
if indent
fragmentIndent = indent[0]
break
else if '\n' in pastFragment.code
break
code = "\n#{fragmentIndent}" + (
for commentFragment in fragment.precedingComments
if commentFragment.isHereComment and commentFragment.multiline
multident commentFragment.code, fragmentIndent, no
else
commentFragment.code
).join("\n#{fragmentIndent}").replace /^(\s*)$/gm, ''
for pastFragment, pastFragmentIndex in fragments[0...(fragmentIndex + 1)] by -1
newLineIndex = pastFragment.code.lastIndexOf '\n'
if newLineIndex is -1
Keep searching previous fragments until we can’t go back any further, either because there are no fragments left or we’ve discovered that we’re in a code block that is interpolated inside a string.
if pastFragmentIndex is 0
pastFragment.code = '\n' + pastFragment.code
newLineIndex = 0
else if pastFragment.isStringWithInterpolations and pastFragment.code is '{'
code = code[1..] + '\n' # Move newline to end.
newLineIndex = 1
else
continue
delete fragment.precedingComments
pastFragment.code = pastFragment.code[0...newLineIndex] +
code + pastFragment.code[newLineIndex..]
break
Yes, this is awfully similar to the previous if
block, but if you
look closely you’ll find lots of tiny differences that make this
confusing if it were abstracted into a function that both blocks share.
if fragment.followingComments
Does the first trailing comment follow at the end of a line of code,
like ; // Comment
, or does it start a new line after a line of code?
trail = fragment.followingComments[0].trail
fragmentIndent = ''
Find the indent of the next line of code, if we have any non-trailing comments to output. We need to first find the next newline, as these comments will be output after that; and then the indent of the line that follows the next newline.
unless trail and fragment.followingComments.length is 1
onNextLine = no
for upcomingFragment in fragments[fragmentIndex...]
unless onNextLine
if '\n' in upcomingFragment.code
onNextLine = yes
else
continue
else
indent = /^ {2,}/m.exec upcomingFragment.code
if indent
fragmentIndent = indent[0]
break
else if '\n' in upcomingFragment.code
break
Is this comment following the indent inserted by bare mode? If so, there’s no need to indent this further.
code = if fragmentIndex is 1 and /^\s+$/.test fragments[0].code
''
else if trail
' '
else
"\n#{fragmentIndent}"
Assemble properly indented comments.
code += (
for commentFragment in fragment.followingComments
if commentFragment.isHereComment and commentFragment.multiline
multident commentFragment.code, fragmentIndent, no
else
commentFragment.code
).join("\n#{fragmentIndent}").replace /^(\s*)$/gm, ''
for upcomingFragment, upcomingFragmentIndex in fragments[fragmentIndex...]
newLineIndex = upcomingFragment.code.indexOf '\n'
if newLineIndex is -1
Keep searching upcoming fragments until we can’t go any further, either because there are no fragments left or we’ve discovered that we’re in a code block that is interpolated inside a string.
if upcomingFragmentIndex is fragments.length - 1
upcomingFragment.code = upcomingFragment.code + '\n'
newLineIndex = upcomingFragment.code.length
else if upcomingFragment.isStringWithInterpolations and upcomingFragment.code is '}'
code = "#{code}\n"
newLineIndex = 0
else
continue
delete fragment.followingComments
Avoid inserting extra blank lines.
code = code.replace /^\n/, '' if upcomingFragment.code is '\n'
upcomingFragment.code = upcomingFragment.code[0...newLineIndex] +
code + upcomingFragment.code[newLineIndex..]
break
fragments
Wrap up the given nodes as a Block, unless it already happens to be one.
@wrap: (nodes) ->
return nodes[0] if nodes.length is 1 and nodes[0] instanceof Block
new Block nodes
astNode: (o) ->
if (o.level? and o.level isnt LEVEL_TOP) and @expressions.length
return (new Sequence(@expressions).withLocationDataFrom @).ast o
super o
astType: ->
if @isRootBlock
'Program'
else if @isClassBody
'ClassBody'
else
'BlockStatement'
astProperties: (o) ->
checkForDirectives = del o, 'checkForDirectives'
sniffDirectives @expressions, notFinalExpression: checkForDirectives if @isRootBlock or checkForDirectives
directives = []
body = []
for expression in @expressions
expressionAst = expression.ast o
Ignore generated PassthroughLiteral
if not expressionAst?
continue
else if expression instanceof Directive
directives.push expressionAst
If an expression is a statement, it can be added to the body as is.
else if expression.isStatementAst o
body.push expressionAst
Otherwise, we need to wrap it in an ExpressionStatement
AST node.
else
body.push Object.assign
type: 'ExpressionStatement'
expression: expressionAst
,
expression.astLocationData()
return {
For now, we’re not including sourceType
on the Program
AST node.
Its value could be either 'script'
or 'module'
, and there’s no way
for CoffeeScript to always know which it should be. The presence of an
import
or export
statement in source code would imply that it should
be a module
, but a project may consist of mostly such files and also
an outlier file that lacks import
or export
but is still imported
into the project and therefore expects to be treated as a module
.
Determining the value of sourceType
is essentially the same challenge
posed by determining the parse goal of a JavaScript file, also module
or script
, and so if Node figures out a way to do so for .js
files
then CoffeeScript can copy Node’s algorithm.
sourceType: ‘module’
body, directives
}
astLocationData: ->
return if @isRootBlock and not @locationData?
super()
A directive e.g. ‘use strict’. Currently only used during AST generation.
exports.Directive = class Directive extends Base
constructor: (@value) ->
super()
astProperties: (o) ->
return
value: Object.assign {},
@value.ast o
type: 'DirectiveLiteral'
Literal
is a base class for static values that can be passed through
directly into JavaScript without translation, such as: strings, numbers,
true
, false
, null
…
exports.Literal = class Literal extends Base
constructor: (@value) ->
super()
shouldCache: NO
assigns: (name) ->
name is @value
compileNode: (o) ->
[@makeCode @value]
astProperties: ->
return
value: @value
toString: ->
This is only intended for debugging.
" #{if @isStatement() then super() else @constructor.name}: #{@value}"
exports.NumberLiteral = class NumberLiteral extends Literal
constructor: (@value, {@parsedValue} = {}) ->
super()
unless @parsedValue?
if isNumber @value
@parsedValue = @value
@value = "#{@value}"
else
@parsedValue = parseNumber @value
isBigInt: ->
/n$/.test @value
astType: ->
if @isBigInt()
'BigIntLiteral'
else
'NumericLiteral'
astProperties: ->
return
value:
if @isBigInt()
@parsedValue.toString()
else
@parsedValue
extra:
rawValue:
if @isBigInt()
@parsedValue.toString()
else
@parsedValue
raw: @value
exports.InfinityLiteral = class InfinityLiteral extends NumberLiteral
constructor: (@value, {@originalValue = 'Infinity'} = {}) ->
super()
compileNode: ->
[@makeCode '2e308']
astNode: (o) ->
unless @originalValue is 'Infinity'
return new NumberLiteral(@value).withLocationDataFrom(@).ast o
super o
astType: -> 'Identifier'
astProperties: ->
return
name: 'Infinity'
declaration: no
exports.NaNLiteral = class NaNLiteral extends NumberLiteral
constructor: ->
super 'NaN'
compileNode: (o) ->
code = [@makeCode '0/0']
if o.level >= LEVEL_OP then @wrapInParentheses code else code
astType: -> 'Identifier'
astProperties: ->
return
name: 'NaN'
declaration: no
exports.StringLiteral = class StringLiteral extends Literal
constructor: (@originalValue, {@quote, @initialChunk, @finalChunk, @indent, @double, @heregex} = {}) ->
super ''
@quote = null if @quote is '///'
@fromSourceString = @quote?
@quote ?= '"'
heredoc = @isFromHeredoc()
val = @originalValue
if @heregex
val = val.replace HEREGEX_OMIT, '$1$2'
val = replaceUnicodeCodePointEscapes val, flags: @heregex.flags
else
val = val.replace STRING_OMIT, '$1'
val =
unless @fromSourceString
val
else if heredoc
indentRegex = /// \n#{@indent} ///g if @indent
val = val.replace indentRegex, '\n' if indentRegex
val = val.replace LEADING_BLANK_LINE, '' if @initialChunk
val = val.replace TRAILING_BLANK_LINE, '' if @finalChunk
val
else
val.replace SIMPLE_STRING_OMIT, (match, offset) =>
if (@initialChunk and offset is 0) or
(@finalChunk and offset + match.length is val.length)
''
else
' '
@delimiter = @quote.charAt 0
@value = makeDelimitedLiteral val, {
@delimiter
@double
}
@unquotedValueForTemplateLiteral = makeDelimitedLiteral val, {
delimiter: '`'
@double
escapeNewlines: no
includeDelimiters: no
convertTrailingNullEscapes: yes
}
@unquotedValueForJSX = makeDelimitedLiteral val, {
@double
escapeNewlines: no
includeDelimiters: no
escapeDelimiter: no
}
compileNode: (o) ->
return StringWithInterpolations.fromStringLiteral(@).compileNode o if @shouldGenerateTemplateLiteral()
return [@makeCode @unquotedValueForJSX] if @jsx
super o
StringLiteral
s can represent either entire literal strings
or pieces of text inside of e.g. an interpolated string.
When parsed as the former but needing to be treated as the latter
(e.g. the string part of a tagged template literal), this will return
a copy of the StringLiteral
with the quotes trimmed from its location
data (like it would have if parsed as part of an interpolated string).
withoutQuotesInLocationData: ->
endsWithNewline = @originalValue[-1..] is '\n'
locationData = Object.assign {}, @locationData
locationData.first_column += @quote.length
if endsWithNewline
locationData.last_line -= 1
locationData.last_column =
if locationData.last_line is locationData.first_line
locationData.first_column + @originalValue.length - '\n'.length
else
@originalValue[...-1].length - '\n'.length - @originalValue[...-1].lastIndexOf('\n')
else
locationData.last_column -= @quote.length
locationData.last_column_exclusive -= @quote.length
locationData.range = [
locationData.range[0] + @quote.length
locationData.range[1] - @quote.length
]
copy = new StringLiteral @originalValue, {@quote, @initialChunk, @finalChunk, @indent, @double, @heregex}
copy.locationData = locationData
copy
isFromHeredoc: ->
@quote.length is 3
shouldGenerateTemplateLiteral: ->
@isFromHeredoc()
astNode: (o) ->
return StringWithInterpolations.fromStringLiteral(@).ast o if @shouldGenerateTemplateLiteral()
super o
astProperties: ->
return
value: @originalValue
extra:
raw: "#{@delimiter}#{@originalValue}#{@delimiter}"
exports.RegexLiteral = class RegexLiteral extends Literal
constructor: (value, {@delimiter = '/', @heregexCommentTokens = []} = {}) ->
super ''
heregex = @delimiter is '///'
endDelimiterIndex = value.lastIndexOf '/'
@flags = value[endDelimiterIndex + 1..]
val = @originalValue = value[1...endDelimiterIndex]
val = val.replace HEREGEX_OMIT, '$1$2' if heregex
val = replaceUnicodeCodePointEscapes val, {@flags}
@value = "#{makeDelimitedLiteral val, delimiter: '/'}#{@flags}"
REGEX_REGEX: /// ^ / (.*) / \w* $ ///
astType: -> 'RegExpLiteral'
astProperties: (o) ->
[, pattern] = @REGEX_REGEX.exec @value
return {
value: undefined
pattern, @flags, @delimiter
originalPattern: @originalValue
extra:
raw: @value
originalRaw: "#{@delimiter}#{@originalValue}#{@delimiter}#{@flags}"
rawValue: undefined
comments:
for heregexCommentToken in @heregexCommentTokens
if heregexCommentToken.here
new HereComment(heregexCommentToken).ast o
else
new LineComment(heregexCommentToken).ast o
}
exports.PassthroughLiteral = class PassthroughLiteral extends Literal
constructor: (@originalValue, {@here, @generated} = {}) ->
super ''
@value = @originalValue.replace /\\+(`|$)/g, (string) ->
string
is always a value like ‘`‘, ‘\`‘, ‘\\`‘, etc.
By reducing it to its latter half, we turn ‘`‘ to ‘', '\\\
‘ to ‘`‘, etc.
string[-Math.ceil(string.length / 2)..]
astNode: (o) ->
return null if @generated
super o
astProperties: ->
return {
value: @originalValue
here: !!@here
}
exports.IdentifierLiteral = class IdentifierLiteral extends Literal
isAssignable: YES
eachName: (iterator) ->
iterator @
astType: ->
if @jsx
'JSXIdentifier'
else
'Identifier'
astProperties: ->
return
name: @value
declaration: !!@isDeclaration
exports.PropertyName = class PropertyName extends Literal
isAssignable: YES
astType: ->
if @jsx
'JSXIdentifier'
else
'Identifier'
astProperties: ->
return
name: @value
declaration: no
exports.ComputedPropertyName = class ComputedPropertyName extends PropertyName
compileNode: (o) ->
[@makeCode('['), @value.compileToFragments(o, LEVEL_LIST)..., @makeCode(']')]
astNode: (o) ->
@value.ast o
exports.StatementLiteral = class StatementLiteral extends Literal
isStatement: YES
makeReturn: THIS
jumps: (o) ->
return this if @value is 'break' and not (o?.loop or o?.block)
return this if @value is 'continue' and not o?.loop
compileNode: (o) ->
[@makeCode "#{@tab}#{@value};"]
astType: ->
switch @value
when 'continue' then 'ContinueStatement'
when 'break' then 'BreakStatement'
when 'debugger' then 'DebuggerStatement'
exports.ThisLiteral = class ThisLiteral extends Literal
constructor: (value) ->
super 'this'
@shorthand = value is '@'
compileNode: (o) ->
code = if o.scope.method?.bound then o.scope.method.context else @value
[@makeCode code]
astType: -> 'ThisExpression'
astProperties: ->
return
shorthand: @shorthand
exports.UndefinedLiteral = class UndefinedLiteral extends Literal
constructor: ->
super 'undefined'
compileNode: (o) ->
[@makeCode if o.level >= LEVEL_ACCESS then '(void 0)' else 'void 0']
astType: -> 'Identifier'
astProperties: ->
return
name: @value
declaration: no
exports.NullLiteral = class NullLiteral extends Literal
constructor: ->
super 'null'
exports.BooleanLiteral = class BooleanLiteral extends Literal
constructor: (value, {@originalValue} = {}) ->
super value
@originalValue ?= @value
astProperties: ->
value: if @value is 'true' then yes else no
name: @originalValue
exports.DefaultLiteral = class DefaultLiteral extends Literal
astType: -> 'Identifier'
astProperties: ->
return
name: 'default'
declaration: no
A return
is a pureStatement—wrapping it in a closure wouldn’t make sense.
exports.Return = class Return extends Base
constructor: (@expression, {@belongsToFuncDirectiveReturn} = {}) ->
super()
children: ['expression']
isStatement: YES
makeReturn: THIS
jumps: THIS
compileToFragments: (o, level) ->
expr = @expression?.makeReturn()
if expr and expr not instanceof Return then expr.compileToFragments o, level else super o, level
compileNode: (o) ->
answer = []
TODO: If we call expression.compile()
here twice, we’ll sometimes
get back different results!
if @expression
answer = @expression.compileToFragments o, LEVEL_PAREN
unshiftAfterComments answer, @makeCode "#{@tab}return "
Since the return
got indented by @tab
, preceding comments that are
multiline need to be indented.
for fragment in answer
if fragment.isHereComment and '\n' in fragment.code
fragment.code = multident fragment.code, @tab
else if fragment.isLineComment
fragment.code = "#{@tab}#{fragment.code}"
else
break
else
answer.push @makeCode "#{@tab}return"
answer.push @makeCode ';'
answer
checkForPureStatementInExpression: ->
don’t flag return
from await return
/yield return
as invalid.
return if @belongsToFuncDirectiveReturn
super()
astType: -> 'ReturnStatement'
astProperties: (o) ->
argument: @expression?.ast(o, LEVEL_PAREN) ? null
Parent class for YieldReturn
/AwaitReturn
.
exports.FuncDirectiveReturn = class FuncDirectiveReturn extends Return
constructor: (expression, {@returnKeyword}) ->
super expression
compileNode: (o) ->
@checkScope o
super o
checkScope: (o) ->
unless o.scope.parent?
@error "#{@keyword} can only occur inside functions"
isStatementAst: NO
astNode: (o) ->
@checkScope o
new Op @keyword,
new Return @expression, belongsToFuncDirectiveReturn: yes
.withLocationDataFrom(
if @expression?
locationData: mergeLocationData @returnKeyword.locationData, @expression.locationData
else
@returnKeyword
)
.withLocationDataFrom @
.ast o
yield return
works exactly like return
, except that it turns the function
into a generator.
exports.YieldReturn = class YieldReturn extends FuncDirectiveReturn
keyword: 'yield'
exports.AwaitReturn = class AwaitReturn extends FuncDirectiveReturn
keyword: 'await'
A value, variable or literal or parenthesized, indexed or dotted into, or vanilla.
exports.Value = class Value extends Base
constructor: (base, props, tag, isDefaultValue = no) ->
super()
return base if not props and base instanceof Value
@base = base
@properties = props or []
@tag = tag
@[tag] = yes if tag
@isDefaultValue = isDefaultValue
If this is a @foo =
assignment, if there are comments on @
move them
to be on foo
.
if @base?.comments and @base instanceof ThisLiteral and @properties[0]?.name?
moveComments @base, @properties[0].name
children: ['base', 'properties']
Add a property (or properties ) Access
to the list.
add: (props) ->
@properties = @properties.concat props
@forceUpdateLocation = yes
this
hasProperties: ->
@properties.length isnt 0
bareLiteral: (type) ->
not @properties.length and @base instanceof type
Some boolean checks for the benefit of other nodes.
isArray : -> @bareLiteral(Arr)
isRange : -> @bareLiteral(Range)
shouldCache : -> @hasProperties() or @base.shouldCache()
isAssignable : (opts) -> @hasProperties() or @base.isAssignable opts
isNumber : -> @bareLiteral(NumberLiteral)
isString : -> @bareLiteral(StringLiteral)
isRegex : -> @bareLiteral(RegexLiteral)
isUndefined : -> @bareLiteral(UndefinedLiteral)
isNull : -> @bareLiteral(NullLiteral)
isBoolean : -> @bareLiteral(BooleanLiteral)
isAtomic : ->
for node in @properties.concat @base
return no if node.soak or node instanceof Call or node instanceof Op and node.operator is 'do'
yes
isNotCallable : -> @isNumber() or @isString() or @isRegex() or
@isArray() or @isRange() or @isSplice() or @isObject() or
@isUndefined() or @isNull() or @isBoolean()
isStatement : (o) -> not @properties.length and @base.isStatement o
isJSXTag : -> @base instanceof JSXTag
assigns : (name) -> not @properties.length and @base.assigns name
jumps : (o) -> not @properties.length and @base.jumps o
isObject: (onlyGenerated) ->
return no if @properties.length
(@base instanceof Obj) and (not onlyGenerated or @base.generated)
isElision: ->
return no unless @base instanceof Arr
@base.hasElision()
isSplice: ->
[..., lastProperty] = @properties
lastProperty instanceof Slice
looksStatic: (className) ->
return no unless ((thisLiteral = @base) instanceof ThisLiteral or (name = @base).value is className) and
@properties.length is 1 and @properties[0].name?.value isnt 'prototype'
return
staticClassName: thisLiteral ? name
The value can be unwrapped as its inner node, if there are no attached properties.
unwrap: ->
if @properties.length then this else @base
A reference has base part (this
value) and name part.
We cache them separately for compiling complex expressions.
a()[b()] ?= c
-> (_base = a())[_name = b()] ? _base[_name] = c
cacheReference: (o) ->
[..., name] = @properties
if @properties.length < 2 and not @base.shouldCache() and not name?.shouldCache()
return [this, this] # `a` `a.b`
base = new Value @base, @properties[...-1]
if base.shouldCache() # `a().b`
bref = new IdentifierLiteral o.scope.freeVariable 'base'
base = new Value new Parens new Assign bref, base
return [base, bref] unless name # `a()`
if name.shouldCache() # `a[b()]`
nref = new IdentifierLiteral o.scope.freeVariable 'name'
name = new Index new Assign nref, name.index
nref = new Index nref
[base.add(name), new Value(bref or base.base, [nref or name])]
We compile a value to JavaScript by compiling and joining each property.
Things get much more interesting if the chain of properties has soak
operators ?.
interspersed. Then we have to take care not to accidentally
evaluate anything twice when building the soak chain.
compileNode: (o) ->
@base.front = @front
props = @properties
if props.length and @base.cached?
Cached fragments enable correct order of the compilation,
and reuse of variables in the scope.
