cue: add token package
Change-Id: I5315cf8f30c2d7fc39423840fe10838e5fe75303
diff --git a/cue/token/position.go b/cue/token/position.go
new file mode 100644
index 0000000..f5ec34f
--- /dev/null
+++ b/cue/token/position.go
@@ -0,0 +1,564 @@
+// Copyright 2018 The CUE Authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+// http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+package token
+
+import (
+ "fmt"
+ "sort"
+ "sync"
+)
+
+// -----------------------------------------------------------------------------
+// Positions
+
+// Position describes an arbitrary source position
+// including the file, line, and column location.
+// A Position is valid if the line number is > 0.
+type Position struct {
+ Filename string // filename, if any
+ Offset int // offset, starting at 0
+ Line int // line number, starting at 1
+ Column int // column number, starting at 1 (byte count)
+ // RelPos Pos // relative position information
+}
+
+// IsValid reports whether the position is valid.
+func (pos *Position) IsValid() bool { return pos.Line > 0 }
+
+// String returns a string in one of several forms:
+//
+// file:line:column valid position with file name
+// line:column valid position without file name
+// file invalid position with file name
+// - invalid position without file name
+//
+func (pos Position) String() string {
+ s := pos.Filename
+ if pos.IsValid() {
+ if s != "" {
+ s += ":"
+ }
+ s += fmt.Sprintf("%d:%d", pos.Line, pos.Column)
+ }
+ if s == "" {
+ s = "-"
+ }
+ return s
+}
+
+// Pos is a compact encoding of a source position within a file set, as well as
+// relative positioning information. It can be converted into a Position for a
+// more convenient, but much larger, representation.
+//
+// The Pos value for a given file is a number in the range [base, base+size],
+// where base and size are specified when adding the file to the file set via
+// AddFile.
+//
+// To create the Pos value for a specific source offset (measured in bytes),
+// first add the respective file to the current file set using FileSet.AddFile
+// and then call File.Pos(offset) for that file. Given a Pos value p for a
+// specific file set fset, the corresponding Position value is obtained by
+// calling fset.Position(p).
+//
+// Pos values can be compared directly with the usual comparison operators: If
+// two Pos values p and q are in the same file, comparing p and q is equivalent
+// to comparing the respective source file offsets. If p and q are in different
+// files, p < q is true if the file implied by p was added to the respective
+// file set before the file implied by cue.
+type Pos int
+
+// NoPos is the zero value for Pos; there is no file and line information
+// associated with it, and NoPos().IsValid() is false. NoPos is always
+// smaller than any other Pos value. The corresponding Position value
+// for NoPos is the zero value for Position.
+const NoPos Pos = 0
+
+// RelPos indicates the relative position of token to the previous token.
+type RelPos int
+
+const (
+ // NoRelPos indicates no relative position is specified.
+ NoRelPos RelPos = iota
+
+ // Elided indicates that the token for which this position is defined is
+ // not rendered at all.
+ Elided
+
+ // NoSpace indicates there is no whitespace after this token.
+ NoSpace
+
+ // Blank means there is horizontal space after this token.
+ Blank
+
+ // Newline means there is a single newline after this token.
+ Newline
+
+ // NewSection means there are two or more newlines after this token.
+ NewSection
+
+ relMask Pos = 0xf
+ relShift = 4
+)
+
+var relNames = []string{
+ "invalid", "elided", "nospace", "blank", "newline", "section",
+}
+
+func (p RelPos) String() string { return relNames[p] }
+
+// HasRelPos repors whether p has a relative position.
+func (p Pos) HasRelPos() bool {
+ return p&relMask != 0
+
+}
+
+func (p Pos) Add(n int) Pos {
+ return p + toPos(index(n))
+}
+
+// IsValid reports whether the position is valid.
+func (p Pos) IsValid() bool {
+ return p != NoPos
+}
+
+// IsNewline reports whether the relative information suggests this node should
+// be printed on a new lien.
+func (p Pos) IsNewline() bool {
+ return p.RelPos() >= Newline
+}
+
+func (p Pos) WithRel(rel RelPos) Pos {
+ return p&^relMask | Pos(rel)
+}
+
+func (p Pos) RelPos() RelPos {
+ return RelPos(p & relMask)
+}
+
+func (p Pos) index() index {
+ return index(p) >> relShift
+}
+
+func toPos(x index) Pos {
+ return (Pos(x) << relShift)
+}
+
+// -----------------------------------------------------------------------------
+// File
+
+type index int
+
+// A File is a handle for a file belonging to a FileSet.