Example:
a(x = 5).b(-> x = 6)
should compile in the same order as
a(x = 5); b(-> x = 6)
(see issue #4437, https://github.com/jashkenas/coffeescript/issues/4437)
fragments = @base.cached
else
fragments = @base.compileToFragments o, (if props.length then LEVEL_ACCESS else null)
if props.length and SIMPLENUM.test fragmentsToText fragments
fragments.push @makeCode '.'
for prop in props
fragments.push (prop.compileToFragments o)...
fragments
Unfold a soak into an If
: a?.b
-> a.b if a?
unfoldSoak: (o) ->
@unfoldedSoak ?= do =>
ifn = @base.unfoldSoak o
if ifn
ifn.body.properties.push @properties...
return ifn
for prop, i in @properties when prop.soak
prop.soak = off
fst = new Value @base, @properties[...i]
snd = new Value @base, @properties[i..]
if fst.shouldCache()
ref = new IdentifierLiteral o.scope.freeVariable 'ref'
fst = new Parens new Assign ref, fst
snd.base = ref
return new If new Existence(fst), snd, soak: on
no
eachName: (iterator, {checkAssignability = yes} = {}) ->
if @hasProperties()
iterator @
else if not checkAssignability or @base.isAssignable()
@base.eachName iterator
else
@error 'tried to assign to unassignable value'
For AST generation, we need an object
that’s this Value
minus its last
property, if it has properties.
object: ->
return @ unless @hasProperties()
Get all properties except the last one; for a Value
with only one
property, initialProperties
is an empty array.
initialProperties = @properties[0.[email protected] - 1]
Create the object
that becomes the new “base” for the split-off final
property.
object = new Value @base, initialProperties, @tag, @isDefaultValue
Add location data to our new node, so that it has correct location data for source maps or later conversion into AST location data.
object.locationData =
if initialProperties.length is 0
This new Value
has only one property, so the location data is just
that of the parent Value
’s base.
@base.locationData
else
This new Value
has multiple properties, so the location data spans
from the parent Value
’s base to the last property that’s included
in this new node (a.k.a. the second-to-last property of the parent).
mergeLocationData @base.locationData, initialProperties[initialProperties.length - 1].locationData
object
containsSoak: ->
return no unless @hasProperties()
for property in @properties when property.soak
return yes
return yes if @base instanceof Call and @base.soak
no
astNode: (o) ->
If the Value
has no properties, the AST node is just whatever this
node’s base
is.
return @base.ast o unless @hasProperties()
Otherwise, call Base::ast
which in turn calls the astType
and
astProperties
methods below.
super o
astType: ->
if @isJSXTag()
'JSXMemberExpression'
else if @containsSoak()
'OptionalMemberExpression'
else
'MemberExpression'
If this Value
has properties, the last property (e.g. c
in a.b.c
)
becomes the property
, and the preceding properties (e.g. a.b
) become
a child Value
node assigned to the object
property.
astProperties: (o) ->
[..., property] = @properties
property.name.jsx = yes if @isJSXTag()
computed = property instanceof Index or property.name?.unwrap() not instanceof PropertyName
return {
object: @object().ast o, LEVEL_ACCESS
property: property.ast o, (LEVEL_PAREN if computed)
computed
optional: !!property.soak
shorthand: !!property.shorthand
}
astLocationData: ->
return super() unless @isJSXTag()
don’t include leading < of JSX tag in location data
mergeAstLocationData(
jisonLocationDataToAstLocationData(@base.tagNameLocationData),
jisonLocationDataToAstLocationData(@properties[@properties.length - 1].locationData)
)
exports.MetaProperty = class MetaProperty extends Base
constructor: (@meta, @property) ->
super()
children: ['meta', 'property']
checkValid: (o) ->
if @meta.value is 'new'
if @property instanceof Access and @property.name.value is 'target'
unless o.scope.parent?
@error "new.target can only occur inside functions"
else
@error "the only valid meta property for new is new.target"
else if @meta.value is 'import'
unless @property instanceof Access and @property.name.value is 'meta'
@error "the only valid meta property for import is import.meta"
compileNode: (o) ->
@checkValid o
fragments = []
fragments.push @meta.compileToFragments(o, LEVEL_ACCESS)...
fragments.push @property.compileToFragments(o)...
fragments
astProperties: (o) ->
@checkValid o
return
meta: @meta.ast o, LEVEL_ACCESS
property: @property.ast o
Comment delimited by ###
(becoming /* */
).
exports.HereComment = class HereComment extends Base
constructor: ({ @content, @newLine, @unshift, @locationData }) ->
super()
compileNode: (o) ->
multiline = '\n' in @content
Unindent multiline comments. They will be reindented later.
if multiline
indent = null
for line in @content.split '\n'
leadingWhitespace = /^\s*/.exec(line)[0]
if not indent or leadingWhitespace.length < indent.length
indent = leadingWhitespace
@content = @content.replace /// \n #{indent} ///g, '\n' if indent
hasLeadingMarks = /\n\s*[#|\*]/.test @content
@content = @content.replace /^([ \t]*)#(?=\s)/gm, ' *' if hasLeadingMarks
@content = "/*#{@content}#{if hasLeadingMarks then ' ' else ''}*/"
fragment = @makeCode @content
fragment.newLine = @newLine
fragment.unshift = @unshift
fragment.multiline = multiline
Don’t rely on fragment.type
, which can break when the compiler is minified.
fragment.isComment = fragment.isHereComment = yes
fragment
astType: -> 'CommentBlock'
astProperties: ->
return
value: @content
Comment running from #
to the end of a line (becoming //
).
exports.LineComment = class LineComment extends Base
constructor: ({ @content, @newLine, @unshift, @locationData, @precededByBlankLine }) ->
super()
compileNode: (o) ->
fragment = @makeCode(if /^\s*$/.test @content then '' else "#{if @precededByBlankLine then "\n#{o.indent}" else ''}//#{@content}")
fragment.newLine = @newLine
fragment.unshift = @unshift
fragment.trail = not @newLine and not @unshift
Don’t rely on fragment.type
, which can break when the compiler is minified.
fragment.isComment = fragment.isLineComment = yes
fragment
astType: -> 'CommentLine'
astProperties: ->
return
value: @content
exports.JSXIdentifier = class JSXIdentifier extends IdentifierLiteral
astType: -> 'JSXIdentifier'
exports.JSXTag = class JSXTag extends JSXIdentifier
constructor: (value, {
@tagNameLocationData
@closingTagOpeningBracketLocationData
@closingTagSlashLocationData
@closingTagNameLocationData
@closingTagClosingBracketLocationData
}) ->
super value
astProperties: ->
return
name: @value
exports.JSXExpressionContainer = class JSXExpressionContainer extends Base
constructor: (@expression, {locationData} = {}) ->
super()
@expression.jsxAttribute = yes
@locationData = locationData ? @expression.locationData
children: ['expression']
compileNode: (o) ->
@expression.compileNode(o)
astProperties: (o) ->
return
expression: astAsBlockIfNeeded @expression, o
exports.JSXEmptyExpression = class JSXEmptyExpression extends Base
exports.JSXText = class JSXText extends Base
constructor: (stringLiteral) ->
super()
@value = stringLiteral.unquotedValueForJSX
@locationData = stringLiteral.locationData
astProperties: ->
return {
@value
extra:
raw: @value
}
exports.JSXAttribute = class JSXAttribute extends Base
constructor: ({@name, value}) ->
super()
@value =
if value?
value = value.base
if value instanceof StringLiteral and not value.shouldGenerateTemplateLiteral()
value
else
new JSXExpressionContainer value
else
null
@value?.comments = value.comments
children: ['name', 'value']
compileNode: (o) ->
compiledName = @name.compileToFragments o, LEVEL_LIST
return compiledName unless @value?
val = @value.compileToFragments o, LEVEL_LIST
compiledName.concat @makeCode('='), val
astProperties: (o) ->
name = @name
if ':' in name.value
name = new JSXNamespacedName name
return
name: name.ast o
value: @value?.ast(o) ? null
exports.JSXAttributes = class JSXAttributes extends Base
constructor: (arr) ->
super()
@attributes = []
for object in arr.objects
@checkValidAttribute object
{base} = object
if base instanceof IdentifierLiteral
attribute with no value eg disabled
attribute = new JSXAttribute name: new JSXIdentifier(base.value).withLocationDataAndCommentsFrom base
attribute.locationData = base.locationData
@attributes.push attribute
else if not base.generated
object spread attribute eg {…props}
attribute = base.properties[0]
attribute.jsx = yes
attribute.locationData = base.locationData
@attributes.push attribute
else
Obj containing attributes with values eg a=”b” c={d}
for property in base.properties
{variable, value} = property
attribute = new JSXAttribute {
name: new JSXIdentifier(variable.base.value).withLocationDataAndCommentsFrom variable.base
value
}
attribute.locationData = property.locationData
@attributes.push attribute
@locationData = arr.locationData
children: ['attributes']
Catch invalid attributes: <div {a:”b”, props} {props} “value” />
checkValidAttribute: (object) ->
{base: attribute} = object
properties = attribute?.properties or []
if not (attribute instanceof Obj or attribute instanceof IdentifierLiteral) or (attribute instanceof Obj and not attribute.generated and (properties.length > 1 or not (properties[0] instanceof Splat)))
object.error """
Unexpected token. Allowed JSX attributes are: id="val", src={source}, {props...} or attribute.
"""
compileNode: (o) ->
fragments = []
for attribute in @attributes
fragments.push @makeCode ' '
fragments.push attribute.compileToFragments(o, LEVEL_TOP)...
fragments
astNode: (o) ->
attribute.ast(o) for attribute in @attributes
exports.JSXNamespacedName = class JSXNamespacedName extends Base
constructor: (tag) ->
super()
[namespace, name] = tag.value.split ':'
@namespace = new JSXIdentifier(namespace).withLocationDataFrom locationData: extractSameLineLocationDataFirst(namespace.length) tag.locationData
@name = new JSXIdentifier(name ).withLocationDataFrom locationData: extractSameLineLocationDataLast(name.length ) tag.locationData
@locationData = tag.locationData
children: ['namespace', 'name']
astProperties: (o) ->
return
namespace: @namespace.ast o
name: @name.ast o
Node for a JSX element
exports.JSXElement = class JSXElement extends Base
constructor: ({@tagName, @attributes, @content}) ->
super()
children: ['tagName', 'attributes', 'content']
compileNode: (o) ->
@content?.base.jsx = yes
fragments = [@makeCode('<')]
fragments.push (tag = @tagName.compileToFragments(o, LEVEL_ACCESS))...
fragments.push @attributes.compileToFragments(o)...
if @content
fragments.push @makeCode('>')
fragments.push @content.compileNode(o, LEVEL_LIST)...
fragments.push [@makeCode('</'), tag..., @makeCode('>')]...
else
fragments.push @makeCode(' />')
fragments
isFragment: ->
[email protected]
astNode: (o) ->
The location data spanning the opening element < … > is captured by the generated Arr which contains the element’s attributes
@openingElementLocationData = jisonLocationDataToAstLocationData @attributes.locationData
tagName = @tagName.base
tagName.locationData = tagName.tagNameLocationData
if @content?
@closingElementLocationData = mergeAstLocationData(
jisonLocationDataToAstLocationData tagName.closingTagOpeningBracketLocationData
jisonLocationDataToAstLocationData tagName.closingTagClosingBracketLocationData
)
super o
astType: ->
if @isFragment()
'JSXFragment'
else
'JSXElement'
elementAstProperties: (o) ->
tagNameAst = =>
tag = @tagName.unwrap()
if tag?.value and ':' in tag.value
tag = new JSXNamespacedName tag
tag.ast o
openingElement = Object.assign {
type: 'JSXOpeningElement'
name: tagNameAst()
selfClosing: not @closingElementLocationData?
attributes: @attributes.ast o
}, @openingElementLocationData
closingElement = null
if @closingElementLocationData?
closingElement = Object.assign {
type: 'JSXClosingElement'
name: Object.assign(
tagNameAst(),
jisonLocationDataToAstLocationData @tagName.base.closingTagNameLocationData
)
}, @closingElementLocationData
if closingElement.name.type in ['JSXMemberExpression', 'JSXNamespacedName']
rangeDiff = closingElement.range[0] - openingElement.range[0] + '/'.length
columnDiff = closingElement.loc.start.column - openingElement.loc.start.column + '/'.length
shiftAstLocationData = (node) =>
node.range = [
node.range[0] + rangeDiff
node.range[1] + rangeDiff
]
node.start += rangeDiff
node.end += rangeDiff
node.loc.start =
line: @closingElementLocationData.loc.start.line
column: node.loc.start.column + columnDiff
node.loc.end =
line: @closingElementLocationData.loc.start.line
column: node.loc.end.column + columnDiff
if closingElement.name.type is 'JSXMemberExpression'
currentExpr = closingElement.name
while currentExpr.type is 'JSXMemberExpression'
shiftAstLocationData currentExpr unless currentExpr is closingElement.name
shiftAstLocationData currentExpr.property
currentExpr = currentExpr.object
shiftAstLocationData currentExpr
else # JSXNamespacedName
shiftAstLocationData closingElement.name.namespace
shiftAstLocationData closingElement.name.name
{openingElement, closingElement}
fragmentAstProperties: (o) ->
openingFragment = Object.assign {
type: 'JSXOpeningFragment'
}, @openingElementLocationData
closingFragment = Object.assign {
type: 'JSXClosingFragment'
}, @closingElementLocationData
{openingFragment, closingFragment}
contentAst: (o) ->
return [] unless @content and not @content.base.isEmpty?()
content = @content.unwrapAll()
children =
if content instanceof StringLiteral
[new JSXText content]
else # StringWithInterpolations
for element in @content.unwrapAll().extractElements o, includeInterpolationWrappers: yes, isJsx: yes
if element instanceof StringLiteral
new JSXText element
else # Interpolation
{expression} = element
unless expression?
emptyExpression = new JSXEmptyExpression()
emptyExpression.locationData = emptyExpressionLocationData {
interpolationNode: element
openingBrace: '{'
closingBrace: '}'
}
new JSXExpressionContainer emptyExpression, locationData: element.locationData
else
unwrapped = expression.unwrapAll()
if unwrapped instanceof JSXElement and
distinguish <a><b /></a>
from <a>{<b />}</a>
unwrapped.locationData.range[0] is element.locationData.range[0]
unwrapped
else
new JSXExpressionContainer unwrapped, locationData: element.locationData
child.ast(o) for child in children when not (child instanceof JSXText and child.value.length is 0)
astProperties: (o) ->
Object.assign(
if @isFragment()
@fragmentAstProperties o
else
@elementAstProperties o
,
children: @contentAst o
)
astLocationData: ->
if @closingElementLocationData?
mergeAstLocationData @openingElementLocationData, @closingElementLocationData
else
@openingElementLocationData
Node for a function invocation.
exports.Call = class Call extends Base
constructor: (@variable, @args = [], @soak, @token) ->
super()
@implicit = @args.implicit
@isNew = no
if @variable instanceof Value and @variable.isNotCallable()
@variable.error "literal is not a function"
if @variable.base instanceof JSXTag
return new JSXElement(
tagName: @variable
attributes: new JSXAttributes @args[0].base
content: @args[1]
)
@variable
never gets output as a result of this node getting created as
part of RegexWithInterpolations
, so for that case move any comments to
the args
property that gets passed into RegexWithInterpolations
via
the grammar.
if @variable.base?.value is 'RegExp' and @args.length isnt 0
moveComments @variable, @args[0]
children: ['variable', 'args']
When setting the location, we sometimes need to update the start location to
account for a newly-discovered new
operator to the left of us. This
expands the range on the left, but not the right.
updateLocationDataIfMissing: (locationData) ->
if @locationData and @needsUpdatedStartLocation
@locationData = Object.assign {},
@locationData,
first_line: locationData.first_line
first_column: locationData.first_column
range: [
locationData.range[0]
@locationData.range[1]
]
base = @variable?.base or @variable
if base.needsUpdatedStartLocation
@variable.locationData = Object.assign {},
@variable.locationData,
first_line: locationData.first_line
first_column: locationData.first_column
range: [
locationData.range[0]
@variable.locationData.range[1]
]
base.updateLocationDataIfMissing locationData
delete @needsUpdatedStartLocation
super locationData
Tag this invocation as creating a new instance.
newInstance: ->
base = @variable?.base or @variable
if base instanceof Call and not base.isNew
base.newInstance()
else
@isNew = true
@needsUpdatedStartLocation = true
this
Soaked chained invocations unfold into if/else ternary structures.
unfoldSoak: (o) ->
if @soak
if @variable instanceof Super
left = new Literal @variable.compile o
rite = new Value left
@variable.error "Unsupported reference to 'super'" unless @variable.accessor?
else
return ifn if ifn = unfoldSoak o, this, 'variable'
[left, rite] = new Value(@variable).cacheReference o
rite = new Call rite, @args
rite.isNew = @isNew
left = new Literal "typeof #{ left.compile o } === \"function\""
return new If left, new Value(rite), soak: yes
call = this
list = []
loop
if call.variable instanceof Call
list.push call
call = call.variable
continue
break unless call.variable instanceof Value
list.push call
break unless (call = call.variable.base) instanceof Call
for call in list.reverse()
if ifn
if call.variable instanceof Call
call.variable = ifn
else
call.variable.base = ifn
ifn = unfoldSoak o, call, 'variable'
ifn
Compile a vanilla function call.
compileNode: (o) ->
@checkForNewSuper()
@variable?.front = @front
compiledArgs = []
If variable is Accessor
fragments are cached and used later
in Value::compileNode
to ensure correct order of the compilation,
and reuse of variables in the scope.
Example:
a(x = 5).b(-> x = 6)
should compile in the same order as
a(x = 5); b(-> x = 6)
(see issue #4437, https://github.com/jashkenas/coffeescript/issues/4437)
varAccess = @variable?.properties?[0] instanceof Access
argCode = (arg for arg in (@args || []) when arg instanceof Code)
if argCode.length > 0 and varAccess and not @variable.base.cached
[cache] = @variable.base.cache o, LEVEL_ACCESS, -> no
@variable.base.cached = cache
for arg, argIndex in @args
if argIndex then compiledArgs.push @makeCode ", "
compiledArgs.push (arg.compileToFragments o, LEVEL_LIST)...
fragments = []
if @isNew
fragments.push @makeCode 'new '
fragments.push @variable.compileToFragments(o, LEVEL_ACCESS)...
fragments.push @makeCode('('), compiledArgs..., @makeCode(')')
fragments
checkForNewSuper: ->
if @isNew
@variable.error "Unsupported reference to 'super'" if @variable instanceof Super
containsSoak: ->
return yes if @soak
return yes if @variable?.containsSoak?()
no
astNode: (o) ->
if @soak and @variable instanceof Super and o.scope.namedMethod()?.ctor
@variable.error "Unsupported reference to 'super'"
@checkForNewSuper()
super o
astType: ->
if @isNew
'NewExpression'
else if @containsSoak()
'OptionalCallExpression'
else
'CallExpression'
astProperties: (o) ->
return
callee: @variable.ast o, LEVEL_ACCESS
arguments: arg.ast(o, LEVEL_LIST) for arg in @args
optional: !!@soak
implicit: !!@implicit
Takes care of converting super()
calls into calls against the prototype’s
function of the same name.
When expressions
are set the call will be compiled in such a way that the
expressions are evaluated without altering the return value of the SuperCall
expression.
exports.SuperCall = class SuperCall extends Call
children: Call::children.concat ['expressions']
isStatement: (o) ->
@expressions?.length and o.level is LEVEL_TOP
compileNode: (o) ->
return super o unless @expressions?.length
superCall = new Literal fragmentsToText super o
replacement = new Block @expressions.slice()
if o.level > LEVEL_TOP
If we might be in an expression we need to cache and return the result
[superCall, ref] = superCall.cache o, null, YES
replacement.push ref
replacement.unshift superCall
replacement.compileToFragments o, if o.level is LEVEL_TOP then o.level else LEVEL_LIST
exports.Super = class Super extends Base
constructor: (@accessor, @superLiteral) ->
super()
children: ['accessor']
compileNode: (o) ->
@checkInInstanceMethod o
method = o.scope.namedMethod()
unless method.ctor? or @accessor?