+// A File has a name, size, and line offset table.
+type File struct {
+ set *FileSet
+ name string // file name as provided to AddFile
+ base index // Pos index range for this file is [base...base+size]
+ size index // file size as provided to AddFile
+
+ // lines and infos are protected by set.mutex
+ lines []index // lines contains the offset of the first character for each line (the first entry is always 0)
+ infos []lineInfo
+}
+
+// Name returns the file name of file f as registered with AddFile.
+func (f *File) Name() string {
+ return f.name
+}
+
+// Base returns the base offset of file f as registered with AddFile.
+func (f *File) Base() int {
+ return int(f.base)
+}
+
+// Size returns the size of file f as registered with AddFile.
+func (f *File) Size() int {
+ return int(f.size)
+}
+
+// LineCount returns the number of lines in file f.
+func (f *File) LineCount() int {
+ f.set.mutex.RLock()
+ n := len(f.lines)
+ f.set.mutex.RUnlock()
+ return n
+}
+
+// AddLine adds the line offset for a new line.
+// The line offset must be larger than the offset for the previous line
+// and smaller than the file size; otherwise the line offset is ignored.
+//
+func (f *File) AddLine(offset int) {
+ x := index(offset)
+ f.set.mutex.Lock()
+ if i := len(f.lines); (i == 0 || f.lines[i-1] < x) && x < f.size {
+ f.lines = append(f.lines, x)
+ }
+ f.set.mutex.Unlock()
+}
+
+// MergeLine merges a line with the following line. It is akin to replacing
+// the newline character at the end of the line with a space (to not change the
+// remaining offsets). To obtain the line number, consult e.g. Position.Line.
+// MergeLine will panic if given an invalid line number.
+//
+func (f *File) MergeLine(line int) {
+ if line <= 0 {
+ panic("illegal line number (line numbering starts at 1)")
+ }
+ f.set.mutex.Lock()
+ defer f.set.mutex.Unlock()
+ if line >= len(f.lines) {
+ panic("illegal line number")
+ }
+ // To merge the line numbered <line> with the line numbered <line+1>,
+ // we need to remove the entry in lines corresponding to the line
+ // numbered <line+1>. The entry in lines corresponding to the line
+ // numbered <line+1> is located at index <line>, since indices in lines
+ // are 0-based and line numbers are 1-based.
+ copy(f.lines[line:], f.lines[line+1:])
+ f.lines = f.lines[:len(f.lines)-1]
+}
+
+// SetLines sets the line offsets for a file and reports whether it succeeded.
+// The line offsets are the offsets of the first character of each line;
+// for instance for the content "ab\nc\n" the line offsets are {0, 3}.
+// An empty file has an empty line offset table.
+// Each line offset must be larger than the offset for the previous line
+// and smaller than the file size; otherwise SetLines fails and returns
+// false.
+// Callers must not mutate the provided slice after SetLines returns.
+//
+func (f *File) SetLines(lines []int) bool {
+ // verify validity of lines table
+ size := f.size
+ for i, offset := range lines {
+ if i > 0 && offset <= lines[i-1] || size <= index(offset) {
+ return false
+ }
+ }
+
+ // set lines table
+ f.set.mutex.Lock()
+ f.lines = f.lines[:0]
+ for _, l := range lines {
+ f.lines = append(f.lines, index(l))
+ }
+ f.set.mutex.Unlock()
+ return true
+}
+
+// SetLinesForContent sets the line offsets for the given file content.
+// It ignores position-altering //line comments.
+func (f *File) SetLinesForContent(content []byte) {
+ var lines []index
+ line := index(0)
+ for offset, b := range content {
+ if line >= 0 {
+ lines = append(lines, line)
+ }
+ line = -1
+ if b == '\n' {
+ line = index(offset) + 1
+ }
+ }
+
+ // set lines table
+ f.set.mutex.Lock()
+ f.lines = lines
+ f.set.mutex.Unlock()
+}
+
+// A lineInfo object describes alternative file and line number
+// information (such as provided via a //line comment in a .go
+// file) for a given file offset.
+type lineInfo struct {
+ // fields are exported to make them accessible to gob
+ Offset int
+ Filename string
+ Line int
+}
+
+// AddLineInfo adds alternative file and line number information for
+// a given file offset. The offset must be larger than the offset for
+// the previously added alternative line info and smaller than the
+// file size; otherwise the information is ignored.