{name, variable} = method
if name.shouldCache() or (name instanceof Index and name.index.isAssignable())
nref = new IdentifierLiteral o.scope.parent.freeVariable 'name'
name.index = new Assign nref, name.index
@accessor = if nref? then new Index nref else name
if @accessor?.name?.comments
A super()
call gets compiled to e.g. super.method()
, which means
the method
property name gets compiled for the first time here, and
again when the method:
property of the class gets compiled. Since
this compilation happens first, comments attached to method:
would
get incorrectly output near super.method()
, when we want them to
get output on the second pass when method:
is output. So set them
aside during this compilation pass, and put them back on the object so
that they’re there for the later compilation.
salvagedComments = @accessor.name.comments
delete @accessor.name.comments
fragments = (new Value (new Literal 'super'), if @accessor then [ @accessor ] else [])
.compileToFragments o
attachCommentsToNode salvagedComments, @accessor.name if salvagedComments
fragments
checkInInstanceMethod: (o) ->
method = o.scope.namedMethod()
@error 'cannot use super outside of an instance method' unless method?.isMethod
astNode: (o) ->
@checkInInstanceMethod o
if @accessor?
return (
new Value(
new Super().withLocationDataFrom (@superLiteral ? @)
[@accessor]
).withLocationDataFrom @
).ast o
super o
Regexes with interpolations are in fact just a variation of a Call
(a
RegExp()
call to be precise) with a StringWithInterpolations
inside.
exports.RegexWithInterpolations = class RegexWithInterpolations extends Base
constructor: (@call, {@heregexCommentTokens = []} = {}) ->
super()
children: ['call']
compileNode: (o) ->
@call.compileNode o
astType: -> 'InterpolatedRegExpLiteral'
astProperties: (o) ->
interpolatedPattern: @call.args[0].ast o
flags: @call.args[1]?.unwrap().originalValue ? ''
comments:
for heregexCommentToken in @heregexCommentTokens
if heregexCommentToken.here
new HereComment(heregexCommentToken).ast o
else
new LineComment(heregexCommentToken).ast o
exports.TaggedTemplateCall = class TaggedTemplateCall extends Call
constructor: (variable, arg, soak) ->
arg = StringWithInterpolations.fromStringLiteral arg if arg instanceof StringLiteral
super variable, [ arg ], soak
compileNode: (o) ->
@variable.compileToFragments(o, LEVEL_ACCESS).concat @args[0].compileToFragments(o, LEVEL_LIST)
astType: -> 'TaggedTemplateExpression'
astProperties: (o) ->
return
tag: @variable.ast o, LEVEL_ACCESS
quasi: @args[0].ast o, LEVEL_LIST
Node to extend an object’s prototype with an ancestor object.
After goog.inherits
from the
Closure Library.
exports.Extends = class Extends extends Base
constructor: (@child, @parent) ->
super()
children: ['child', 'parent']
Hooks one constructor into another’s prototype chain.
compileToFragments: (o) ->
new Call(new Value(new Literal utility 'extend', o), [@child, @parent]).compileToFragments o
A .
access into a property of a value, or the ::
shorthand for
an access into the object’s prototype.
exports.Access = class Access extends Base
constructor: (@name, {@soak, @shorthand} = {}) ->
super()
children: ['name']
compileToFragments: (o) ->
name = @name.compileToFragments o
node = @name.unwrap()
if node instanceof PropertyName
[@makeCode('.'), name...]
else
[@makeCode('['), name..., @makeCode(']')]
shouldCache: NO
astNode: (o) ->
Babel doesn’t have an AST node for Access
, but rather just includes
this Access node’s child name
Identifier node as the property
of
the MemberExpression
node.
@name.ast o
A [ ... ]
indexed access into an array or object.
exports.Index = class Index extends Base
constructor: (@index) ->
super()
children: ['index']
compileToFragments: (o) ->
[].concat @makeCode("["), @index.compileToFragments(o, LEVEL_PAREN), @makeCode("]")
shouldCache: ->
@index.shouldCache()
astNode: (o) ->
Babel doesn’t have an AST node for Index
, but rather just includes
this Index node’s child index
Identifier node as the property
of
the MemberExpression
node. The fact that the MemberExpression
’s
property
is an Index means that computed
is true
for the
MemberExpression
.
@index.ast o
A range literal. Ranges can be used to extract portions (slices) of arrays, to specify a range for comprehensions, or as a value, to be expanded into the corresponding array of integers at runtime.
exports.Range = class Range extends Base
children: ['from', 'to']
constructor: (@from, @to, tag) ->
super()
@exclusive = tag is 'exclusive'
@equals = if @exclusive then '' else '='
Compiles the range’s source variables – where it starts and where it ends. But only if they need to be cached to avoid double evaluation.
compileVariables: (o) ->
o = merge o, top: true
shouldCache = del o, 'shouldCache'
[@fromC, @fromVar] = @cacheToCodeFragments @from.cache o, LEVEL_LIST, shouldCache
[@toC, @toVar] = @cacheToCodeFragments @to.cache o, LEVEL_LIST, shouldCache
[@step, @stepVar] = @cacheToCodeFragments step.cache o, LEVEL_LIST, shouldCache if step = del o, 'step'
@fromNum = if @from.isNumber() then parseNumber @fromVar else null
@toNum = if @to.isNumber() then parseNumber @toVar else null
@stepNum = if step?.isNumber() then parseNumber @stepVar else null
When compiled normally, the range returns the contents of the for loop needed to iterate over the values in the range. Used by comprehensions.
compileNode: (o) ->
@compileVariables o unless @fromVar
return @compileArray(o) unless o.index
Set up endpoints.
known = @fromNum? and @toNum?
idx = del o, 'index'
idxName = del o, 'name'
namedIndex = idxName and idxName isnt idx
varPart =
if known and not namedIndex
"var #{idx} = #{@fromC}"
else
"#{idx} = #{@fromC}"
varPart += ", #{@toC}" if @toC isnt @toVar
varPart += ", #{@step}" if @step isnt @stepVar
[lt, gt] = ["#{idx} <#{@equals}", "#{idx} >#{@equals}"]
Generate the condition.
[from, to] = [@fromNum, @toNum]
Always check if the step
isn’t zero to avoid the infinite loop.
stepNotZero = "#{ @stepNum ? @stepVar } !== 0"
stepCond = "#{ @stepNum ? @stepVar } > 0"
lowerBound = "#{lt} #{ if known then to else @toVar }"
upperBound = "#{gt} #{ if known then to else @toVar }"
condPart =
if @step?
if @stepNum? and @stepNum isnt 0
if @stepNum > 0 then "#{lowerBound}" else "#{upperBound}"
else
"#{stepNotZero} && (#{stepCond} ? #{lowerBound} : #{upperBound})"
else
if known
"#{ if from <= to then lt else gt } #{to}"
else
"(#{@fromVar} <= #{@toVar} ? #{lowerBound} : #{upperBound})"
cond = if @stepVar then "#{@stepVar} > 0" else "#{@fromVar} <= #{@toVar}"
Generate the step.
stepPart = if @stepVar
"#{idx} += #{@stepVar}"
else if known
if namedIndex
if from <= to then "++#{idx}" else "--#{idx}"
else
if from <= to then "#{idx}++" else "#{idx}--"
else
if namedIndex
"#{cond} ? ++#{idx} : --#{idx}"
else
"#{cond} ? #{idx}++ : #{idx}--"
varPart = "#{idxName} = #{varPart}" if namedIndex
stepPart = "#{idxName} = #{stepPart}" if namedIndex
The final loop body.
[@makeCode "#{varPart}; #{condPart}; #{stepPart}"]
When used as a value, expand the range into the equivalent array.
compileArray: (o) ->
known = @fromNum? and @toNum?
if known and Math.abs(@fromNum - @toNum) <= 20
range = [@fromNum..@toNum]
range.pop() if @exclusive
return [@makeCode "[#{ range.join(', ') }]"]
idt = @tab + TAB
i = o.scope.freeVariable 'i', single: true, reserve: no
result = o.scope.freeVariable 'results', reserve: no
pre = "\n#{idt}var #{result} = [];"
if known
o.index = i
body = fragmentsToText @compileNode o
else
vars = "#{i} = #{@fromC}" + if @toC isnt @toVar then ", #{@toC}" else ''
cond = "#{@fromVar} <= #{@toVar}"
body = "var #{vars}; #{cond} ? #{i} <#{@equals} #{@toVar} : #{i} >#{@equals} #{@toVar}; #{cond} ? #{i}++ : #{i}--"
post = "{ #{result}.push(#{i}); }\n#{idt}return #{result};\n#{o.indent}"
hasArgs = (node) -> node?.contains isLiteralArguments
args = ', arguments' if hasArgs(@from) or hasArgs(@to)
[@makeCode "(function() {#{pre}\n#{idt}for (#{body})#{post}}).apply(this#{args ? ''})"]
astProperties: (o) ->
return {
from: @from?.ast(o) ? null
to: @to?.ast(o) ? null
@exclusive
}
An array slice literal. Unlike JavaScript’s Array#slice
, the second parameter
specifies the index of the end of the slice, just as the first parameter
is the index of the beginning.
exports.Slice = class Slice extends Base
children: ['range']
constructor: (@range) ->
super()
We have to be careful when trying to slice through the end of the array,
9e9
is used because not all implementations respect undefined
or 1/0
.
9e9
should be safe because 9e9
> 2**32
, the max array length.
compileNode: (o) ->
{to, from} = @range
Handle an expression in the property access, e.g. a[!b in c..]
.
if from?.shouldCache()
from = new Value new Parens from
if to?.shouldCache()
to = new Value new Parens to
fromCompiled = from?.compileToFragments(o, LEVEL_PAREN) or [@makeCode '0']
if to
compiled = to.compileToFragments o, LEVEL_PAREN
compiledText = fragmentsToText compiled
if not (not @range.exclusive and +compiledText is -1)
toStr = ', ' + if @range.exclusive
compiledText
else if to.isNumber()
"#{+compiledText + 1}"
else
compiled = to.compileToFragments o, LEVEL_ACCESS
"+#{fragmentsToText compiled} + 1 || 9e9"
[@makeCode ".slice(#{ fragmentsToText fromCompiled }#{ toStr or '' })"]
astNode: (o) ->
@range.ast o
An object literal, nothing fancy.
exports.Obj = class Obj extends Base
constructor: (props, @generated = no) ->
super()
@objects = @properties = props or []
children: ['properties']
isAssignable: (opts) ->
for prop in @properties
Check for reserved words.
message = isUnassignable prop.unwrapAll().value
prop.error message if message
prop = prop.value if prop instanceof Assign and
prop.context is 'object' and
prop.value?.base not instanceof Arr
return no unless prop.isAssignable opts
yes
shouldCache: ->
not @isAssignable()
Check if object contains splat.
hasSplat: ->
return yes for prop in @properties when prop instanceof Splat
no
Move rest property to the end of the list.
{a, rest..., b} = obj
-> {a, b, rest...} = obj
foo = ({a, rest..., b}) ->
-> foo = {a, b, rest...}) ->
reorderProperties: ->
props = @properties
splatProps = @getAndCheckSplatProps()
splatProp = props.splice splatProps[0], 1
@objects = @properties = [].concat props, splatProp
compileNode: (o) ->
@reorderProperties() if @hasSplat() and @lhs
props = @properties
if @generated
for node in props when node instanceof Value
node.error 'cannot have an implicit value in an implicit object'
idt = o.indent += TAB
lastNode = @lastNode @properties
If this object is the left-hand side of an assignment, all its children are too.
@propagateLhs()
isCompact = yes
for prop in @properties
if prop instanceof Assign and prop.context is 'object'
isCompact = no
answer = []
answer.push @makeCode if isCompact then '' else '\n'
for prop, i in props
join = if i is props.length - 1
''
else if isCompact
', '
else if prop is lastNode
'\n'
else
',\n'
indent = if isCompact then '' else idt
key = if prop instanceof Assign and prop.context is 'object'
prop.variable
else if prop instanceof Assign
prop.operatorToken.error "unexpected #{prop.operatorToken.value}" unless @lhs
prop.variable
else
prop
if key instanceof Value and key.hasProperties()
key.error 'invalid object key' if prop.context is 'object' or not key.this
key = key.properties[0].name
prop = new Assign key, prop, 'object'
if key is prop
if prop.shouldCache()
[key, value] = prop.base.cache o
key = new PropertyName key.value if key instanceof IdentifierLiteral
prop = new Assign key, value, 'object'
else if key instanceof Value and key.base instanceof ComputedPropertyName
{ [foo()] }
output as { [ref = foo()]: ref }
.
if prop.base.value.shouldCache()
[key, value] = prop.base.value.cache o
key = new ComputedPropertyName key.value if key instanceof IdentifierLiteral
prop = new Assign key, value, 'object'
else
{ [expression] }
output as { [expression]: expression }
.
prop = new Assign key, prop.base.value, 'object'
else if not prop.bareLiteral?(IdentifierLiteral) and prop not instanceof Splat
prop = new Assign prop, prop, 'object'
if indent then answer.push @makeCode indent
answer.push prop.compileToFragments(o, LEVEL_TOP)...
if join then answer.push @makeCode join
answer.push @makeCode if isCompact then '' else "\n#{@tab}"
answer = @wrapInBraces answer
if @front then @wrapInParentheses answer else answer
getAndCheckSplatProps: ->
return unless @hasSplat() and @lhs
props = @properties
splatProps = (i for prop, i in props when prop instanceof Splat)
props[splatProps[1]].error "multiple spread elements are disallowed" if splatProps?.length > 1
splatProps
assigns: (name) ->
for prop in @properties when prop.assigns name then return yes
no
eachName: (iterator) ->
for prop in @properties
prop = prop.value if prop instanceof Assign and prop.context is 'object'
prop = prop.unwrapAll()
prop.eachName iterator if prop.eachName?
Convert “bare” properties to ObjectProperty
s (or Splat
s).
expandProperty: (property) ->
{variable, context, operatorToken} = property
key = if property instanceof Assign and context is 'object'
variable
else if property instanceof Assign
operatorToken.error "unexpected #{operatorToken.value}" unless @lhs
variable
else
property
if key instanceof Value and key.hasProperties()
key.error 'invalid object key' unless context isnt 'object' and key.this
if property instanceof Assign
return new ObjectProperty fromAssign: property
else
return new ObjectProperty key: property
return new ObjectProperty(fromAssign: property) unless key is property
return property if property instanceof Splat
new ObjectProperty key: property
expandProperties: ->
@expandProperty(property) for property in @properties
propagateLhs: (setLhs) ->
@lhs = yes if setLhs
return unless @lhs
for property in @properties
if property instanceof Assign and property.context is 'object'
{value} = property
unwrappedValue = value.unwrapAll()
if unwrappedValue instanceof Arr or unwrappedValue instanceof Obj
unwrappedValue.propagateLhs yes
else if unwrappedValue instanceof Assign
unwrappedValue.nestedLhs = yes
else if property instanceof Assign
Shorthand property with default, e.g. {a = 1} = b
.
property.nestedLhs = yes
else if property instanceof Splat
property.propagateLhs yes
astNode: (o) ->
@getAndCheckSplatProps()
super o
astType: ->
if @lhs
'ObjectPattern'
else
'ObjectExpression'
astProperties: (o) ->
return
implicit: !!@generated
properties:
property.ast(o) for property in @expandProperties()
exports.ObjectProperty = class ObjectProperty extends Base
constructor: ({key, fromAssign}) ->
super()
if fromAssign
{variable: @key, value, context} = fromAssign
if context is 'object'
All non-shorthand properties (i.e. includes :
).
@value = value
else
Left-hand-side shorthand with default e.g. {a = 1} = b
.
@value = fromAssign
@shorthand = yes
@locationData = fromAssign.locationData
else
Shorthand without default e.g. {a}
or {@a}
or {[a]}
.
@key = key
@shorthand = yes
@locationData = key.locationData
astProperties: (o) ->
isComputedPropertyName = (@key instanceof Value and @key.base instanceof ComputedPropertyName) or @key.unwrap() instanceof StringWithInterpolations
keyAst = @key.ast o, LEVEL_LIST
return
key:
if keyAst?.declaration
Object.assign {}, keyAst, declaration: no
else
keyAst
value: @value?.ast(o, LEVEL_LIST) ? keyAst
shorthand: !!@shorthand
computed: !!isComputedPropertyName
method: no
An array literal.
exports.Arr = class Arr extends Base
constructor: (objs, @lhs = no) ->
super()
@objects = objs or []
@propagateLhs()
children: ['objects']
hasElision: ->
return yes for obj in @objects when obj instanceof Elision
no
isAssignable: (opts) ->
{allowExpansion, allowNontrailingSplat, allowEmptyArray = no} = opts ? {}
return allowEmptyArray unless @objects.length
for obj, i in @objects
return no if not allowNontrailingSplat and obj instanceof Splat and i + 1 isnt @objects.length
return no unless (allowExpansion and obj instanceof Expansion) or (obj.isAssignable(opts) and (not obj.isAtomic or obj.isAtomic()))
yes
shouldCache: ->
not @isAssignable()
compileNode: (o) ->
return [@makeCode '[]'] unless @objects.length
o.indent += TAB
fragmentIsElision = ([ fragment ]) ->
fragment.type is 'Elision' and fragment.code.trim() is ','
Detect if Elision
s at the beginning of the array are processed (e.g. [, , , a]).
passedElision = no
answer = []
for obj, objIndex in @objects
unwrappedObj = obj.unwrapAll()
Let compileCommentFragments
know to intersperse block comments
into the fragments created when compiling this array.
if unwrappedObj.comments and
unwrappedObj.comments.filter((comment) -> not comment.here).length is 0
unwrappedObj.includeCommentFragments = YES
compiledObjs = (obj.compileToFragments o, LEVEL_LIST for obj in @objects)
olen = compiledObjs.length
If compiledObjs
includes newlines, we will output this as a multiline
array (i.e. with a newline and indentation after the [
). If an element
contains line comments, that should also trigger multiline output since
by definition line comments will introduce newlines into our output.
The exception is if only the first element has line comments; in that
case, output as the compact form if we otherwise would have, so that the
first element’s line comments get output before or after the array.
includesLineCommentsOnNonFirstElement = no
for fragments, index in compiledObjs
for fragment in fragments
if fragment.isHereComment
fragment.code = fragment.code.trim()
else if index isnt 0 and includesLineCommentsOnNonFirstElement is no and hasLineComments fragment
includesLineCommentsOnNonFirstElement = yes
Add ‘, ‘ if all Elisions
from the beginning of the array are processed (e.g. [, , , a]) and
element isn’t Elision
or last element is Elision
(e.g. [a,,b,,])
if index isnt 0 and passedElision and (not fragmentIsElision(fragments) or index is olen - 1)
answer.push @makeCode ', '
passedElision = passedElision or not fragmentIsElision fragments
answer.push fragments...
if includesLineCommentsOnNonFirstElement or '\n' in fragmentsToText(answer)
for fragment, fragmentIndex in answer
if fragment.isHereComment
fragment.code = "#{multident(fragment.code, o.indent, no)}\n#{o.indent}"
else if fragment.code is ', ' and not fragment?.isElision and fragment.type not in ['StringLiteral', 'StringWithInterpolations']
fragment.code = ",\n#{o.indent}"
answer.unshift @makeCode "[\n#{o.indent}"
answer.push @makeCode "\n#{@tab}]"
else
for fragment in answer when fragment.isHereComment
fragment.code = "#{fragment.code} "
answer.unshift @makeCode '['
answer.push @makeCode ']'
answer
assigns: (name) ->
for obj in @objects when obj.assigns name then return yes
no
eachName: (iterator) ->
for obj in @objects
obj = obj.unwrapAll()
obj.eachName iterator
If this array is the left-hand side of an assignment, all its children are too.
propagateLhs: (setLhs) ->
@lhs = yes if setLhs
return unless @lhs
for object in @objects
object.lhs = yes if object instanceof Splat or object instanceof Expansion
unwrappedObject = object.unwrapAll()
if unwrappedObject instanceof Arr or unwrappedObject instanceof Obj
unwrappedObject.propagateLhs yes
else if unwrappedObject instanceof Assign
unwrappedObject.nestedLhs = yes
astType: ->
if @lhs
'ArrayPattern'
else
'ArrayExpression'
astProperties: (o) ->
return
elements:
object.ast(o, LEVEL_LIST) for object in @objects
The CoffeeScript class definition. Initialize a Class with its name, an optional superclass, and a body.
exports.Class = class Class extends Base
children: ['variable', 'parent', 'body']
constructor: (@variable, @parent, @body) ->
super()
unless @body?
@body = new Block
@hasGeneratedBody = yes
compileNode: (o) ->
@name = @determineName()
executableBody = @walkBody o
Special handling to allow class expr.A extends A
declarations
parentName = @parent.base.value if @parent instanceof Value and not @parent.hasProperties()
@hasNameClash = @name? and @name is parentName
node = @
if executableBody or @hasNameClash
node = new ExecutableClassBody node, executableBody
else if not @name? and o.level is LEVEL_TOP
Anonymous classes are only valid in expressions
node = new Parens node
if @boundMethods.length and @parent
@variable ?= new IdentifierLiteral o.scope.freeVariable '_class'
[@variable, @variableRef] = @variable.cache o unless @variableRef?
if @variable
node = new Assign @variable, node, null, { @moduleDeclaration }
@compileNode = @compileClassDeclaration
try
return node.compileToFragments o
finally
delete @compileNode
compileClassDeclaration: (o) ->
@ctor ?= @makeDefaultConstructor() if @externalCtor or @boundMethods.length
@ctor?.noReturn = true
@proxyBoundMethods() if @boundMethods.length
o.indent += TAB
result = []
result.push @makeCode "class "
result.push @makeCode @name if @name
@compileCommentFragments o, @variable, result if @variable?.comments?
result.push @makeCode ' ' if @name
result.push @makeCode('extends '), @parent.compileToFragments(o)..., @makeCode ' ' if @parent
result.push @makeCode '{'
unless @body.isEmpty()
@body.spaced = true
result.push @makeCode '\n'
result.push @body.compileToFragments(o, LEVEL_TOP)...
result.push @makeCode "\n#{@tab}"
result.push @makeCode '}'
result
Figure out the appropriate name for this class
determineName: ->
return null unless @variable
[..., tail] = @variable.properties
node = if tail
tail instanceof Access and tail.name
else
@variable.base
unless node instanceof IdentifierLiteral or node instanceof PropertyName
return null
name = node.value
unless tail
message = isUnassignable name
@variable.error message if message
if name in JS_FORBIDDEN then "_#{name}" else name
walkBody: (o) ->
@ctor = null
@boundMethods = []
executableBody = null
initializer = []
{ expressions } = @body
i = 0
for expression in expressions.slice()
if expression instanceof Value and expression.isObject true
{ properties } = expression.base
exprs = []
end = 0
start = 0
pushSlice = -> exprs.push new Value new Obj properties[start...end], true if end > start
while assign = properties[end]
if initializerExpression = @addInitializerExpression assign, o
pushSlice()
exprs.push initializerExpression
initializer.push initializerExpression
start = end + 1
end++
pushSlice()
expressions[i..i] = exprs
i += exprs.length
else
if initializerExpression = @addInitializerExpression expression, o
initializer.push initializerExpression
expressions[i] = initializerExpression
i += 1
for method in initializer when method instanceof Code
if method.ctor
method.error 'Cannot define more than one constructor in a class' if @ctor
@ctor = method
else if method.isStatic and method.bound
method.context = @name
else if method.bound
@boundMethods.push method
return unless o.compiling
if initializer.length isnt expressions.length
@body.expressions = (expression.hoist() for expression in initializer)
new Block expressions
Add an expression to the class initializer
This is the key method for determining whether an expression in a class
body should appear in the initializer or the executable body. If the given
node
is valid in a class body the method will return a (new, modified,
or identical) node for inclusion in the class initializer, otherwise
nothing will be returned and the node will appear in the executable body.
At time of writing, only methods (instance and static) are valid in ES
class initializers. As new ES class features (such as class fields) reach
Stage 4, this method will need to be updated to support them. We
additionally allow PassthroughLiteral
s (backticked expressions) in the
initializer as an escape hatch for ES features that are not implemented
(e.g. getters and setters defined via the get
and set
keywords as
opposed to the Object.defineProperty
method).
addInitializerExpression: (node, o) ->
if node.unwrapAll() instanceof PassthroughLiteral
node
else if @validInitializerMethod node
@addInitializerMethod node
else if not o.compiling and @validClassProperty node
@addClassProperty node
else if not o.compiling and @validClassPrototypeProperty node
@addClassPrototypeProperty node
else
null
Checks if the given node is a valid ES class initializer method.
validInitializerMethod: (node) ->
return no unless node instanceof Assign and node.value instanceof Code
return yes if node.context is 'object' and not node.variable.hasProperties()
return node.variable.looksStatic(@name) and (@name or not node.value.bound)
Returns a configured class initializer method
addInitializerMethod: (assign) ->
{ variable, value: method, operatorToken } = assign
method.isMethod = yes
method.isStatic = variable.looksStatic @name
if method.isStatic
method.name = variable.properties[0]
else
methodName = variable.base
method.name = new (if methodName.shouldCache() then Index else Access) methodName
method.name.updateLocationDataIfMissing methodName.locationData
isConstructor =
if methodName instanceof StringLiteral
methodName.originalValue is 'constructor'
else
methodName.value is 'constructor'
method.ctor = (if @parent then 'derived' else 'base') if isConstructor
method.error 'Cannot define a constructor as a bound (fat arrow) function' if method.bound and method.ctor
method.operatorToken = operatorToken
method
validClassProperty: (node) ->
return no unless node instanceof Assign
return node.variable.looksStatic @name
addClassProperty: (assign) ->
{variable, value, operatorToken} = assign
{staticClassName} = variable.looksStatic @name
new ClassProperty({
name: variable.properties[0]
isStatic: yes
staticClassName
value
operatorToken
}).withLocationDataFrom assign
validClassPrototypeProperty: (node) ->
return no unless node instanceof Assign
node.context is 'object' and not node.variable.hasProperties()
addClassPrototypeProperty: (assign) ->
{variable, value} = assign
new ClassPrototypeProperty({
name: variable.base
value
}).withLocationDataFrom assign
makeDefaultConstructor: ->
ctor = @addInitializerMethod new Assign (new Value new PropertyName 'constructor'), new Code
@body.unshift ctor
if @parent
ctor.body.push new SuperCall new Super, [new Splat new IdentifierLiteral 'arguments']
if @externalCtor
applyCtor = new Value @externalCtor, [ new Access new PropertyName 'apply' ]
applyArgs = [ new ThisLiteral, new IdentifierLiteral 'arguments' ]
ctor.body.push new Call applyCtor, applyArgs
ctor.body.makeReturn()
ctor
proxyBoundMethods: ->
@ctor.thisAssignments = for method in @boundMethods
method.classVariable = @variableRef if @parent
name = new Value(new ThisLiteral, [ method.name ])
new Assign name, new Call(new Value(name, [new Access new PropertyName 'bind']), [new ThisLiteral])
null
declareName: (o) ->
return unless (name = @variable?.unwrap()) instanceof IdentifierLiteral
alreadyDeclared = o.scope.find name.value
name.isDeclaration = not alreadyDeclared
isStatementAst: -> yes
astNode: (o) ->
if jumpNode = @body.jumps()
jumpNode.error 'Class bodies cannot contain pure statements'
if argumentsNode = @body.contains isLiteralArguments
argumentsNode.error "Class bodies shouldn't reference arguments"
@declareName o
@name = @determineName()
@body.isClassBody = yes
@body.locationData = zeroWidthLocationDataFromEndLocation @locationData if @hasGeneratedBody
@walkBody o
sniffDirectives @body.expressions
@ctor?.noReturn = yes
super o
astType: (o) ->
if o.level is LEVEL_TOP
'ClassDeclaration'
else
'ClassExpression'
astProperties: (o) ->
return
id: @variable?.ast(o) ? null
superClass: @parent?.ast(o, LEVEL_PAREN) ? null
body: @body.ast o, LEVEL_TOP
exports.ExecutableClassBody = class ExecutableClassBody extends Base
children: [ 'class', 'body' ]
defaultClassVariableName: '_Class'
constructor: (@class, @body = new Block) ->
super()
compileNode: (o) ->
if jumpNode = @body.jumps()
jumpNode.error 'Class bodies cannot contain pure statements'
if argumentsNode = @body.contains isLiteralArguments
argumentsNode.error "Class bodies shouldn't reference arguments"
params = []
args = [new ThisLiteral]
wrapper = new Code params, @body
klass = new Parens new Call (new Value wrapper, [new Access new PropertyName 'call']), args
@body.spaced = true
o.classScope = wrapper.makeScope o.scope
@name = @class.name ? o.classScope.freeVariable @defaultClassVariableName
ident = new IdentifierLiteral @name
directives = @walkBody()
@setContext()
if @class.hasNameClash
parent = new IdentifierLiteral o.classScope.freeVariable 'superClass'
wrapper.params.push new Param parent
args.push @class.parent
@class.parent = parent
if @externalCtor
externalCtor = new IdentifierLiteral o.classScope.freeVariable 'ctor', reserve: no
@class.externalCtor = externalCtor
@externalCtor.variable.base = externalCtor
if @name isnt @class.name
@body.expressions.unshift new Assign (new IdentifierLiteral @name), @class
else
@body.expressions.unshift @class
@body.expressions.unshift directives...
@body.push ident
klass.compileToFragments o
Traverse the class’s children and:
@properties
@properties
walkBody: ->
directives = []
index = 0
while expr = @body.expressions[index]
break unless expr instanceof Value and expr.isString()
if expr.hoisted
index++
else
directives.push @body.expressions.splice(index, 1)...
@traverseChildren false, (child) =>
return false if child instanceof Class or child instanceof HoistTarget
cont = true
if child instanceof Block
for node, i in child.expressions
if node instanceof Value and node.isObject(true)
cont = false
child.expressions[i] = @addProperties node.base.properties
else if node instanceof Assign and node.variable.looksStatic @name
node.value.isStatic = yes
child.expressions = flatten child.expressions
cont
directives
setContext: ->
@body.traverseChildren false, (node) =>
if node instanceof ThisLiteral
node.value = @name
else if node instanceof Code and node.bound and (node.isStatic or not node.name)
node.context = @name
Make class/prototype assignments for invalid ES properties
addProperties: (assigns) ->
result = for assign in assigns
variable = assign.variable
base = variable?.base
value = assign.value
delete assign.context
if base.value is 'constructor'
if value instanceof Code
base.error 'constructors must be defined at the top level of a class body'
The class scope is not available yet, so return the assignment to update later
assign = @externalCtor = new Assign new Value, value
else if not assign.variable.this
name =
if base instanceof ComputedPropertyName
new Index base.value
else
new (if base.shouldCache() then Index else Access) base
prototype = new Access new PropertyName 'prototype'
variable = new Value new ThisLiteral(), [ prototype, name ]
assign.variable = variable
else if assign.value instanceof Code
assign.value.isStatic = true
assign
compact result
exports.ClassProperty = class ClassProperty extends Base
constructor: ({@name, @isStatic, @staticClassName, @value, @operatorToken}) ->
super()
children: ['name', 'value', 'staticClassName']
isStatement: YES
astProperties: (o) ->
return
key: @name.ast o, LEVEL_LIST
value: @value.ast o, LEVEL_LIST
static: !!@isStatic
computed: @name instanceof Index or @name instanceof ComputedPropertyName
operator: @operatorToken?.value ? '='
staticClassName: @staticClassName?.ast(o) ? null
exports.ClassPrototypeProperty = class ClassPrototypeProperty extends Base
constructor: ({@name, @value}) ->
super()
children: ['name', 'value']
isStatement: YES
astProperties: (o) ->
return
key: @name.ast o, LEVEL_LIST
value: @value.ast o, LEVEL_LIST
computed: @name instanceof ComputedPropertyName or @name instanceof StringWithInterpolations
exports.ModuleDeclaration = class ModuleDeclaration extends Base
constructor: (@clause, @source, @assertions) ->
super()
@checkSource()
children: ['clause', 'source', 'assertions']
isStatement: YES
jumps: THIS
makeReturn: THIS
checkSource: ->
if @source? and @source instanceof StringWithInterpolations
@source.error 'the name of the module to be imported from must be an uninterpolated string'
checkScope: (o, moduleDeclarationType) ->
TODO: would be appropriate to flag this error during AST generation (as
well as when compiling to JS). But o.indent
isn’t tracked during AST
generation, and there doesn’t seem to be a current alternative way to track
whether we’re at the “program top-level”.
if o.indent.length isnt 0
@error "#{moduleDeclarationType} statements must be at top-level scope"
astAssertions: (o) ->
if @assertions?.properties?
@assertions.properties.map (assertion) =>
{ start, end, loc, left, right } = assertion.ast(o)
{ type: 'ImportAttribute', start, end, loc, key: left, value: right }
else
[]
exports.ImportDeclaration = class ImportDeclaration extends ModuleDeclaration
compileNode: (o) ->
@checkScope o, 'import'
o.importedSymbols = []
code = []
code.push @makeCode "#{@tab}import "
code.push @clause.compileNode(o)... if @clause?
if @source?.value?
code.push @makeCode ' from ' unless @clause is null
code.push @makeCode @source.value
if @assertions?
code.push @makeCode ' assert '
code.push @assertions.compileToFragments(o)...
code.push @makeCode ';'
code
astNode: (o) ->
o.importedSymbols = []
super o
astProperties: (o) ->
ret =
specifiers: @clause?.ast(o) ? []
source: @source.ast o
assertions: @astAssertions(o)
ret.importKind = 'value' if @clause
ret
exports.ImportClause = class ImportClause extends Base
constructor: (@defaultBinding, @namedImports) ->
super()
children: ['defaultBinding', 'namedImports']
compileNode: (o) ->
code = []
if @defaultBinding?
code.push @defaultBinding.compileNode(o)...
code.push @makeCode ', ' if @namedImports?
if @namedImports?
code.push @namedImports.compileNode(o)...
code
astNode: (o) ->
The AST for ImportClause
is the non-nested list of import specifiers
that will be the specifiers
property of an ImportDeclaration
AST
compact flatten [
@defaultBinding?.ast o
@namedImports?.ast o
]
exports.ExportDeclaration = class ExportDeclaration extends ModuleDeclaration
compileNode: (o) ->
@checkScope o, 'export'
@checkForAnonymousClassExport()
code = []
code.push @makeCode "#{@tab}export "
code.push @makeCode 'default ' if @ instanceof ExportDefaultDeclaration
if @ not instanceof ExportDefaultDeclaration and
(@clause instanceof Assign or @clause instanceof Class)
code.push @makeCode 'var '
@clause.moduleDeclaration = 'export'
if @clause.body? and @clause.body instanceof Block
code = code.concat @clause.compileToFragments o, LEVEL_TOP
else
code = code.concat @clause.compileNode o
if @source?.value?
code.push @makeCode " from #{@source.value}"
if @assertions?
code.push @makeCode ' assert '
code.push @assertions.compileToFragments(o)...
code.push @makeCode ';'
code
Prevent exporting an anonymous class; all exported members must be named
checkForAnonymousClassExport: ->
if @ not instanceof ExportDefaultDeclaration and @clause instanceof Class and not @clause.variable
@clause.error 'anonymous classes cannot be exported'
astNode: (o) ->
@checkForAnonymousClassExport()
super o
exports.ExportNamedDeclaration = class ExportNamedDeclaration extends ExportDeclaration
astProperties: (o) ->
ret =
source: @source?.ast(o) ? null
assertions: @astAssertions(o)
exportKind: 'value'
clauseAst = @clause.ast o
if @clause instanceof ExportSpecifierList
ret.specifiers = clauseAst
ret.declaration = null
else
ret.specifiers = []
ret.declaration = clauseAst
ret
exports.ExportDefaultDeclaration = class ExportDefaultDeclaration extends ExportDeclaration
astProperties: (o) ->
return
declaration: @clause.ast o
assertions: @astAssertions(o)
exports.ExportAllDeclaration = class ExportAllDeclaration extends ExportDeclaration
astProperties: (o) ->
return
source: @source.ast o
assertions: @astAssertions(o)
exportKind: 'value'
exports.ModuleSpecifierList = class ModuleSpecifierList extends Base
constructor: (@specifiers) ->
super()
children: ['specifiers']
compileNode: (o) ->
code = []
o.indent += TAB
compiledList = (specifier.compileToFragments o, LEVEL_LIST for specifier in @specifiers)
if @specifiers.length isnt 0
code.push @makeCode "{\n#{o.indent}"
for fragments, index in compiledList
code.push @makeCode(",\n#{o.indent}") if index
code.push fragments...
code.push @makeCode "\n}"
else
code.push @makeCode '{}'
code
astNode: (o) ->
specifier.ast(o) for specifier in @specifiers
exports.ImportSpecifierList = class ImportSpecifierList extends ModuleSpecifierList
exports.ExportSpecifierList = class ExportSpecifierList extends ModuleSpecifierList
exports.ModuleSpecifier = class ModuleSpecifier extends Base
constructor: (@original, @alias, @moduleDeclarationType) ->
super()
if @original.comments or @alias?.comments
@comments = []
@comments.push @original.comments... if @original.comments
@comments.push @alias.comments... if @alias?.comments
The name of the variable entering the local scope
@identifier = if @alias? then @alias.value else @original.value
children: ['original', 'alias']
compileNode: (o) ->
@addIdentifierToScope o
code = []
code.push @makeCode @original.value
code.push @makeCode " as #{@alias.value}" if @alias?
code
addIdentifierToScope: (o) ->
o.scope.find @identifier, @moduleDeclarationType
astNode: (o) ->
@addIdentifierToScope o
super o
exports.ImportSpecifier = class ImportSpecifier extends ModuleSpecifier
constructor: (imported, local) ->
super imported, local, 'import'
addIdentifierToScope: (o) ->
Per the spec, symbols can’t be imported multiple times
(e.g. import { foo, foo } from 'lib'
is invalid)
if @identifier in o.importedSymbols or o.scope.check(@identifier)
@error "'#{@identifier}' has already been declared"
else
o.importedSymbols.push @identifier
super o
astProperties: (o) ->
originalAst = @original.ast o
return
imported: originalAst
local: @alias?.ast(o) ? originalAst
importKind: null
exports.ImportDefaultSpecifier = class ImportDefaultSpecifier extends ImportSpecifier
astProperties: (o) ->
return
local: @original.ast o
exports.ImportNamespaceSpecifier = class ImportNamespaceSpecifier extends ImportSpecifier
astProperties: (o) ->
return
local: @alias.ast o
exports.ExportSpecifier = class ExportSpecifier extends ModuleSpecifier
constructor: (local, exported) ->
super local, exported, 'export'
astProperties: (o) ->
originalAst = @original.ast o
return
local: originalAst
exported: @alias?.ast(o) ? originalAst
exports.DynamicImport = class DynamicImport extends Base
compileNode: ->
[@makeCode 'import']
astType: -> 'Import'
exports.DynamicImportCall = class DynamicImportCall extends Call
compileNode: (o) ->
@checkArguments()
super o
checkArguments: ->
unless 1 <= @args.length <= 2
@error 'import() accepts either one or two arguments'
astNode: (o) ->
@checkArguments()
super o
The Assign is used to assign a local variable to value, or to set the property of an object – including within object literals.
exports.Assign = class Assign extends Base
constructor: (@variable, @value, @context, options = {}) ->
super()
{@param, @subpattern, @operatorToken, @moduleDeclaration, @originalContext = @context} = options
@propagateLhs()
children: ['variable', 'value']
isAssignable: YES
isStatement: (o) ->
o?.level is LEVEL_TOP and @context? and (@moduleDeclaration or "?" in @context)
checkNameAssignability: (o, varBase) ->
if o.scope.type(varBase.value) is 'import'
varBase.error "'#{varBase.value}' is read-only"
assigns: (name) ->
@[if @context is 'object' then 'value' else 'variable'].assigns name
unfoldSoak: (o) ->
unfoldSoak o, this, 'variable'
addScopeVariables: (o, {
During AST generation, we need to allow assignment to these constructs
that are considered “unassignable” during compile-to-JS, while still
flagging things like [null] = b
.
allowAssignmentToExpansion = no,
allowAssignmentToNontrailingSplat = no,
allowAssignmentToEmptyArray = no,
allowAssignmentToComplexSplat = no
} = {}) ->
return unless not @context or @context is '**='
varBase = @variable.unwrapAll()
if not varBase.isAssignable {
allowExpansion: allowAssignmentToExpansion
allowNontrailingSplat: allowAssignmentToNontrailingSplat
allowEmptyArray: allowAssignmentToEmptyArray
allowComplexSplat: allowAssignmentToComplexSplat
}
@variable.error "'#{@variable.compile o}' can't be assigned"
varBase.eachName (name) =>
return if name.hasProperties?()
message = isUnassignable name.value
name.error message if message
moduleDeclaration
can be 'import'
or 'export'
.
@checkNameAssignability o, name
if @moduleDeclaration
o.scope.add name.value, @moduleDeclaration
name.isDeclaration = yes
else if @param
o.scope.add name.value,
if @param is 'alwaysDeclare'
'var'
else
'param'
else
alreadyDeclared = o.scope.find name.value
name.isDeclaration ?= not alreadyDeclared
If this assignment identifier has one or more herecomments
attached, output them as part of the declarations line (unless
other herecomments are already staged there) for compatibility
with Flow typing. Don’t do this if this assignment is for a
class, e.g. ClassName = class ClassName {
, as Flow requires
the comment to be between the class name and the {
.
if name.comments and not o.scope.comments[name.value] and
@value not instanceof Class and
name.comments.every((comment) -> comment.here and not comment.multiline)
commentsNode = new IdentifierLiteral name.value
commentsNode.comments = name.comments
commentFragments = []
@compileCommentFragments o, commentsNode, commentFragments
o.scope.comments[name.value] = commentFragments
Compile an assignment, delegating to compileDestructuring
or
compileSplice
if appropriate. Keep track of the name of the base object
we’ve been assigned to, for correct internal references. If the variable
has not been seen yet within the current scope, declare it.
compileNode: (o) ->
isValue = @variable instanceof Value
if isValue
If @variable
is an array or an object, we’re destructuring;
if it’s also isAssignable()
, the destructuring syntax is supported
in ES and we can output it as is; otherwise we @compileDestructuring
and convert this ES-unsupported destructuring into acceptable output.
if @variable.isArray() or @variable.isObject()
unless @variable.isAssignable()
if @variable.isObject() and @variable.base.hasSplat()
return @compileObjectDestruct o
else
return @compileDestructuring o
return @compileSplice o if @variable.isSplice()
return @compileConditional o if @isConditional()
return @compileSpecialMath o if @context in ['//=', '%%=']
@addScopeVariables o
if @value instanceof Code
if @value.isStatic
@value.name = @variable.properties[0]
else if @variable.properties?.length >= 2
[properties..., prototype, name] = @variable.properties
@value.name = name if prototype.name?.value is 'prototype'
val = @value.compileToFragments o, LEVEL_LIST
compiledName = @variable.compileToFragments o, LEVEL_LIST
if @context is 'object'
if @variable.shouldCache()
compiledName.unshift @makeCode '['
compiledName.push @makeCode ']'
return compiledName.concat @makeCode(': '), val
answer = compiledName.concat @makeCode(" #{ @context or '=' } "), val
Per https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/Destructuring_assignment#Assignment_without_declaration, if we’re destructuring without declaring, the destructuring assignment must be wrapped in parentheses. The assignment is wrapped in parentheses if ‘o.level’ has lower precedence than LEVEL_LIST (3) (i.e. LEVEL_COND (4), LEVEL_OP (5) or LEVEL_ACCESS (6)), or if we’re destructuring object, e.g. {a,b} = obj.
if o.level > LEVEL_LIST or isValue and @variable.base instanceof Obj and not @nestedLhs and not (@param is yes)
@wrapInParentheses answer
else
answer
Object rest property is not assignable: {{a}...}
compileObjectDestruct: (o) ->
@variable.base.reorderProperties()
{properties: props} = @variable.base
[..., splat] = props
splatProp = splat.name
assigns = []
refVal = new Value new IdentifierLiteral o.scope.freeVariable 'ref'
props.splice -1, 1, new Splat refVal
assigns.push new Assign(new Value(new Obj props), @value).compileToFragments o, LEVEL_LIST
assigns.push new Assign(new Value(splatProp), refVal).compileToFragments o, LEVEL_LIST
@joinFragmentArrays assigns, ', '
Brief implementation of recursive pattern matching, when assigning array or object literals to a value. Peeks at their properties to assign inner names.
compileDestructuring: (o) ->
top = o.level is LEVEL_TOP
{value} = this
{objects} = @variable.base
olen = objects.length
Special-case for {} = a
and [] = a
(empty patterns).