+//
+// AddLineInfo is typically used to register alternative position
+// information for //line filename:line comments in source files.
+//
+func (f *File) AddLineInfo(offset int, filename string, line int) {
+ x := index(offset)
+ f.set.mutex.Lock()
+ if i := len(f.infos); i == 0 || index(f.infos[i-1].Offset) < x && x < f.size {
+ f.infos = append(f.infos, lineInfo{offset, filename, line})
+ }
+ f.set.mutex.Unlock()
+}
+
+// Pos returns the Pos value for the given file offset;
+// the offset must be <= f.Size().
+// f.Pos(f.Offset(p)) == p.
+//
+func (f *File) Pos(offset int, rel RelPos) Pos {
+ if index(offset) > f.size {
+ panic("illegal file offset")
+ }
+ return toPos(f.base+index(offset)) + Pos(rel)
+}
+
+// Offset returns the offset for the given file position p;
+// p must be a valid Pos value in that file.
+// f.Offset(f.Pos(offset)) == offset.
+//
+func (f *File) Offset(p Pos) int {
+ x := p.index()
+ if x < f.base || x > f.base+index(f.size) {
+ panic("illegal Pos value")
+ }
+ return int(x - f.base)
+}
+
+// Line returns the line number for the given file position p;
+// p must be a Pos value in that file or NoPos.
+//
+func (f *File) Line(p Pos) int {
+ return f.Position(p).Line
+}
+
+func searchLineInfos(a []lineInfo, x int) int {
+ return sort.Search(len(a), func(i int) bool { return a[i].Offset > x }) - 1
+}
+
+// unpack returns the filename and line and column number for a file offset.
+// If adjusted is set, unpack will return the filename and line information
+// possibly adjusted by //line comments; otherwise those comments are ignored.
+//
+func (f *File) unpack(offset index, adjusted bool) (filename string, line, column int) {
+ filename = f.name
+ if i := searchInts(f.lines, offset); i >= 0 {
+ line, column = int(i+1), int(offset-f.lines[i]+1)
+ }
+ if adjusted && len(f.infos) > 0 {
+ // almost no files have extra line infos
+ if i := searchLineInfos(f.infos, int(offset)); i >= 0 {
+ alt := &f.infos[i]
+ filename = alt.Filename
+ if i := searchInts(f.lines, index(alt.Offset)); i >= 0 {
+ line += alt.Line - i - 1
+ }
+ }
+ }
+ return
+}
+
+func (f *File) position(p Pos, adjusted bool) (pos Position) {
+ offset := p.index() - f.base
+ pos.Offset = int(offset)
+ pos.Filename, pos.Line, pos.Column = f.unpack(offset, adjusted)
+ return
+}
+
+// PositionFor returns the Position value for the given file position p.
+// If adjusted is set, the position may be adjusted by position-altering
+// //line comments; otherwise those comments are ignored.
+// p must be a Pos value in f or NoPos.
+//
+func (f *File) PositionFor(p Pos, adjusted bool) (pos Position) {
+ x := p.index()
+ if p != NoPos {
+ if x < f.base || x > f.base+f.size {
+ panic("illegal Pos value")
+ }
+ pos = f.position(p, adjusted)
+ }
+ return
+}
+
+// Position returns the Position value for the given file position p.
+// Calling f.Position(p) is equivalent to calling f.PositionFor(p, true).
+//
+func (f *File) Position(p Pos) (pos Position) {
+ return f.PositionFor(p, true)
+}
+
+// A FileSet represents a set of source files.
+// Methods of file sets are synchronized; multiple goroutines
+// may invoke them concurrently.
+type FileSet struct {
+ mutex sync.RWMutex // protects the file set
+ base int // base offset for the next file
+ files []*File // list of files in the order added to the set
+ last *File // cache of last file looked up
+}
+
+// NewFileSet creates a new file set.
+func NewFileSet() *FileSet {
+ return &FileSet{
+ base: 1, // 0 == NoPos
+ }
+}
+
+// Base returns the minimum base offset that must be provided to
+// AddFile when adding the next file.
+func (s *FileSet) Base() int {
+ s.mutex.RLock()
+ b := s.base
+ s.mutex.RUnlock()
+ return b
+
+}
+
+// AddFile adds a new file with a given filename, base offset, and file size
+// to the file set s and returns the file. Multiple files may have the same
+// name. The base offset must not be smaller than the FileSet's Base(), and
+// size must not be negative. As a special case, if a negative base is provided,
+// the current value of the FileSet's Base() is used instead.