Compile to simply a
.
if olen is 0
code = value.compileToFragments o
return if o.level >= LEVEL_OP then @wrapInParentheses code else code
[obj] = objects
@disallowLoneExpansion()
{splats, expans, splatsAndExpans} = @getAndCheckSplatsAndExpansions()
isSplat = splats?.length > 0
isExpans = expans?.length > 0
vvar = value.compileToFragments o, LEVEL_LIST
vvarText = fragmentsToText vvar
assigns = []
pushAssign = (variable, val) =>
assigns.push new Assign(variable, val, null, param: @param, subpattern: yes).compileToFragments o, LEVEL_LIST
if isSplat
splatVar = objects[splats[0]].name.unwrap()
if splatVar instanceof Arr or splatVar instanceof Obj
splatVarRef = new IdentifierLiteral o.scope.freeVariable 'ref'
objects[splats[0]].name = splatVarRef
splatVarAssign = -> pushAssign new Value(splatVar), splatVarRef
At this point, there are several things to destructure. So the fn()
in
{a, b} = fn()
must be cached, for example. Make vvar into a simple
variable if it isn’t already.
if value.unwrap() not instanceof IdentifierLiteral or @variable.assigns(vvarText)
ref = o.scope.freeVariable 'ref'
assigns.push [@makeCode(ref + ' = '), vvar...]
vvar = [@makeCode ref]
vvarText = ref
slicer = (type) -> (vvar, start, end = no) ->
vvar = new IdentifierLiteral vvar unless vvar instanceof Value
args = [vvar, new NumberLiteral(start)]
args.push new NumberLiteral end if end
slice = new Value (new IdentifierLiteral utility type, o), [new Access new PropertyName 'call']
new Value new Call slice, args
Helper which outputs [].slice
code.
compSlice = slicer "slice"
Helper which outputs [].splice
code.
compSplice = slicer "splice"
Check if objects
array contains any instance of Assign
, e.g. {a:1}.
hasObjAssigns = (objs) ->
(i for obj, i in objs when obj instanceof Assign and obj.context is 'object')
Check if objects
array contains any unassignable object.
objIsUnassignable = (objs) ->
return yes for obj in objs when not obj.isAssignable()
no
objects
are complex when there is object assign ({a:1}),
unassignable object, or just a single node.
complexObjects = (objs) ->
hasObjAssigns(objs).length or objIsUnassignable(objs) or olen is 1
“Complex” objects
are processed in a loop.
Examples: [a, b, {c, r…}, d], [a, …, {b, r…}, c, d]
loopObjects = (objs, vvar, vvarTxt) =>
for obj, i in objs
Elision
can be skipped.
continue if obj instanceof Elision
If obj
is {a: 1}
if obj instanceof Assign and obj.context is 'object'
{variable: {base: idx}, value: vvar} = obj
{variable: vvar} = vvar if vvar instanceof Assign
idx =
if vvar.this
vvar.properties[0].name
else
new PropertyName vvar.unwrap().value
acc = idx.unwrap() instanceof PropertyName
vval = new Value value, [new (if acc then Access else Index) idx]
else
obj
is [a…], {a…} or a
vvar = switch
when obj instanceof Splat then new Value obj.name
else obj
vval = switch
when obj instanceof Splat then compSlice(vvarTxt, i)
else new Value new Literal(vvarTxt), [new Index new NumberLiteral i]
message = isUnassignable vvar.unwrap().value
vvar.error message if message
pushAssign vvar, vval
“Simple” objects
can be split and compiled to arrays, [a, b, c] = arr, [a, b, c…] = arr
assignObjects = (objs, vvar, vvarTxt) =>
vvar = new Value new Arr(objs, yes)
vval = if vvarTxt instanceof Value then vvarTxt else new Value new Literal(vvarTxt)
pushAssign vvar, vval
processObjects = (objs, vvar, vvarTxt) ->
if complexObjects objs
loopObjects objs, vvar, vvarTxt
else
assignObjects objs, vvar, vvarTxt
In case there is Splat
or Expansion
in objects
,
we can split array in two simple subarrays.
Splat
[a, b, c…, d, e] can be split into [a, b, c…] and [d, e].
Expansion
[a, b, …, c, d] can be split into [a, b] and [c, d].
Examples:
a) Splat
CS: [a, b, c…, d, e] = arr
JS: [a, b, …c] = arr, [d, e] = splice.call(c, -2)
b) Expansion
CS: [a, b, …, d, e] = arr
JS: [a, b] = arr, [d, e] = slice.call(arr, -2)
if splatsAndExpans.length
expIdx = splatsAndExpans[0]
leftObjs = objects.slice 0, expIdx + (if isSplat then 1 else 0)
rightObjs = objects.slice expIdx + 1
processObjects leftObjs, vvar, vvarText if leftObjs.length isnt 0
if rightObjs.length isnt 0
Slice or splice objects
.
refExp = switch
when isSplat then compSplice new Value(objects[expIdx].name), rightObjs.length * -1
when isExpans then compSlice vvarText, rightObjs.length * -1
if complexObjects rightObjs
restVar = refExp
refExp = o.scope.freeVariable 'ref'
assigns.push [@makeCode(refExp + ' = '), restVar.compileToFragments(o, LEVEL_LIST)...]
processObjects rightObjs, vvar, refExp
else
There is no Splat
or Expansion
in objects
.
processObjects objects, vvar, vvarText
splatVarAssign?()
assigns.push vvar unless top or @subpattern
fragments = @joinFragmentArrays assigns, ', '
if o.level < LEVEL_LIST then fragments else @wrapInParentheses fragments
Disallow [...] = a
for some reason. (Could be equivalent to [] = a
?)
disallowLoneExpansion: ->
return unless @variable.base instanceof Arr
{objects} = @variable.base
return unless objects?.length is 1
[loneObject] = objects
if loneObject instanceof Expansion
loneObject.error 'Destructuring assignment has no target'
Show error if there is more than one Splat
, or Expansion
.
Examples: [a, b, c…, d, e, f…], [a, b, …, c, d, …], [a, b, …, c, d, e…]
getAndCheckSplatsAndExpansions: ->
return {splats: [], expans: [], splatsAndExpans: []} unless @variable.base instanceof Arr
{objects} = @variable.base
Count all Splats
: [a, b, c…, d, e]
splats = (i for obj, i in objects when obj instanceof Splat)
Count all Expansions
: [a, b, …, c, d]
expans = (i for obj, i in objects when obj instanceof Expansion)
Combine splats and expansions.
splatsAndExpans = [splats..., expans...]
if splatsAndExpans.length > 1
Sort ‘splatsAndExpans’ so we can show error at first disallowed token.
objects[splatsAndExpans.sort()[1]].error "multiple splats/expansions are disallowed in an assignment"
{splats, expans, splatsAndExpans}
When compiling a conditional assignment, take care to ensure that the operands are only evaluated once, even though we have to reference them more than once.
compileConditional: (o) ->
[left, right] = @variable.cacheReference o
Disallow conditional assignment of undefined variables.
if not left.properties.length and left.base instanceof Literal and
left.base not instanceof ThisLiteral and not o.scope.check left.base.value
@throwUnassignableConditionalError left.base.value
if "?" in @context
o.isExistentialEquals = true
new If(new Existence(left), right, type: 'if').addElse(new Assign(right, @value, '=')).compileToFragments o
else
fragments = new Op(@context[...-1], left, new Assign(right, @value, '=')).compileToFragments o
if o.level <= LEVEL_LIST then fragments else @wrapInParentheses fragments
Convert special math assignment operators like a //= b
to the equivalent
extended form a = a ** b
and then compiles that.
compileSpecialMath: (o) ->
[left, right] = @variable.cacheReference o
new Assign(left, new Op(@context[...-1], right, @value)).compileToFragments o
Compile the assignment from an array splice literal, using JavaScript’s
Array#splice
method.
compileSplice: (o) ->
{range: {from, to, exclusive}} = @variable.properties.pop()
unwrappedVar = @variable.unwrapAll()
if unwrappedVar.comments
moveComments unwrappedVar, @
delete @variable.comments
name = @variable.compile o
if from
[fromDecl, fromRef] = @cacheToCodeFragments from.cache o, LEVEL_OP
else
fromDecl = fromRef = '0'
if to
if from?.isNumber() and to.isNumber()
to = to.compile(o) - fromRef
to += 1 unless exclusive
else
to = to.compile(o, LEVEL_ACCESS) + ' - ' + fromRef
to += ' + 1' unless exclusive
else
to = "9e9"
[valDef, valRef] = @value.cache o, LEVEL_LIST
answer = [].concat @makeCode("#{utility 'splice', o}.apply(#{name}, [#{fromDecl}, #{to}].concat("), valDef, @makeCode(")), "), valRef
if o.level > LEVEL_TOP then @wrapInParentheses answer else answer
eachName: (iterator) ->
@variable.unwrapAll().eachName iterator
isDefaultAssignment: -> @param or @nestedLhs
propagateLhs: ->
return unless @variable?.isArray?() or @variable?.isObject?()
This is the left-hand side of an assignment; let Arr
and Obj
know that, so that those nodes know that they’re assignable as
destructured variables.
@variable.base.propagateLhs yes
throwUnassignableConditionalError: (name) ->
@variable.error "the variable \"#{name}\" can't be assigned with #{@context} because it has not been declared before"
isConditional: ->
@context in ['||=', '&&=', '?=']
isStatementAst: NO
astNode: (o) ->
@disallowLoneExpansion()
@getAndCheckSplatsAndExpansions()
if @isConditional()
variable = @variable.unwrap()
if variable instanceof IdentifierLiteral and not o.scope.check variable.value
@throwUnassignableConditionalError variable.value
@addScopeVariables o, allowAssignmentToExpansion: yes, allowAssignmentToNontrailingSplat: yes, allowAssignmentToEmptyArray: yes, allowAssignmentToComplexSplat: yes
super o
astType: ->
if @isDefaultAssignment()
'AssignmentPattern'
else
'AssignmentExpression'
astProperties: (o) ->
ret =
right: @value.ast o, LEVEL_LIST
left: @variable.ast o, LEVEL_LIST
unless @isDefaultAssignment()
ret.operator = @originalContext ? '='
ret
exports.FuncGlyph = class FuncGlyph extends Base
constructor: (@glyph) ->
super()
A function definition. This is the only node that creates a new Scope. When for the purposes of walking the contents of a function body, the Code has no children – they’re within the inner scope.
exports.Code = class Code extends Base
constructor: (params, body, @funcGlyph, @paramStart) ->
super()
@params = params or []
@body = body or new Block
@bound = @funcGlyph?.glyph is '=>'
@isGenerator = no
@isAsync = no
@isMethod = no
@body.traverseChildren no, (node) =>
if (node instanceof Op and node.isYield()) or node instanceof YieldReturn
@isGenerator = yes
if (node instanceof Op and node.isAwait()) or node instanceof AwaitReturn
@isAsync = yes
if node instanceof For and node.isAwait()
@isAsync = yes
@propagateLhs()
children: ['params', 'body']
isStatement: -> @isMethod
jumps: NO
makeScope: (parentScope) -> new Scope parentScope, @body, this
Compilation creates a new scope unless explicitly asked to share with the outer scope. Handles splat parameters in the parameter list by setting such parameters to be the final parameter in the function definition, as required per the ES2015 spec. If the CoffeeScript function definition had parameters after the splat, they are declared via expressions in the function body.
compileNode: (o) ->
@checkForAsyncOrGeneratorConstructor()
if @bound
@context = o.scope.method.context if o.scope.method?.bound
@context = 'this' unless @context
@updateOptions o
params = []
exprs = []
thisAssignments = @thisAssignments?.slice() ? []
paramsAfterSplat = []
haveSplatParam = no
haveBodyParam = no
@checkForDuplicateParams()
@disallowLoneExpansionAndMultipleSplats()
Separate this
assignments.
@eachParamName (name, node, param, obj) ->
if node.this
name = node.properties[0].name.value
name = "_#{name}" if name in JS_FORBIDDEN
target = new IdentifierLiteral o.scope.freeVariable name, reserve: no
Param
is object destructuring with a default value: ({@prop = 1}) ->
In a case when the variable name is already reserved, we have to assign
a new variable name to the destructured variable: ({prop:prop1 = 1}) ->
replacement =
if param.name instanceof Obj and obj instanceof Assign and
obj.operatorToken.value is '='
new Assign (new IdentifierLiteral name), target, 'object' #, operatorToken: new Literal ':'
else
target
param.renameParam node, replacement
thisAssignments.push new Assign node, target
Parse the parameters, adding them to the list of parameters to put in the
function definition; and dealing with splats or expansions, including
adding expressions to the function body to declare all parameter
variables that would have been after the splat/expansion parameter.
If we encounter a parameter that needs to be declared in the function
body for any reason, for example it’s destructured with this
, also
declare and assign all subsequent parameters in the function body so that
any non-idempotent parameters are evaluated in the correct order.
for param, i in @params
Was ...
used with this parameter? Splat/expansion parameters cannot
have default values, so we need not worry about that.
if param.splat or param instanceof Expansion
haveSplatParam = yes
if param.splat
if param.name instanceof Arr or param.name instanceof Obj
Splat arrays are treated oddly by ES; deal with them the legacy way in the function body. TODO: Should this be handled in the function parameter list, and if so, how?
splatParamName = o.scope.freeVariable 'arg'
params.push ref = new Value new IdentifierLiteral splatParamName
exprs.push new Assign new Value(param.name), ref
else
params.push ref = param.asReference o
splatParamName = fragmentsToText ref.compileNodeWithoutComments o
if param.shouldCache()
exprs.push new Assign new Value(param.name), ref
else # `param` is an Expansion
splatParamName = o.scope.freeVariable 'args'
params.push new Value new IdentifierLiteral splatParamName
o.scope.parameter splatParamName
Parse all other parameters; if a splat paramater has not yet been encountered, add these other parameters to the list to be output in the function definition.
else
if param.shouldCache() or haveBodyParam
param.assignedInBody = yes
haveBodyParam = yes
This parameter cannot be declared or assigned in the parameter
list. So put a reference in the parameter list and add a statement
to the function body assigning it, e.g.
(arg) => { var a = arg.a; }
, with a default value if it has one.
if param.value?
condition = new Op '===', param, new UndefinedLiteral
ifTrue = new Assign new Value(param.name), param.value
exprs.push new If condition, ifTrue
else
exprs.push new Assign new Value(param.name), param.asReference(o), null, param: 'alwaysDeclare'
If this parameter comes before the splat or expansion, it will go in the function definition parameter list.
unless haveSplatParam
If this parameter has a default value, and it hasn’t already been
set by the shouldCache()
block above, define it as a statement in
the function body. This parameter comes after the splat parameter,
so we can’t define its default value in the parameter list.
if param.shouldCache()
ref = param.asReference o
else
if param.value? and not param.assignedInBody
ref = new Assign new Value(param.name), param.value, null, param: yes
else
ref = param
Add this parameter’s reference(s) to the function scope.
if param.name instanceof Arr or param.name instanceof Obj
This parameter is destructured.
param.name.lhs = yes
unless param.shouldCache()
param.name.eachName (prop) ->
o.scope.parameter prop.value
else
This compilation of the parameter is only to get its name to add to the scope name tracking; since the compilation output here isn’t kept for eventual output, don’t include comments in this compilation, so that they get output the “real” time this param is compiled.
paramToAddToScope = if param.value? then param else ref
o.scope.parameter fragmentsToText paramToAddToScope.compileToFragmentsWithoutComments o
params.push ref
else
paramsAfterSplat.push param
If this parameter had a default value, since it’s no longer in the function parameter list we need to assign its default value (if necessary) as an expression in the body.
if param.value? and not param.shouldCache()
condition = new Op '===', param, new UndefinedLiteral
ifTrue = new Assign new Value(param.name), param.value
exprs.push new If condition, ifTrue
Add this parameter to the scope, since it wouldn’t have been added yet since it was skipped earlier.
o.scope.add param.name.value, 'var', yes if param.name?.value?
If there were parameters after the splat or expansion parameter, those parameters need to be assigned in the body of the function.
if paramsAfterSplat.length isnt 0
Create a destructured assignment, e.g. [a, b, c] = [args..., b, c]
exprs.unshift new Assign new Value(
new Arr [new Splat(new IdentifierLiteral(splatParamName)), (param.asReference o for param in paramsAfterSplat)...]
), new Value new IdentifierLiteral splatParamName
Add new expressions to the function body
wasEmpty = @body.isEmpty()
@disallowSuperInParamDefaults()
@checkSuperCallsInConstructorBody()
@body.expressions.unshift thisAssignments... unless @expandCtorSuper thisAssignments
@body.expressions.unshift exprs...
if @isMethod and @bound and not @isStatic and @classVariable
boundMethodCheck = new Value new Literal utility 'boundMethodCheck', o
@body.expressions.unshift new Call(boundMethodCheck, [new Value(new ThisLiteral), @classVariable])
@body.makeReturn() unless wasEmpty or @noReturn
JavaScript doesn’t allow bound (=>
) functions to also be generators.
This is usually caught via Op::compileContinuation
, but double-check:
if @bound and @isGenerator
yieldNode = @body.contains (node) -> node instanceof Op and node.operator is 'yield'
(yieldNode or @).error 'yield cannot occur inside bound (fat arrow) functions'
Assemble the output
modifiers = []
modifiers.push 'static' if @isMethod and @isStatic
modifiers.push 'async' if @isAsync
unless @isMethod or @bound
modifiers.push "function#{if @isGenerator then '*' else ''}"
else if @isGenerator
modifiers.push '*'
signature = [@makeCode '(']
Block comments between a function name and (
get output between
function
and (
.
if @paramStart?.comments?