+//
+// Adding the file will set the file set's Base() value to base + size + 1
+// as the minimum base value for the next file. The following relationship
+// exists between a Pos value p for a given file offset offs:
+//
+// int(p) = base + offs
+//
+// with offs in the range [0, size] and thus p in the range [base, base+size].
+// For convenience, File.Pos may be used to create file-specific position
+// values from a file offset.
+func (s *FileSet) AddFile(filename string, base, size int) *File {
+ s.mutex.Lock()
+ defer s.mutex.Unlock()
+ if base < 0 {
+ base = s.base
+ }
+ if base < s.base || size < 0 {
+ panic("illegal base or size")
+ }
+ // base >= s.base && size >= 0
+ f := &File{s, filename, index(base), index(size), []index{0}, nil}
+ base += size + 1 // +1 because EOF also has a position
+ if base < 0 {
+ panic("token.Pos offset overflow (> 2G of source code in file set)")
+ }
+ // add the file to the file set
+ s.base = base
+ s.files = append(s.files, f)
+ s.last = f
+ return f
+}
+
+// Iterate calls f for the files in the file set in the order they were added
+// until f returns false.
+//
+func (s *FileSet) Iterate(f func(*File) bool) {
+ for i := 0; ; i++ {
+ var file *File
+ s.mutex.RLock()
+ if i < len(s.files) {
+ file = s.files[i]
+ }
+ s.mutex.RUnlock()
+ if file == nil || !f(file) {
+ break
+ }
+ }
+}
+
+func searchFiles(a []*File, x index) int {
+ return sort.Search(len(a), func(i int) bool { return a[i].base > x }) - 1
+}
+
+func (s *FileSet) file(p Pos) *File {
+ x := p.index()
+ s.mutex.RLock()
+ // common case: p is in last file
+ if f := s.last; f != nil && f.base <= x && x <= f.base+f.size {
+ s.mutex.RUnlock()
+ return f
+ }
+ // p is not in last file - search all files
+ if i := searchFiles(s.files, x); i >= 0 {
+ f := s.files[i]
+ // f.base <= int(p) by definition of searchFiles
+ if x <= f.base+f.size {
+ s.mutex.RUnlock()
+ s.mutex.Lock()
+ s.last = f // race is ok - s.last is only a cache
+ s.mutex.Unlock()
+ return f
+ }
+ }
+ s.mutex.RUnlock()
+ return nil
+}
+
+// File returns the file that contains the position p.
+// If no such file is found (for instance for p == NoPos),
+// the result is nil.
+//
+func (s *FileSet) File(p Pos) (f *File) {
+ if p.index() != 0 {
+ f = s.file(p)
+ }
+ return
+}
+
+// PositionFor converts a Pos p in the fileset into a Position value.
+// If adjusted is set, the position may be adjusted by position-altering
+// //line comments; otherwise those comments are ignored.
+// p must be a Pos value in s or NoPos.
+//
+func (s *FileSet) PositionFor(p Pos, adjusted bool) (pos Position) {
+ if p.index() != 0 {
+ if f := s.file(p); f != nil {
+ s.mutex.RLock()
+ pos = f.position(p, adjusted)
+ s.mutex.RUnlock()
+ }
+ }
+ return
+}
+
+// Position converts a Pos p in the fileset into a Position value.
+// Calling s.Position(p) is equivalent to calling s.PositionFor(p, true).
+//
+func (s *FileSet) Position(p Pos) (pos Position) {
+ return s.PositionFor(p, true)
+}
+
+// -----------------------------------------------------------------------------
+// Helper functions
+
+func searchInts(a []index, x index) int {
+ // This function body is a manually inlined version of:
+ //
+ // return sort.Search(len(a), func(i int) bool { return a[i] > x }) - 1
+ //
+ // With better compiler optimizations, this may not be needed in the
+ // future, but at the moment this change improves the go/printer
+ // benchmark performance by ~30%. This has a direct impact on the
+ // speed of gofmt and thus seems worthwhile (2011-04-29).
+ // TODO(gri): Remove this when compilers have caught up.