@compileCommentFragments o, @paramStart, signature
for param, i in params
signature.push @makeCode ', ' if i isnt 0
signature.push @makeCode '...' if haveSplatParam and i is params.length - 1
Compile this parameter, but if any generated variables get created
(e.g. ref
), shift those into the parent scope since we can’t put a
var
line inside a function parameter list.
scopeVariablesCount = o.scope.variables.length
signature.push param.compileToFragments(o, LEVEL_PAREN)...
if scopeVariablesCount isnt o.scope.variables.length
generatedVariables = o.scope.variables.splice scopeVariablesCount
o.scope.parent.variables.push generatedVariables...
signature.push @makeCode ')'
Block comments between )
and ->
/=>
get output between )
and {
.
if @funcGlyph?.comments?
comment.unshift = no for comment in @funcGlyph.comments
@compileCommentFragments o, @funcGlyph, signature
body = @body.compileWithDeclarations o unless @body.isEmpty()
We need to compile the body before method names to ensure super
references are handled.
if @isMethod
[methodScope, o.scope] = [o.scope, o.scope.parent]
name = @name.compileToFragments o
name.shift() if name[0].code is '.'
o.scope = methodScope
answer = @joinFragmentArrays (@makeCode m for m in modifiers), ' '
answer.push @makeCode ' ' if modifiers.length and name
answer.push name... if name
answer.push signature...
answer.push @makeCode ' =>' if @bound and not @isMethod
answer.push @makeCode ' {'
answer.push @makeCode('\n'), body..., @makeCode("\n#{@tab}") if body?.length
answer.push @makeCode '}'
return indentInitial answer, @ if @isMethod
if @front or (o.level >= LEVEL_ACCESS) then @wrapInParentheses answer else answer
updateOptions: (o) ->
o.scope = del(o, 'classScope') or @makeScope o.scope
o.scope.shared = del(o, 'sharedScope')
o.indent += TAB
delete o.bare
delete o.isExistentialEquals
checkForDuplicateParams: ->
paramNames = []
@eachParamName (name, node, param) ->
node.error "multiple parameters named '#{name}'" if name in paramNames
paramNames.push name
eachParamName: (iterator) ->
param.eachName iterator for param in @params
Short-circuit traverseChildren
method to prevent it from crossing scope
boundaries unless crossScope
is true
.
traverseChildren: (crossScope, func) ->
super(crossScope, func) if crossScope
Short-circuit replaceInContext
method to prevent it from crossing context boundaries. Bound
functions have the same context.
replaceInContext: (child, replacement) ->
if @bound
super child, replacement
else
false
disallowSuperInParamDefaults: ({forAst} = {}) ->
return false unless @ctor
@eachSuperCall Block.wrap(@params), (superCall) ->
superCall.error "'super' is not allowed in constructor parameter defaults"
, checkForThisBeforeSuper: not forAst
checkSuperCallsInConstructorBody: ->
return false unless @ctor
seenSuper = @eachSuperCall @body, (superCall) =>
superCall.error "'super' is only allowed in derived class constructors" if @ctor is 'base'
seenSuper
flagThisParamInDerivedClassConstructorWithoutCallingSuper: (param) ->
param.error "Can't use @params in derived class constructors without calling super"
checkForAsyncOrGeneratorConstructor: ->
if @ctor
@name.error 'Class constructor may not be async' if @isAsync
@name.error 'Class constructor may not be a generator' if @isGenerator
disallowLoneExpansionAndMultipleSplats: ->
seenSplatParam = no
for param in @params
Was ...
used with this parameter? (Only one such parameter is allowed
per function.)
if param.splat or param instanceof Expansion
if seenSplatParam
param.error 'only one splat or expansion parameter is allowed per function definition'
else if param instanceof Expansion and @params.length is 1
param.error 'an expansion parameter cannot be the only parameter in a function definition'
seenSplatParam = yes
expandCtorSuper: (thisAssignments) ->
return false unless @ctor
seenSuper = @eachSuperCall @body, (superCall) =>
superCall.expressions = thisAssignments
haveThisParam = thisAssignments.length and thisAssignments.length isnt @thisAssignments?.length
if @ctor is 'derived' and not seenSuper and haveThisParam
param = thisAssignments[0].variable
@flagThisParamInDerivedClassConstructorWithoutCallingSuper param
seenSuper
Find all super calls in the given context node;
returns true
if iterator
is called.
eachSuperCall: (context, iterator, {checkForThisBeforeSuper = yes} = {}) ->
seenSuper = no
context.traverseChildren yes, (child) =>
if child instanceof SuperCall
super
in a constructor (the only super
without an accessor)
cannot be given an argument with a reference to this
, as that would
be referencing this
before calling super
.
unless child.variable.accessor
childArgs = child.args.filter (arg) ->
arg not instanceof Class and (arg not instanceof Code or arg.bound)
Block.wrap(childArgs).traverseChildren yes, (node) =>
node.error "Can't call super with @params in derived class constructors" if node.this
seenSuper = yes
iterator child
else if checkForThisBeforeSuper and child instanceof ThisLiteral and @ctor is 'derived' and not seenSuper
child.error "Can't reference 'this' before calling super in derived class constructors"
super
has the same target in bound (arrow) functions, so check them too
child not instanceof SuperCall and (child not instanceof Code or child.bound)
seenSuper
propagateLhs: ->
for param in @params
{name} = param
if name instanceof Arr or name instanceof Obj
name.propagateLhs yes
else if param instanceof Expansion
param.lhs = yes
astAddParamsToScope: (o) ->
@eachParamName (name) ->
o.scope.add name, 'param'
astNode: (o) ->
@updateOptions o
@checkForAsyncOrGeneratorConstructor()
@checkForDuplicateParams()
@disallowSuperInParamDefaults forAst: yes
@disallowLoneExpansionAndMultipleSplats()
seenSuper = @checkSuperCallsInConstructorBody()
if @ctor is 'derived' and not seenSuper
@eachParamName (name, node) =>
if node.this
@flagThisParamInDerivedClassConstructorWithoutCallingSuper node
@astAddParamsToScope o
@body.makeReturn null, yes unless @body.isEmpty() or @noReturn
super o
astType: ->
if @isMethod
'ClassMethod'
else if @bound
'ArrowFunctionExpression'
else
'FunctionExpression'
paramForAst: (param) ->
return param if param instanceof Expansion
{name, value, splat} = param
if splat
new Splat name, lhs: yes, postfix: splat.postfix
.withLocationDataFrom param
else if value?
new Assign name, value, null, param: yes
.withLocationDataFrom locationData: mergeLocationData name.locationData, value.locationData
else
name
methodAstProperties: (o) ->
getIsComputed = =>
return yes if @name instanceof Index
return yes if @name instanceof ComputedPropertyName
return yes if @name.name instanceof ComputedPropertyName
no
return
static: !!@isStatic
key: @name.ast o
computed: getIsComputed()
kind:
if @ctor
'constructor'
else
'method'
operator: @operatorToken?.value ? '='
staticClassName: @isStatic.staticClassName?.ast(o) ? null
bound: !!@bound
astProperties: (o) ->
return Object.assign
params: @paramForAst(param).ast(o) for param in @params
body: @body.ast (Object.assign {}, o, checkForDirectives: yes), LEVEL_TOP
generator: !!@isGenerator
async: !!@isAsync
We never generate named functions, so specify id
as null
, which
matches the Babel AST for anonymous function expressions/arrow functions
id: null
hasIndentedBody: @body.locationData.first_line > @funcGlyph?.locationData.first_line
,
if @isMethod then @methodAstProperties o else {}
astLocationData: ->
functionLocationData = super()
return functionLocationData unless @isMethod
astLocationData = mergeAstLocationData @name.astLocationData(), functionLocationData
if @isStatic.staticClassName?
astLocationData = mergeAstLocationData @isStatic.staticClassName.astLocationData(), astLocationData
astLocationData
A parameter in a function definition. Beyond a typical JavaScript parameter, these parameters can also attach themselves to the context of the function, as well as be a splat, gathering up a group of parameters into an array.
exports.Param = class Param extends Base
constructor: (@name, @value, @splat) ->
super()
message = isUnassignable @name.unwrapAll().value
@name.error message if message
if @name instanceof Obj and @name.generated
token = @name.objects[0].operatorToken
token.error "unexpected #{token.value}"
children: ['name', 'value']
compileToFragments: (o) ->
@name.compileToFragments o, LEVEL_LIST
compileToFragmentsWithoutComments: (o) ->
@name.compileToFragmentsWithoutComments o, LEVEL_LIST
asReference: (o) ->
return @reference if @reference
node = @name
if node.this
name = node.properties[0].name.value
name = "_#{name}" if name in JS_FORBIDDEN
node = new IdentifierLiteral o.scope.freeVariable name
else if node.shouldCache()
node = new IdentifierLiteral o.scope.freeVariable 'arg'
node = new Value node
node.updateLocationDataIfMissing @locationData
@reference = node
shouldCache: ->
@name.shouldCache()
Iterates the name or names of a Param
.
In a sense, a destructured parameter represents multiple JS parameters. This
method allows to iterate them all.
The iterator
function will be called as iterator(name, node)
where
name
is the name of the parameter and node
is the AST node corresponding
to that name.
eachName: (iterator, name = @name) ->
checkAssignabilityOfLiteral = (literal) ->
message = isUnassignable literal.value
if message
literal.error message
unless literal.isAssignable()
literal.error "'#{literal.value}' can't be assigned"
atParam = (obj, originalObj = null) => iterator "@#{obj.properties[0].name.value}", obj, @, originalObj
if name instanceof Call
name.error "Function invocation can't be assigned"
foo
if name instanceof Literal
checkAssignabilityOfLiteral name
return iterator name.value, name, @
@foo
return atParam name if name instanceof Value
for obj in name.objects ? []
Save original obj.
nObj = obj
if obj instanceof Assign and not obj.context?
obj = obj.variable
{foo:bar}
if obj instanceof Assign
… possibly with a default value
if obj.value instanceof Assign
obj = obj.value.variable
else
obj = obj.value
@eachName iterator, obj.unwrap()
[xs...]
else if obj instanceof Splat
node = obj.name.unwrap()
iterator node.value, node, @
else if obj instanceof Value
[{a}]
if obj.isArray() or obj.isObject()
@eachName iterator, obj.base
{@foo}
else if obj.this
atParam obj, nObj
else
checkAssignabilityOfLiteral obj.base
iterator obj.base.value, obj.base, @
else if obj instanceof Elision
obj
else if obj not instanceof Expansion
obj.error "illegal parameter #{obj.compile()}"
return
Rename a param by replacing the given AST node for a name with a new node. This needs to ensure that the the source for object destructuring does not change.
renameParam: (node, newNode) ->
isNode = (candidate) -> candidate is node
replacement = (node, parent) =>
if parent instanceof Obj
key = node
key = node.properties[0].name if node.this
No need to assign a new variable for the destructured variable if the variable isn’t reserved.
Examples:
({@foo}) ->
should compile to ({foo}) { this.foo = foo}
foo = 1; ({@foo}) ->
should compile to foo = 1; ({foo:foo1}) { this.foo = foo1 }
if node.this and key.value is newNode.value
new Value newNode
else
new Assign new Value(key), newNode, 'object'
else
newNode
@replaceInContext isNode, replacement
A splat, either as a parameter to a function, an argument to a call, or as part of a destructuring assignment.
exports.Splat = class Splat extends Base
constructor: (name, {@lhs, @postfix = true} = {}) ->
super()
@name = if name.compile then name else new Literal name
children: ['name']
shouldCache: -> no
isAssignable: ({allowComplexSplat = no} = {})->
return allowComplexSplat if @name instanceof Obj or @name instanceof Parens
@name.isAssignable() and (not @name.isAtomic or @name.isAtomic())
assigns: (name) ->
@name.assigns name
compileNode: (o) ->
compiledSplat = [@makeCode('...'), @name.compileToFragments(o, LEVEL_OP)...]
return compiledSplat unless @jsx
return [@makeCode('{'), compiledSplat..., @makeCode('}')]
unwrap: -> @name
propagateLhs: (setLhs) ->
@lhs = yes if setLhs
return unless @lhs
@name.propagateLhs? yes
astType: ->
if @jsx
'JSXSpreadAttribute'
else if @lhs
'RestElement'
else
'SpreadElement'
astProperties: (o) -> {
argument: @name.ast o, LEVEL_OP
@postfix
}
Used to skip values inside an array destructuring (pattern matching) or parameter list.
exports.Expansion = class Expansion extends Base
shouldCache: NO
compileNode: (o) ->
@throwLhsError()
asReference: (o) ->
this
eachName: (iterator) ->
throwLhsError: ->
@error 'Expansion must be used inside a destructuring assignment or parameter list'
astNode: (o) ->
unless @lhs
@throwLhsError()
super o
astType: -> 'RestElement'
astProperties: ->
return
argument: null
Array elision element (for example, [,a, , , b, , c, ,]).
exports.Elision = class Elision extends Base
isAssignable: YES
shouldCache: NO
compileToFragments: (o, level) ->
fragment = super o, level
fragment.isElision = yes
fragment
compileNode: (o) ->
[@makeCode ', ']
asReference: (o) ->
this
eachName: (iterator) ->
astNode: ->
null
A while loop, the only sort of low-level loop exposed by CoffeeScript. From it, all other loops can be manufactured. Useful in cases where you need more flexibility or more speed than a comprehension can provide.
exports.While = class While extends Base
constructor: (@condition, {invert: @inverted, @guard, @isLoop} = {}) ->
super()
children: ['condition', 'guard', 'body']
isStatement: YES
makeReturn: (results, mark) ->
return super(results, mark) if results
@returns = not @jumps()
if mark
@body.makeReturn(results, mark) if @returns
return
this
addBody: (@body) ->
this
jumps: ->
{expressions} = @body
return no unless expressions.length
for node in expressions
return jumpNode if jumpNode = node.jumps loop: yes
no
The main difference from a JavaScript while is that the CoffeeScript while can be used as a part of a larger expression – while loops may return an array containing the computed result of each iteration.
compileNode: (o) ->
o.indent += TAB
set = ''
{body} = this
if body.isEmpty()
body = @makeCode ''
else
if @returns
body.makeReturn rvar = o.scope.freeVariable 'results'
set = "#{@tab}#{rvar} = [];\n"
if @guard
if body.expressions.length > 1
body.expressions.unshift new If (new Parens @guard).invert(), new StatementLiteral "continue"
else
body = Block.wrap [new If @guard, body] if @guard
body = [].concat @makeCode("\n"), (body.compileToFragments o, LEVEL_TOP), @makeCode("\n#{@tab}")
answer = [].concat @makeCode(set + @tab + "while ("), @processedCondition().compileToFragments(o, LEVEL_PAREN),
@makeCode(") {"), body, @makeCode("}")
if @returns
answer.push @makeCode "\n#{@tab}return #{rvar};"
answer
processedCondition: ->
@processedConditionCache ?= if @inverted then @condition.invert() else @condition
astType: -> 'WhileStatement'
astProperties: (o) ->
return
test: @condition.ast o, LEVEL_PAREN
body: @body.ast o, LEVEL_TOP
guard: @guard?.ast(o) ? null
inverted: !!@inverted
postfix: !!@postfix
loop: !!@isLoop
Simple Arithmetic and logical operations. Performs some conversion from CoffeeScript operations into their JavaScript equivalents.
exports.Op = class Op extends Base
constructor: (op, first, second, flip, {@invertOperator, @originalOperator = op} = {}) ->
super()
if op is 'new'
if ((firstCall = unwrapped = first.unwrap()) instanceof Call or (firstCall = unwrapped.base) instanceof Call) and not firstCall.do and not firstCall.isNew
return new Value firstCall.newInstance(), if firstCall is unwrapped then [] else unwrapped.properties
first = new Parens first unless first instanceof Parens or first.unwrap() instanceof IdentifierLiteral or first.hasProperties?()
call = new Call first, []
call.locationData = @locationData
call.isNew = yes
return call
@operator = CONVERSIONS[op] or op
@first = first
@second = second
@flip = !!flip
if @operator in ['--', '++']
message = isUnassignable @first.unwrapAll().value
@first.error message if message
return this
The map of conversions from CoffeeScript to JavaScript symbols.
CONVERSIONS =
'==': '==='
'!=': '!=='
'of': 'in'
'yieldfrom': 'yield*'
The map of invertible operators.
INVERSIONS =
'!==': '==='
'===': '!=='
children: ['first', 'second']
isNumber: ->
@isUnary() and @operator in ['+', '-'] and
@first instanceof Value and @first.isNumber()
isAwait: ->
@operator is 'await'
isYield: ->
@operator in ['yield', 'yield*']
isUnary: ->
not @second
shouldCache: ->
not @isNumber()
Am I capable of Python-style comparison chaining?
isChainable: ->
@operator in ['<', '>', '>=', '<=', '===', '!==']
isChain: ->
@isChainable() and @first.isChainable()
invert: ->
if @isInOperator()
@invertOperator = '!'
return @
if @isChain()
allInvertable = yes
curr = this
while curr and curr.operator
allInvertable and= (curr.operator of INVERSIONS)
curr = curr.first
return new Parens(this).invert() unless allInvertable
curr = this
while curr and curr.operator
curr.invert = !curr.invert
curr.operator = INVERSIONS[curr.operator]
curr = curr.first
this
else if op = INVERSIONS[@operator]
@operator = op
if @first.unwrap() instanceof Op
@first.invert()
this
else if @second
new Parens(this).invert()
else if @operator is '!' and (fst = @first.unwrap()) instanceof Op and
fst.operator in ['!', 'in', 'instanceof']
fst
else
new Op '!', this
unfoldSoak: (o) ->
@operator in ['++', '--', 'delete'] and unfoldSoak o, this, 'first'
generateDo: (exp) ->
passedParams = []
func = if exp instanceof Assign and (ref = exp.value.unwrap()) instanceof Code
ref
else
exp
for param in func.params or []
if param.value
passedParams.push param.value
delete param.value
else
passedParams.push param
call = new Call exp, passedParams
call.do = yes
call
isInOperator: ->
@originalOperator is 'in'
compileNode: (o) ->
if @isInOperator()
inNode = new In @first, @second
return (if @invertOperator then inNode.invert() else inNode).compileNode o
if @invertOperator
@invertOperator = null
return @invert().compileNode(o)
return Op::generateDo(@first).compileNode o if @operator is 'do'
isChain = @isChain()
In chains, there’s no need to wrap bare obj literals in parens, as the chained expression is wrapped.
@first.front = @front unless isChain
@checkDeleteOperand o
return @compileContinuation o if @isYield() or @isAwait()
return @compileUnary o if @isUnary()
return @compileChain o if isChain
switch @operator
when '?' then @compileExistence o, @second.isDefaultValue
when '//' then @compileFloorDivision o
when '%%' then @compileModulo o
else
lhs = @first.compileToFragments o, LEVEL_OP
rhs = @second.compileToFragments o, LEVEL_OP
answer = [].concat lhs, @makeCode(" #{@operator} "), rhs
if o.level <= LEVEL_OP then answer else @wrapInParentheses answer
Mimic Python’s chained comparisons when multiple comparison operators are used sequentially. For example:
bin/coffee -e 'console.log 50 < 65 > 10'
true
compileChain: (o) ->
[@first.second, shared] = @first.second.cache o
fst = @first.compileToFragments o, LEVEL_OP
fragments = fst.concat @makeCode(" #{if @invert then '&&' else '||'} "),
(shared.compileToFragments o), @makeCode(" #{@operator} "), (@second.compileToFragments o, LEVEL_OP)
@wrapInParentheses fragments
Keep reference to the left expression, unless this an existential assignment
compileExistence: (o, checkOnlyUndefined) ->
if @first.shouldCache()
ref = new IdentifierLiteral o.scope.freeVariable 'ref'
fst = new Parens new Assign ref, @first
else
fst = @first
ref = fst
new If(new Existence(fst, checkOnlyUndefined), ref, type: 'if').addElse(@second).compileToFragments o
Compile a unary Op.
compileUnary: (o) ->
parts = []
op = @operator
parts.push [@makeCode op]
if op is '!' and @first instanceof Existence
@first.negated = not @first.negated
return @first.compileToFragments o
if o.level >= LEVEL_ACCESS
return (new Parens this).compileToFragments o
plusMinus = op in ['+', '-']
parts.push [@makeCode(' ')] if op in ['typeof', 'delete'] or
plusMinus and @first instanceof Op and @first.operator is op
if plusMinus and @first instanceof Op
@first = new Parens @first
parts.push @first.compileToFragments o, LEVEL_OP
parts.reverse() if @flip
@joinFragmentArrays parts, ''
compileContinuation: (o) ->
parts = []
op = @operator
@checkContinuation o unless @isAwait()
if 'expression' in Object.keys(@first) and not (@first instanceof Throw)
parts.push @first.expression.compileToFragments o, LEVEL_OP if @first.expression?
else
parts.push [@makeCode "("] if o.level >= LEVEL_PAREN
parts.push [@makeCode op]
parts.push [@makeCode " "] if @first.base?.value isnt ''
parts.push @first.compileToFragments o, LEVEL_OP
parts.push [@makeCode ")"] if o.level >= LEVEL_PAREN
@joinFragmentArrays parts, ''
checkContinuation: (o) ->
unless o.scope.parent?
@error "#{@operator} can only occur inside functions"
if o.scope.method?.bound and o.scope.method.isGenerator
@error 'yield cannot occur inside bound (fat arrow) functions'
compileFloorDivision: (o) ->
floor = new Value new IdentifierLiteral('Math'), [new Access new PropertyName 'floor']
second = if @second.shouldCache() then new Parens @second else @second
div = new Op '/', @first, second
new Call(floor, [div]).compileToFragments o
compileModulo: (o) ->
mod = new Value new Literal utility 'modulo', o
new Call(mod, [@first, @second]).compileToFragments o
toString: (idt) ->
super idt, @constructor.name + ' ' + @operator
checkDeleteOperand: (o) ->
if @operator is 'delete' and o.scope.check(@first.unwrapAll().value)
@error 'delete operand may not be argument or var'
astNode: (o) ->
@checkContinuation o if @isYield()
@checkDeleteOperand o
super o
astType: ->
return 'AwaitExpression' if @isAwait()
return 'YieldExpression' if @isYield()
return 'ChainedComparison' if @isChain()
switch @operator
when '||', '&&', '?' then 'LogicalExpression'
when '++', '--' then 'UpdateExpression'
else
if @isUnary() then 'UnaryExpression'
else 'BinaryExpression'
operatorAst: ->
"#{if @invertOperator then "#{@invertOperator} " else ''}#{@originalOperator}"
chainAstProperties: (o) ->
operators = [@operatorAst()]
operands = [@second]
currentOp = @first
loop
operators.unshift currentOp.operatorAst()
operands.unshift currentOp.second
currentOp = currentOp.first
unless currentOp.isChainable()
operands.unshift currentOp
break
return {
operators
operands: (operand.ast(o, LEVEL_OP) for operand in operands)
}
astProperties: (o) ->
return @chainAstProperties(o) if @isChain()
firstAst = @first.ast o, LEVEL_OP
secondAst = @second?.ast o, LEVEL_OP
operatorAst = @operatorAst()
switch
when @isUnary()
argument =
if @isYield() and @first.unwrap().value is ''
null
else
firstAst
return {argument} if @isAwait()
return {
argument
delegate: @operator is 'yield*'
} if @isYield()
return {
argument
operator: operatorAst
prefix: !@flip
}
else
return
left: firstAst
right: secondAst
operator: operatorAst
exports.In = class In extends Base
constructor: (@object, @array) ->
super()
children: ['object', 'array']
invert: NEGATE
compileNode: (o) ->
if @array instanceof Value and @array.isArray() and @array.base.objects.length
for obj in @array.base.objects when obj instanceof Splat
hasSplat = yes
break
compileOrTest
only if we have an array literal with no splats
return @compileOrTest o unless hasSplat
@compileLoopTest o
compileOrTest: (o) ->
[sub, ref] = @object.cache o, LEVEL_OP
[cmp, cnj] = if @negated then [' !== ', ' && '] else [' === ', ' || ']
tests = []
for item, i in @array.base.objects
if i then tests.push @makeCode cnj
tests = tests.concat (if i then ref else sub), @makeCode(cmp), item.compileToFragments(o, LEVEL_ACCESS)
if o.level < LEVEL_OP then tests else @wrapInParentheses tests
compileLoopTest: (o) ->
[sub, ref] = @object.cache o, LEVEL_LIST
fragments = [].concat @makeCode(utility('indexOf', o) + ".call("), @array.compileToFragments(o, LEVEL_LIST),
@makeCode(", "), ref, @makeCode(") " + if @negated then '< 0' else '>= 0')
return fragments if fragmentsToText(sub) is fragmentsToText(ref)
fragments = sub.concat @makeCode(', '), fragments
if o.level < LEVEL_LIST then fragments else @wrapInParentheses fragments
toString: (idt) ->
super idt, @constructor.name + if @negated then '!' else ''
A classic try/catch/finally block.
exports.Try = class Try extends Base
constructor: (@attempt, @catch, @ensure, @finallyTag) ->
super()
children: ['attempt', 'catch', 'ensure']
isStatement: YES
jumps: (o) -> @attempt.jumps(o) or @catch?.jumps(o)
makeReturn: (results, mark) ->
if mark
@attempt?.makeReturn results, mark
@catch?.makeReturn results, mark
return
@attempt = @attempt.makeReturn results if @attempt
@catch = @catch .makeReturn results if @catch
this
Compilation is more or less as you would expect – the finally clause is optional, the catch is not.
compileNode: (o) ->
originalIndent = o.indent
o.indent += TAB
tryPart = @attempt.compileToFragments o, LEVEL_TOP
catchPart = if @catch
@catch.compileToFragments merge(o, indent: originalIndent), LEVEL_TOP
else unless @ensure or @catch
generatedErrorVariableName = o.scope.freeVariable 'error', reserve: no
[@makeCode(" catch (#{generatedErrorVariableName}) {}")]
else
[]
ensurePart = if @ensure then ([].concat @makeCode(" finally {\n"), @ensure.compileToFragments(o, LEVEL_TOP),
@makeCode("\n#{@tab}}")) else []
[].concat @makeCode("#{@tab}try {\n"),
tryPart,
@makeCode("\n#{@tab}}"), catchPart, ensurePart
astType: -> 'TryStatement'
astProperties: (o) ->
return
block: @attempt.ast o, LEVEL_TOP
handler: @catch?.ast(o) ? null
finalizer:
if @ensure?
Object.assign @ensure.ast(o, LEVEL_TOP),
Include finally
keyword in location data.
mergeAstLocationData(
jisonLocationDataToAstLocationData(@finallyTag.locationData),
@ensure.astLocationData()
)
else
null
exports.Catch = class Catch extends Base
constructor: (@recovery, @errorVariable) ->
super()
@errorVariable?.unwrap().propagateLhs? yes
children: ['recovery', 'errorVariable']
isStatement: YES
jumps: (o) -> @recovery.jumps o
makeReturn: (results, mark) ->
ret = @recovery.makeReturn results, mark
return if mark
@recovery = ret
this
compileNode: (o) ->
o.indent += TAB
generatedErrorVariableName = o.scope.freeVariable 'error', reserve: no
placeholder = new IdentifierLiteral generatedErrorVariableName
@checkUnassignable()
if @errorVariable
@recovery.unshift new Assign @errorVariable, placeholder
[].concat @makeCode(" catch ("), placeholder.compileToFragments(o), @makeCode(") {\n"),
@recovery.compileToFragments(o, LEVEL_TOP), @makeCode("\n#{@tab}}")
checkUnassignable: ->
if @errorVariable
message = isUnassignable @errorVariable.unwrapAll().value
@errorVariable.error message if message
astNode: (o) ->
@checkUnassignable()
@errorVariable?.eachName (name) ->
alreadyDeclared = o.scope.find name.value
name.isDeclaration = not alreadyDeclared
super o
astType: -> 'CatchClause'
astProperties: (o) ->
return
param: @errorVariable?.ast(o) ? null
body: @recovery.ast o, LEVEL_TOP
Simple node to throw an exception.
exports.Throw = class Throw extends Base
constructor: (@expression) ->
super()
children: ['expression']
isStatement: YES
jumps: NO
A Throw is already a return, of sorts…
makeReturn: THIS
compileNode: (o) ->
fragments = @expression.compileToFragments o, LEVEL_LIST
unshiftAfterComments fragments, @makeCode 'throw '
fragments.unshift @makeCode @tab
fragments.push @makeCode ';'
fragments
astType: -> 'ThrowStatement'
astProperties: (o) ->
return
argument: @expression.ast o, LEVEL_LIST
Checks a variable for existence – not null
and not undefined
. This is
similar to .nil?
in Ruby, and avoids having to consult a JavaScript truth
table. Optionally only check if a variable is not undefined
.
exports.Existence = class Existence extends Base
constructor: (@expression, onlyNotUndefined = no) ->
super()
@comparisonTarget = if onlyNotUndefined then 'undefined' else 'null'
salvagedComments = []
@expression.traverseChildren yes, (child) ->
if child.comments
for comment in child.comments
salvagedComments.push comment unless comment in salvagedComments
delete child.comments
attachCommentsToNode salvagedComments, @
moveComments @expression, @
children: ['expression']
invert: NEGATE
compileNode: (o) ->
@expression.front = @front
code = @expression.compile o, LEVEL_OP
if @expression.unwrap() instanceof IdentifierLiteral and not o.scope.check code
[cmp, cnj] = if @negated then ['===', '||'] else ['!==', '&&']
code = "typeof #{code} #{cmp} \"undefined\"" + if @comparisonTarget isnt 'undefined' then " #{cnj} #{code} #{cmp} #{@comparisonTarget}" else ''
else
We explicity want to use loose equality (==
) when comparing against null
,
so that an existence check roughly corresponds to a check for truthiness.
Do not change this to ===
for null
, as this will break mountains of
existing code. When comparing only against undefined
, however, we want to
use ===
because this use case is for parity with ES2015+ default values,
which only get assigned when the variable is undefined
(but not null
).
cmp = if @comparisonTarget is 'null'
if @negated then '==' else '!='
else # `undefined`
if @negated then '===' else '!=='
code = "#{code} #{cmp} #{@comparisonTarget}"
[@makeCode(if o.level <= LEVEL_COND then code else "(#{code})")]
astType: -> 'UnaryExpression'
astProperties: (o) ->
return
argument: @expression.ast o
operator: '?'
prefix: no
An extra set of parentheses, specified explicitly in the source. At one time we tried to clean up the results by detecting and removing redundant parentheses, but no longer – you can put in as many as you please.
Parentheses are a good way to force any statement to become an expression.
exports.Parens = class Parens extends Base
constructor: (@body) ->
super()
children: ['body']
unwrap: -> @body
shouldCache: -> @body.shouldCache()
compileNode: (o) ->
expr = @body.unwrap()
If these parentheses are wrapping an IdentifierLiteral
followed by a
block comment, output the parentheses (or put another way, don’t optimize
away these redundant parentheses). This is because Flow requires
parentheses in certain circumstances to distinguish identifiers followed
by comment-based type annotations from JavaScript labels.
shouldWrapComment = expr.comments?.some(
(comment) -> comment.here and not comment.unshift and not comment.newLine)
if expr instanceof Value and expr.isAtomic() and not @jsxAttribute and not shouldWrapComment
expr.front = @front
return expr.compileToFragments o
fragments = expr.compileToFragments o, LEVEL_PAREN
bare = o.level < LEVEL_OP and not shouldWrapComment and (
expr instanceof Op and not expr.isInOperator() or expr.unwrap() instanceof Call or
(expr instanceof For and expr.returns)
) and (o.level < LEVEL_COND or fragments.length <= 3)
return @wrapInBraces fragments if @jsxAttribute
if bare then fragments else @wrapInParentheses fragments
astNode: (o) -> @body.unwrap().ast o, LEVEL_PAREN
exports.StringWithInterpolations = class StringWithInterpolations extends Base
constructor: (@body, {@quote, @startQuote, @jsxAttribute} = {}) ->
super()
@fromStringLiteral: (stringLiteral) ->
updatedString = stringLiteral.withoutQuotesInLocationData()
updatedStringValue = new Value(updatedString).withLocationDataFrom updatedString
new StringWithInterpolations Block.wrap([updatedStringValue]), quote: stringLiteral.quote, jsxAttribute: stringLiteral.jsxAttribute
.withLocationDataFrom stringLiteral
children: ['body']
unwrap
returns this
to stop ancestor nodes reaching in to grab @body,
and using @body.compileNode. StringWithInterpolations.compileNode
is
the custom logic to output interpolated strings as code.
unwrap: -> this
shouldCache: -> @body.shouldCache()
extractElements: (o, {includeInterpolationWrappers, isJsx} = {}) ->
Assumes that expr
is Block
expr = @body.unwrap()
elements = []
salvagedComments = []
expr.traverseChildren no, (node) =>
if node instanceof StringLiteral
if node.comments
salvagedComments.push node.comments...
delete node.comments
elements.push node
return yes
else if node instanceof Interpolation
if salvagedComments.length isnt 0
for comment in salvagedComments
comment.unshift = yes
comment.newLine = yes
attachCommentsToNode salvagedComments, node
if (unwrapped = node.expression?.unwrapAll()) instanceof PassthroughLiteral and unwrapped.generated and not (isJsx and o.compiling)
if o.compiling
commentPlaceholder = new StringLiteral('').withLocationDataFrom node
commentPlaceholder.comments = unwrapped.comments
(commentPlaceholder.comments ?= []).push node.comments... if node.comments
elements.push new Value commentPlaceholder
else
empty = new Interpolation().withLocationDataFrom node
empty.comments = node.comments
elements.push empty
else if node.expression or includeInterpolationWrappers
(node.expression?.comments ?= []).push node.comments... if node.comments
elements.push if includeInterpolationWrappers then node else node.expression
return no
else if node.comments
This node is getting discarded, but salvage its comments.
if elements.length isnt 0 and elements[elements.length - 1] not instanceof StringLiteral
for comment in node.comments
comment.unshift = no
comment.newLine = yes
attachCommentsToNode node.comments, elements[elements.length - 1]
else
salvagedComments.push node.comments...
delete node.comments
return yes
elements
compileNode: (o) ->
@comments ?= @startQuote?.comments
if @jsxAttribute
wrapped = new Parens new StringWithInterpolations @body
wrapped.jsxAttribute = yes
return wrapped.compileNode o
elements = @extractElements o, isJsx: @jsx
fragments = []
fragments.push @makeCode '`' unless @jsx
for element in elements
if element instanceof StringLiteral
unquotedElementValue = if @jsx then element.unquotedValueForJSX else element.unquotedValueForTemplateLiteral
fragments.push @makeCode unquotedElementValue
else
fragments.push @makeCode '$' unless @jsx
code = element.compileToFragments(o, LEVEL_PAREN)
if not @isNestedTag(element) or
code.some((fragment) -> fragment.comments?.some((comment) -> comment.here is no))
code = @wrapInBraces code
Flag the {
and }
fragments as having been generated by this
StringWithInterpolations
node, so that compileComments
knows
to treat them as bounds. But the braces are unnecessary if all of
the enclosed comments are /* */
comments. Don’t trust
fragment.type
, which can report minified variable names when
this compiler is minified.
code[0].isStringWithInterpolations = yes
code[code.length - 1].isStringWithInterpolations = yes
fragments.push code...
fragments.push @makeCode '`' unless @jsx
fragments
isNestedTag: (element) ->
call = element.unwrapAll?()
@jsx and call instanceof JSXElement
astType: -> 'TemplateLiteral'
astProperties: (o) ->
elements = @extractElements o, includeInterpolationWrappers: yes
[..., last] = elements
quasis = []
expressions = []
for element, index in elements
if element instanceof StringLiteral
quasis.push new TemplateElement(
element.originalValue
tail: element is last
).withLocationDataFrom(element).ast o
else # Interpolation
{expression} = element
node =
unless expression?
emptyInterpolation = new EmptyInterpolation()
emptyInterpolation.locationData = emptyExpressionLocationData {
interpolationNode: element
openingBrace: '#{'
closingBrace: '}'
}
emptyInterpolation
else
expression.unwrapAll()
expressions.push astAsBlockIfNeeded node, o
{expressions, quasis, @quote}
exports.TemplateElement = class TemplateElement extends Base
constructor: (@value, {@tail} = {}) ->
super()
astProperties: ->
return
value:
raw: @value
tail: !!@tail
exports.Interpolation = class Interpolation extends Base
constructor: (@expression) ->
super()
children: ['expression']
Represents the contents of an empty interpolation (e.g. #{}
).
Only used during AST generation.
exports.EmptyInterpolation = class EmptyInterpolation extends Base
constructor: ->
super()
CoffeeScript’s replacement for the for loop is our array and object comprehensions, that compile into for loops here. They also act as an expression, able to return the result of each filtered iteration.
Unlike Python array comprehensions, they can be multi-line, and you can pass the current index of the loop as a second parameter. Unlike Ruby blocks, you can map and filter in a single pass.
exports.For = class For extends While
constructor: (body, source) ->
super()
@addBody body
@addSource source
children: ['body', 'source', 'guard', 'step']
isAwait: -> @await ? no
addBody: (body) ->
@body = Block.wrap [body]
{expressions} = @body
if expressions.length
@body.locationData ?= mergeLocationData expressions[0].locationData, expressions[expressions.length - 1].locationData
this
addSource: (source) ->
{@source = no} = source
attribs = ["name", "index", "guard", "step", "own", "ownTag", "await", "awaitTag", "object", "from"]
@[attr] = source[attr] ? @[attr] for attr in attribs
return this unless @source
@index.error 'cannot use index with for-from' if @from and @index
@ownTag.error "cannot use own with for-#{if @from then 'from' else 'in'}" if @own and not @object
[@name, @index] = [@index, @name] if @object
@index.error 'index cannot be a pattern matching expression' if @index?.isArray?() or @index?.isObject?()
@awaitTag.error 'await must be used with for-from' if @await and not @from
@range = @source instanceof Value and @source.base instanceof Range and not @source.properties.length and not @from
@pattern = @name instanceof Value
@name.unwrap().propagateLhs?(yes) if @pattern
@index.error 'indexes do not apply to range loops' if @range and @index
@name.error 'cannot pattern match over range loops' if @range and @pattern
@returns = no
Move up any comments in the “for
line”, i.e. the line of code with for
,
from any child nodes of that line up to the for
node itself so that these
comments get output, and get output above the for
loop.
for attribute in ['source', 'guard', 'step', 'name', 'index'] when @[attribute]
@[attribute].traverseChildren yes, (node) =>
if node.comments
These comments are buried pretty deeply, so if they happen to be
trailing comments the line they trail will be unrecognizable when
we’re done compiling this for
loop; so just shift them up to
output above the for
line.
comment.newLine = comment.unshift = yes for comment in node.comments
moveComments node, @[attribute]
moveComments @[attribute], @
this
Welcome to the hairiest method in all of CoffeeScript. Handles the inner loop, filtering, stepping, and result saving for array, object, and range comprehensions. Some of the generated code can be shared in common, and some cannot.
compileNode: (o) ->
body = Block.wrap [@body]
[..., last] = body.expressions
@returns = no if last?.jumps() instanceof Return
source = if @range then @source.base else @source
scope = o.scope
name = @name and (@name.compile o, LEVEL_LIST) if not @pattern
index = @index and (@index.compile o, LEVEL_LIST)
scope.find(name) if name and not @pattern
scope.find(index) if index and @index not instanceof Value
rvar = scope.freeVariable 'results' if @returns
if @from
ivar = scope.freeVariable 'x', single: true if @pattern
else
ivar = (@object and index) or scope.freeVariable 'i', single: true
kvar = ((@range or @from) and name) or index or ivar
kvarAssign = if kvar isnt ivar then "#{kvar} = " else ""
if @step and not @range
[step, stepVar] = @cacheToCodeFragments @step.cache o, LEVEL_LIST, shouldCacheOrIsAssignable
stepNum = parseNumber stepVar if @step.isNumber()
name = ivar if @pattern
varPart = ''
guardPart = ''
defPart = ''
idt1 = @tab + TAB
if @range
forPartFragments = source.compileToFragments merge o,
{index: ivar, name, @step, shouldCache: shouldCacheOrIsAssignable}
else
svar = @source.compile o, LEVEL_LIST
if (name or @own) and not @from and @source.unwrap() not instanceof IdentifierLiteral
defPart += "#{@tab}#{ref = scope.freeVariable 'ref'} = #{svar};\n"
svar = ref
if name and not @pattern and not @from
namePart = "#{name} = #{svar}[#{kvar}]"
if not @object and not @from
defPart += "#{@tab}#{step};\n" if step isnt stepVar
down = stepNum < 0
lvar = scope.freeVariable 'len' unless @step and stepNum? and down
declare = "#{kvarAssign}#{ivar} = 0, #{lvar} = #{svar}.length"
declareDown = "#{kvarAssign}#{ivar} = #{svar}.length - 1"
compare = "#{ivar} < #{lvar}"
compareDown = "#{ivar} >= 0"
if @step
if stepNum?
if down
compare = compareDown
declare = declareDown
else
compare = "#{stepVar} > 0 ? #{compare} : #{compareDown}"
declare = "(#{stepVar} > 0 ? (#{declare}) : #{declareDown})"
increment = "#{ivar} += #{stepVar}"
else
increment = "#{if kvar isnt ivar then "++#{ivar}" else "#{ivar}++"}"
forPartFragments = [@makeCode("#{declare}; #{compare}; #{kvarAssign}#{increment}")]
if @returns
resultPart = "#{@tab}#{rvar} = [];\n"
returnResult = "\n#{@tab}return #{rvar};"
body.makeReturn rvar
if @guard
if body.expressions.length > 1
body.expressions.unshift new If (new Parens @guard).invert(), new StatementLiteral "continue"
else
body = Block.wrap [new If @guard, body] if @guard
if @pattern
body.expressions.unshift new Assign @name, if @from then new IdentifierLiteral kvar else new Literal "#{svar}[#{kvar}]"
varPart = "\n#{idt1}#{namePart};" if namePart
if @object
forPartFragments = [@makeCode("#{kvar} in #{svar}")]
guardPart = "\n#{idt1}if (!#{utility 'hasProp', o}.call(#{svar}, #{kvar})) continue;" if @own
else if @from
if @await
forPartFragments = new Op 'await', new Parens new Literal "#{kvar} of #{svar}"
forPartFragments = forPartFragments.compileToFragments o, LEVEL_TOP
else
forPartFragments = [@makeCode("#{kvar} of #{svar}")]
bodyFragments = body.compileToFragments merge(o, indent: idt1), LEVEL_TOP
if bodyFragments and bodyFragments.length > 0
bodyFragments = [].concat @makeCode('\n'), bodyFragments, @makeCode('\n')
fragments = [@makeCode(defPart)]
fragments.push @makeCode(resultPart) if resultPart
forCode = if @await then 'for ' else 'for ('
forClose = if @await then '' else ')'
fragments = fragments.concat @makeCode(@tab), @makeCode( forCode),
forPartFragments, @makeCode("#{forClose} {#{guardPart}#{varPart}"), bodyFragments,
@makeCode(@tab), @makeCode('}')
fragments.push @makeCode(returnResult) if returnResult
fragments
astNode: (o) ->
addToScope = (name) ->
alreadyDeclared = o.scope.find name.value
name.isDeclaration = not alreadyDeclared
@name?.eachName addToScope, checkAssignability: no
@index?.eachName addToScope, checkAssignability: no
super o
astType: -> 'For'
astProperties: (o) ->
return
source: @source?.ast o
body: @body.ast o, LEVEL_TOP
guard: @guard?.ast(o) ? null
name: @name?.ast(o) ? null
index: @index?.ast(o) ? null
step: @step?.ast(o) ? null
postfix: !!@postfix
own: !!@own
await: !!@await
style: switch
when @from then 'from'
when @object then 'of'
when @name then 'in'
else 'range'
A JavaScript switch statement. Converts into a returnable expression on-demand.