+ i, j := 0, len(a)
+ for i < j {
+ h := i + (j-i)/2 // avoid overflow when computing h
+ // i ≤ h < j
+ if a[h] <= x {
+ i = h + 1
+ } else {
+ j = h
+ }
+ }
+ return i - 1
+}
diff --git a/cue/token/position_test.go b/cue/token/position_test.go
new file mode 100644
index 0000000..906cac0
--- /dev/null
+++ b/cue/token/position_test.go
@@ -0,0 +1,336 @@
+// Copyright 2018 The CUE Authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+// http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+package token
+
+import (
+ "fmt"
+ "math/rand"
+ "sync"
+ "testing"
+)
+
+func checkPos(t *testing.T, msg string, got, want Position) {
+ if got.Filename != want.Filename {
+ t.Errorf("%s: got filename = %q; want %q", msg, got.Filename, want.Filename)
+ }
+ if got.Offset != want.Offset {
+ t.Errorf("%s: got offset = %d; want %d", msg, got.Offset, want.Offset)
+ }
+ if got.Line != want.Line {
+ t.Errorf("%s: got line = %d; want %d", msg, got.Line, want.Line)
+ }
+ if got.Column != want.Column {
+ t.Errorf("%s: got column = %d; want %d", msg, got.Column, want.Column)
+ }
+}
+
+func TestNoPos(t *testing.T) {
+ if NoPos.IsValid() {
+ t.Errorf("NoPos should not be valid")
+ }
+ var fset *FileSet
+ checkPos(t, "nil NoPos", fset.Position(NoPos), Position{})
+ fset = NewFileSet()
+ checkPos(t, "fset NoPos", fset.Position(NoPos), Position{})
+}
+
+var tests = []struct {
+ filename string
+ source []byte // may be nil
+ size int
+ lines []int
+}{
+ {"a", []byte{}, 0, []int{}},
+ {"b", []byte("01234"), 5, []int{0}},
+ {"c", []byte("\n\n\n\n\n\n\n\n\n"), 9, []int{0, 1, 2, 3, 4, 5, 6, 7, 8}},
+ {"d", nil, 100, []int{0, 5, 10, 20, 30, 70, 71, 72, 80, 85, 90, 99}},
+ {"e", nil, 777, []int{0, 80, 100, 120, 130, 180, 267, 455, 500, 567, 620}},
+ {"f", []byte("package p\n\nimport \"fmt\""), 23, []int{0, 10, 11}},
+ {"g", []byte("package p\n\nimport \"fmt\"\n"), 24, []int{0, 10, 11}},
+ {"h", []byte("package p\n\nimport \"fmt\"\n "), 25, []int{0, 10, 11, 24}},
+}
+
+func linecol(lines []int, offs int) (int, int) {
+ prevLineOffs := 0
+ for line, lineOffs := range lines {
+ if offs < lineOffs {
+ return line, offs - prevLineOffs + 1
+ }
+ prevLineOffs = lineOffs
+ }
+ return len(lines), offs - prevLineOffs + 1
+}
+
+func verifyPositions(t *testing.T, fset *FileSet, f *File, lines []int) {
+ for offs := 0; offs < f.Size(); offs++ {
+ p := f.Pos(offs, 0)
+ offs2 := f.Offset(p)
+ if offs2 != offs {
+ t.Errorf("%s, Offset: got offset %d; want %d", f.Name(), offs2, offs)
+ }
+ line, col := linecol(lines, offs)
+ msg := fmt.Sprintf("%s (offs = %d, p = %d)", f.Name(), offs, p)
+ checkPos(t, msg, f.Position(f.Pos(offs, 0)), Position{f.Name(), offs, line, col})
+ checkPos(t, msg, fset.Position(p), Position{f.Name(), offs, line, col})
+ }
+}
+
+func makeTestSource(size int, lines []int) []byte {
+ src := make([]byte, size)
+ for _, offs := range lines {
+ if offs > 0 {
+ src[offs-1] = '\n'
+ }
+ }
+ return src
+}
+
+func TestPositions(t *testing.T) {
+ const delta = 7 // a non-zero base offset increment
+ fset := NewFileSet()
+ for _, test := range tests {
+ // verify consistency of test case
+ if test.source != nil && len(test.source) != test.size {
+ t.Errorf("%s: inconsistent test case: got file size %d; want %d", test.filename, len(test.source), test.size)
+ }
+
+ // add file and verify name and size
+ f := fset.AddFile(test.filename, fset.Base()+delta, test.size)
+ if f.Name() != test.filename {