exports.Switch = class Switch extends Base
constructor: (@subject, @cases, @otherwise) ->
super()
children: ['subject', 'cases', 'otherwise']
isStatement: YES
jumps: (o = {block: yes}) ->
for {block} in @cases
return jumpNode if jumpNode = block.jumps o
@otherwise?.jumps o
makeReturn: (results, mark) ->
block.makeReturn(results, mark) for {block} in @cases
@otherwise or= new Block [new Literal 'void 0'] if results
@otherwise?.makeReturn results, mark
this
compileNode: (o) ->
idt1 = o.indent + TAB
idt2 = o.indent = idt1 + TAB
fragments = [].concat @makeCode(@tab + "switch ("),
(if @subject then @subject.compileToFragments(o, LEVEL_PAREN) else @makeCode "false"),
@makeCode(") {\n")
for {conditions, block}, i in @cases
for cond in flatten [conditions]
cond = cond.invert() unless @subject
fragments = fragments.concat @makeCode(idt1 + "case "), cond.compileToFragments(o, LEVEL_PAREN), @makeCode(":\n")
fragments = fragments.concat body, @makeCode('\n') if (body = block.compileToFragments o, LEVEL_TOP).length > 0
break if i is @cases.length - 1 and not @otherwise
expr = @lastNode block.expressions
continue if expr instanceof Return or expr instanceof Throw or (expr instanceof Literal and expr.jumps() and expr.value isnt 'debugger')
fragments.push cond.makeCode(idt2 + 'break;\n')
if @otherwise and @otherwise.expressions.length
fragments.push @makeCode(idt1 + "default:\n"), (@otherwise.compileToFragments o, LEVEL_TOP)..., @makeCode("\n")
fragments.push @makeCode @tab + '}'
fragments
astType: -> 'SwitchStatement'
casesAst: (o) ->
cases = []
for kase, caseIndex in @cases
{conditions: tests, block: consequent} = kase
tests = flatten [tests]
lastTestIndex = tests.length - 1
for test, testIndex in tests
testConsequent =
if testIndex is lastTestIndex
consequent
else
null
caseLocationData = test.locationData
caseLocationData = mergeLocationData caseLocationData, testConsequent.expressions[testConsequent.expressions.length - 1].locationData if testConsequent?.expressions.length
caseLocationData = mergeLocationData caseLocationData, kase.locationData, justLeading: yes if testIndex is 0
caseLocationData = mergeLocationData caseLocationData, kase.locationData, justEnding: yes if testIndex is lastTestIndex
cases.push new SwitchCase(test, testConsequent, trailing: testIndex is lastTestIndex).withLocationDataFrom locationData: caseLocationData
if @otherwise?.expressions.length
cases.push new SwitchCase(null, @otherwise).withLocationDataFrom @otherwise
kase.ast(o) for kase in cases
astProperties: (o) ->
return
discriminant: @subject?.ast(o, LEVEL_PAREN) ? null
cases: @casesAst o
class SwitchCase extends Base
constructor: (@test, @block, {@trailing} = {}) ->
super()
children: ['test', 'block']
astProperties: (o) ->
return
test: @test?.ast(o, LEVEL_PAREN) ? null
consequent: @block?.ast(o, LEVEL_TOP).body ? []
trailing: !!@trailing
exports.SwitchWhen = class SwitchWhen extends Base
constructor: (@conditions, @block) ->
super()
children: ['conditions', 'block']
If/else statements. Acts as an expression by pushing down requested returns to the last line of each clause.
Single-expression Ifs are compiled into conditional operators if possible, because ternaries are already proper expressions, and don’t need conversion.
exports.If = class If extends Base
constructor: (@condition, @body, options = {}) ->
super()
@elseBody = null
@isChain = false
{@soak, @postfix, @type} = options
moveComments @condition, @ if @condition.comments
children: ['condition', 'body', 'elseBody']
bodyNode: -> @body?.unwrap()
elseBodyNode: -> @elseBody?.unwrap()
Rewrite a chain of Ifs to add a default case as the final else.
addElse: (elseBody) ->
if @isChain
@elseBodyNode().addElse elseBody
@locationData = mergeLocationData @locationData, @elseBodyNode().locationData
else
@isChain = elseBody instanceof If
@elseBody = @ensureBlock elseBody
@elseBody.updateLocationDataIfMissing elseBody.locationData
@locationData = mergeLocationData @locationData, @elseBody.locationData if @locationData? and @elseBody.locationData?
this
The If only compiles into a statement if either of its bodies needs to be a statement. Otherwise a conditional operator is safe.
isStatement: (o) ->
o?.level is LEVEL_TOP or
@bodyNode().isStatement(o) or @elseBodyNode()?.isStatement(o)
jumps: (o) -> @body.jumps(o) or @elseBody?.jumps(o)
compileNode: (o) ->
if @isStatement o then @compileStatement o else @compileExpression o
makeReturn: (results, mark) ->
if mark
@body?.makeReturn results, mark
@elseBody?.makeReturn results, mark
return
@elseBody or= new Block [new Literal 'void 0'] if results
@body and= new Block [@body.makeReturn results]
@elseBody and= new Block [@elseBody.makeReturn results]
this
ensureBlock: (node) ->
if node instanceof Block then node else new Block [node]
Compile the If
as a regular if-else statement. Flattened chains
force inner else bodies into statement form.
compileStatement: (o) ->
child = del o, 'chainChild'
exeq = del o, 'isExistentialEquals'
if exeq
return new If(@processedCondition().invert(), @elseBodyNode(), type: 'if').compileToFragments o
indent = o.indent + TAB
cond = @processedCondition().compileToFragments o, LEVEL_PAREN
body = @ensureBlock(@body).compileToFragments merge o, {indent}
ifPart = [].concat @makeCode("if ("), cond, @makeCode(") {\n"), body, @makeCode("\n#{@tab}}")
ifPart.unshift @makeCode @tab unless child
return ifPart unless @elseBody
answer = ifPart.concat @makeCode(' else ')
if @isChain
o.chainChild = yes
answer = answer.concat @elseBody.unwrap().compileToFragments o, LEVEL_TOP
else
answer = answer.concat @makeCode("{\n"), @elseBody.compileToFragments(merge(o, {indent}), LEVEL_TOP), @makeCode("\n#{@tab}}")
answer
Compile the If
as a conditional operator.
compileExpression: (o) ->
cond = @processedCondition().compileToFragments o, LEVEL_COND
body = @bodyNode().compileToFragments o, LEVEL_LIST
alt = if @elseBodyNode() then @elseBodyNode().compileToFragments(o, LEVEL_LIST) else [@makeCode('void 0')]
fragments = cond.concat @makeCode(" ? "), body, @makeCode(" : "), alt
if o.level >= LEVEL_COND then @wrapInParentheses fragments else fragments
unfoldSoak: ->
@soak and this
processedCondition: ->
@processedConditionCache ?= if @type is 'unless' then @condition.invert() else @condition
isStatementAst: (o) ->
o.level is LEVEL_TOP
astType: (o) ->
if @isStatementAst o
'IfStatement'
else
'ConditionalExpression'
astProperties: (o) ->
isStatement = @isStatementAst o
return
test: @condition.ast o, if isStatement then LEVEL_PAREN else LEVEL_COND
consequent:
if isStatement
@body.ast o, LEVEL_TOP
else
@bodyNode().ast o, LEVEL_TOP
alternate:
if @isChain
@elseBody.unwrap().ast o, if isStatement then LEVEL_TOP else LEVEL_COND
else if not isStatement and @elseBody?.expressions?.length is 1
@elseBody.expressions[0].ast o, LEVEL_TOP
else
@elseBody?.ast(o, LEVEL_TOP) ? null
postfix: !!@postfix
inverted: @type is 'unless'
A sequence expression e.g. (a; b)
.
Currently only used during AST generation.
exports.Sequence = class Sequence extends Base
children: ['expressions']
constructor: (@expressions) ->
super()
astNode: (o) ->
return @expressions[0].ast(o) if @expressions.length is 1
super o
astType: -> 'SequenceExpression'
astProperties: (o) ->
return
expressions:
expression.ast(o) for expression in @expressions
UTILITIES =
modulo: -> 'function(a, b) { return (+a % (b = +b) + b) % b; }'
boundMethodCheck: -> "
function(instance, Constructor) {
if (!(instance instanceof Constructor)) {
throw new Error('Bound instance method accessed before binding');
}
}
"
Shortcuts to speed up the lookup time for native functions.
hasProp: -> '{}.hasOwnProperty'
indexOf: -> '[].indexOf'
slice : -> '[].slice'
splice : -> '[].splice'
Levels indicate a node’s position in the AST. Useful for knowing if parens are necessary or superfluous.
LEVEL_TOP = 1 # ...;
LEVEL_PAREN = 2 # (...)
LEVEL_LIST = 3 # [...]
LEVEL_COND = 4 # ... ? x : y
LEVEL_OP = 5 # !...
LEVEL_ACCESS = 6 # ...[0]
Tabs are two spaces for pretty printing.
TAB = ' '
SIMPLENUM = /^[+-]?\d+(?:_\d+)*$/
SIMPLE_STRING_OMIT = /\s*\n\s*/g
LEADING_BLANK_LINE = /^[^\n\S]*\n/
TRAILING_BLANK_LINE = /\n[^\n\S]*$/
STRING_OMIT = ///
((?:\\\\)+) # Consume (and preserve) an even number of backslashes.
| \\[^\S\n]*\n\s* # Remove escaped newlines.
///g
HEREGEX_OMIT = ///
((?:\\\\)+) # Consume (and preserve) an even number of backslashes.
| \\(\s) # Preserve escaped whitespace.
| \s+(?:#.*)? # Remove whitespace and comments.
///g
Helper for ensuring that utility functions are assigned at the top level.
utility = (name, o) ->
{root} = o.scope
if name of root.utilities
root.utilities[name]
else
ref = root.freeVariable name
root.assign ref, UTILITIES[name] o
root.utilities[name] = ref
multident = (code, tab, includingFirstLine = yes) ->
endsWithNewLine = code[code.length - 1] is '\n'
code = (if includingFirstLine then tab else '') + code.replace /\n/g, "$&#{tab}"
code = code.replace /\s+$/, ''
code = code + '\n' if endsWithNewLine
code
Wherever in CoffeeScript 1 we might’ve inserted a makeCode "#{@tab}"
to
indent a line of code, now we must account for the possibility of comments
preceding that line of code. If there are such comments, indent each line of
such comments, and then indent the first following line of code.
indentInitial = (fragments, node) ->
for fragment, fragmentIndex in fragments
if fragment.isHereComment
fragment.code = multident fragment.code, node.tab
else
fragments.splice fragmentIndex, 0, node.makeCode "#{node.tab}"
break
fragments
hasLineComments = (node) ->
return no unless node.comments
for comment in node.comments
return yes if comment.here is no
return no
Move the comments
property from one object to another, deleting it from
the first object.
moveComments = (from, to) ->
return unless from?.comments
attachCommentsToNode from.comments, to
delete from.comments
Sometimes when compiling a node, we want to insert a fragment at the start of an array of fragments; but if the start has one or more comment fragments, we want to insert this fragment after those but before any non-comments.
unshiftAfterComments = (fragments, fragmentToInsert) ->
inserted = no
for fragment, fragmentIndex in fragments when not fragment.isComment
fragments.splice fragmentIndex, 0, fragmentToInsert
inserted = yes
break
fragments.push fragmentToInsert unless inserted
fragments
isLiteralArguments = (node) ->
node instanceof IdentifierLiteral and node.value is 'arguments'
isLiteralThis = (node) ->
node instanceof ThisLiteral or (node instanceof Code and node.bound)
shouldCacheOrIsAssignable = (node) -> node.shouldCache() or node.isAssignable?()
Unfold a node’s child if soak, then tuck the node under created If
unfoldSoak = (o, parent, name) ->
return unless ifn = parent[name].unfoldSoak o
parent[name] = ifn.body
ifn.body = new Value parent
ifn
Constructs a string or regex by escaping certain characters.
makeDelimitedLiteral = (body, {delimiter: delimiterOption, escapeNewlines, double, includeDelimiters = yes, escapeDelimiter = yes, convertTrailingNullEscapes} = {}) ->
body = '(?:)' if body is '' and delimiterOption is '/'
escapeTemplateLiteralCurlies = delimiterOption is '`'
regex = ///
(\\\\) # Escaped backslash.
| (\\0(?=\d)) # Null character mistaken as octal escape.
#{
if convertTrailingNullEscapes
/// | (\\0) $ ///.source # Trailing null character that could be mistaken as octal escape.
else
''
}
#{
if escapeDelimiter
/// | \\?(#{delimiterOption}) ///.source # (Possibly escaped) delimiter.
else
''
}
#{
if escapeTemplateLiteralCurlies
/// | \\?(\$\{) ///.source # `${` inside template literals must be escaped.
else
''
}
| \\?(?:
#{if escapeNewlines then '(\n)|' else ''}
(\r)
| (\u2028)
| (\u2029)
) # (Possibly escaped) newlines.
| (\\.) # Other escapes.
///g
body = body.replace regex, (match, backslash, nul, ...args) ->
trailingNullEscape =
args.shift() if convertTrailingNullEscapes
delimiter =
args.shift() if escapeDelimiter
templateLiteralCurly =
args.shift() if escapeTemplateLiteralCurlies
lf =
args.shift() if escapeNewlines
[cr, ls, ps, other] = args
switch
Ignore escaped backslashes.
when backslash then (if double then backslash + backslash else backslash)
when nul then '\\x00'
when trailingNullEscape then "\\x00"
when delimiter then "\\#{delimiter}"
when templateLiteralCurly then "\\${"
when lf then '\\n'
when cr then '\\r'
when ls then '\\u2028'
when ps then '\\u2029'
when other then (if double then "\\#{other}" else other)
printedDelimiter = if includeDelimiters then delimiterOption else ''
"#{printedDelimiter}#{body}#{printedDelimiter}"
sniffDirectives = (expressions, {notFinalExpression} = {}) ->
index = 0
lastIndex = expressions.length - 1
while index <= lastIndex
break if index is lastIndex and notFinalExpression
expression = expressions[index]
if (unwrapped = expression?.unwrap?()) instanceof PassthroughLiteral and unwrapped.generated
index++
continue
break unless expression instanceof Value and expression.isString() and not expression.unwrap().shouldGenerateTemplateLiteral()
expressions[index] =
new Directive expression
.withLocationDataFrom expression
index++
astAsBlockIfNeeded = (node, o) ->
unwrapped = node.unwrap()
if unwrapped instanceof Block and unwrapped.expressions.length > 1
unwrapped.makeReturn null, yes
unwrapped.ast o, LEVEL_TOP
else
node.ast o, LEVEL_PAREN
Helpers for mergeLocationData
and mergeAstLocationData
below.
lesser = (a, b) -> if a < b then a else b
greater = (a, b) -> if a > b then a else b
isAstLocGreater = (a, b) ->
return yes if a.line > b.line
return no unless a.line is b.line
a.column > b.column
isLocationDataStartGreater = (a, b) ->
return yes if a.first_line > b.first_line
return no unless a.first_line is b.first_line
a.first_column > b.first_column
isLocationDataEndGreater = (a, b) ->
return yes if a.last_line > b.last_line
return no unless a.last_line is b.last_line
a.last_column > b.last_column
Take two nodes’ location data and return a new locationData
object that
encompasses the location data of both nodes. So the new first_line
value
will be the earlier of the two nodes’ first_line
values, the new
last_column
the later of the two nodes’ last_column
values, etc.
If you only want to extend the first node’s location data with the start or
end location data of the second node, pass the justLeading
or justEnding
options. So e.g. if first
’s range is [4, 5] and second
’s range is [1, 10],
you’d get:
mergeLocationData(first, second).range # [1, 10]
mergeLocationData(first, second, justLeading: yes).range # [1, 5]
mergeLocationData(first, second, justEnding: yes).range # [4, 10]
exports.mergeLocationData = mergeLocationData = (locationDataA, locationDataB, {justLeading, justEnding} = {}) ->
return Object.assign(
if justEnding
first_line: locationDataA.first_line
first_column: locationDataA.first_column
else
if isLocationDataStartGreater locationDataA, locationDataB
first_line: locationDataB.first_line
first_column: locationDataB.first_column
else
first_line: locationDataA.first_line
first_column: locationDataA.first_column
,
if justLeading
last_line: locationDataA.last_line
last_column: locationDataA.last_column
last_line_exclusive: locationDataA.last_line_exclusive
last_column_exclusive: locationDataA.last_column_exclusive
else
if isLocationDataEndGreater locationDataA, locationDataB
last_line: locationDataA.last_line
last_column: locationDataA.last_column
last_line_exclusive: locationDataA.last_line_exclusive
last_column_exclusive: locationDataA.last_column_exclusive
else
last_line: locationDataB.last_line
last_column: locationDataB.last_column
last_line_exclusive: locationDataB.last_line_exclusive
last_column_exclusive: locationDataB.last_column_exclusive
,
range: [
if justEnding
locationDataA.range[0]
else
lesser locationDataA.range[0], locationDataB.range[0]
,
if justLeading
locationDataA.range[1]
else
greater locationDataA.range[1], locationDataB.range[1]
]
)
Take two AST nodes, or two AST nodes’ location data objects, and return a new
location data object that encompasses the location data of both nodes. So the
new start
value will be the earlier of the two nodes’ start
values, the
new end
value will be the later of the two nodes’ end
values, etc.
If you only want to extend the first node’s location data with the start or
end location data of the second node, pass the justLeading
or justEnding
options. So e.g. if first
’s range is [4, 5] and second
’s range is [1, 10],
you’d get:
mergeAstLocationData(first, second).range # [1, 10]
mergeAstLocationData(first, second, justLeading: yes).range # [1, 5]
mergeAstLocationData(first, second, justEnding: yes).range # [4, 10]
exports.mergeAstLocationData = mergeAstLocationData = (nodeA, nodeB, {justLeading, justEnding} = {}) ->
return
loc:
start:
if justEnding
nodeA.loc.start
else
if isAstLocGreater nodeA.loc.start, nodeB.loc.start
nodeB.loc.start
else
nodeA.loc.start
end:
if justLeading
nodeA.loc.end
else
if isAstLocGreater nodeA.loc.end, nodeB.loc.end
nodeA.loc.end
else
nodeB.loc.end
range: [
if justEnding
nodeA.range[0]
else
lesser nodeA.range[0], nodeB.range[0]
,
if justLeading
nodeA.range[1]
else
greater nodeA.range[1], nodeB.range[1]
]
start:
if justEnding
nodeA.start
else
lesser nodeA.start, nodeB.start
end:
if justLeading
nodeA.end
else
greater nodeA.end, nodeB.end
Convert Jison-style node class location data to Babel-style location data
exports.jisonLocationDataToAstLocationData = jisonLocationDataToAstLocationData = ({first_line, first_column, last_line_exclusive, last_column_exclusive, range}) ->
return
loc:
start:
line: first_line + 1
column: first_column
end:
line: last_line_exclusive + 1
column: last_column_exclusive
range: [
range[0]
range[1]
]
start: range[0]
end: range[1]
Generate a zero-width location data that corresponds to the end of another node’s location.
zeroWidthLocationDataFromEndLocation = ({range: [, endRange], last_line_exclusive, last_column_exclusive}) -> {
first_line: last_line_exclusive
first_column: last_column_exclusive
last_line: last_line_exclusive
last_column: last_column_exclusive
last_line_exclusive
last_column_exclusive
range: [endRange, endRange]
}
extractSameLineLocationDataFirst = (numChars) -> ({range: [startRange], first_line, first_column}) -> {
first_line
first_column
last_line: first_line
last_column: first_column + numChars - 1
last_line_exclusive: first_line
last_column_exclusive: first_column + numChars
range: [startRange, startRange + numChars]
}
extractSameLineLocationDataLast = (numChars) -> ({range: [, endRange], last_line, last_column, last_line_exclusive, last_column_exclusive}) -> {
first_line: last_line
first_column: last_column - (numChars - 1)
last_line: last_line
last_column: last_column
last_line_exclusive
last_column_exclusive
range: [endRange - numChars, endRange]
}
We don’t currently have a token corresponding to the empty space between interpolation/JSX expression braces, so piece together the location data by trimming the braces from the Interpolation’s location data. Technically the last_line/last_column calculation here could be incorrect if the ending brace is preceded by a newline, but last_line/last_column aren’t used for AST generation anyway.
emptyExpressionLocationData = ({interpolationNode: element, openingBrace, closingBrace}) ->
first_line: element.locationData.first_line
first_column: element.locationData.first_column + openingBrace.length
last_line: element.locationData.last_line
last_column: element.locationData.last_column - closingBrace.length
last_line_exclusive: element.locationData.last_line
last_column_exclusive: element.locationData.last_column
range: [
element.locationData.range[0] + openingBrace.length
element.locationData.range[1] - closingBrace.length
]