+ t.Errorf("got filename %q; want %q", f.Name(), test.filename)
+ }
+ if f.Size() != test.size {
+ t.Errorf("%s: got file size %d; want %d", f.Name(), f.Size(), test.size)
+ }
+ if fset.File(f.Pos(0, 0)) != f {
+ t.Errorf("%s: f.Pos(0, 0) was not found in f", f.Name())
+ }
+
+ // add lines individually and verify all positions
+ for i, offset := range test.lines {
+ f.AddLine(offset)
+ if f.LineCount() != i+1 {
+ t.Errorf("%s, AddLine: got line count %d; want %d", f.Name(), f.LineCount(), i+1)
+ }
+ // adding the same offset again should be ignored
+ f.AddLine(offset)
+ if f.LineCount() != i+1 {
+ t.Errorf("%s, AddLine: got unchanged line count %d; want %d", f.Name(), f.LineCount(), i+1)
+ }
+ verifyPositions(t, fset, f, test.lines[0:i+1])
+ }
+
+ // add lines with SetLines and verify all positions
+ if ok := f.SetLines(test.lines); !ok {
+ t.Errorf("%s: SetLines failed", f.Name())
+ }
+ if f.LineCount() != len(test.lines) {
+ t.Errorf("%s, SetLines: got line count %d; want %d", f.Name(), f.LineCount(), len(test.lines))
+ }
+ verifyPositions(t, fset, f, test.lines)
+
+ // add lines with SetLinesForContent and verify all positions
+ src := test.source
+ if src == nil {
+ // no test source available - create one from scratch
+ src = makeTestSource(test.size, test.lines)
+ }
+ f.SetLinesForContent(src)
+ if f.LineCount() != len(test.lines) {
+ t.Errorf("%s, SetLinesForContent: got line count %d; want %d", f.Name(), f.LineCount(), len(test.lines))
+ }
+ verifyPositions(t, fset, f, test.lines)
+ }
+}
+
+func TestLineInfo(t *testing.T) {
+ fset := NewFileSet()
+ f := fset.AddFile("foo", fset.Base(), 500)
+ lines := []int{0, 42, 77, 100, 210, 220, 277, 300, 333, 401}
+ // add lines individually and provide alternative line information
+ for _, offs := range lines {
+ f.AddLine(offs)
+ f.AddLineInfo(offs, "bar", 42)
+ }
+ // verify positions for all offsets
+ for offs := 0; offs <= f.Size(); offs++ {
+ p := f.Pos(offs, 0)
+ _, col := linecol(lines, offs)
+ msg := fmt.Sprintf("%s (offs = %d, p = %d)", f.Name(), offs, p)
+ checkPos(t, msg, f.Position(f.Pos(offs, 0)), Position{"bar", offs, 42, col})
+ checkPos(t, msg, fset.Position(p), Position{"bar", offs, 42, col})
+ }
+}
+
+func TestFiles(t *testing.T) {
+ fset := NewFileSet()
+ for i, test := range tests {
+ base := fset.Base()
+ if i%2 == 1 {
+ // Setting a negative base is equivalent to
+ // fset.Base(), so test some of each.
+ base = -1
+ }
+ fset.AddFile(test.filename, base, test.size)
+ j := 0
+ fset.Iterate(func(f *File) bool {
+ if f.Name() != tests[j].filename {
+ t.Errorf("got filename = %s; want %s", f.Name(), tests[j].filename)
+ }
+ j++
+ return true
+ })
+ if j != i+1 {
+ t.Errorf("got %d files; want %d", j, i+1)
+ }
+ }
+}
+
+// FileSet.File should return nil if Pos is past the end of the FileSet.
+func TestFileSetPastEnd(t *testing.T) {
+ fset := NewFileSet()
+ for _, test := range tests {
+ fset.AddFile(test.filename, fset.Base(), test.size)
+ }
+ if f := fset.File(toPos(index(fset.Base()))); f != nil {
+ t.Errorf("got %v, want nil", f)
+ }
+}
+
+func TestFileSetCacheUnlikely(t *testing.T) {
+ fset := NewFileSet()
+ offsets := make(map[string]index)
+ for _, test := range tests {
+ offsets[test.filename] = index(fset.Base())
+ fset.AddFile(test.filename, fset.Base(), test.size)
+ }
+ for file, pos := range offsets {
+ f := fset.File(toPos(pos))
+ if f.Name() != file {
+ t.Errorf("got %q at position %d, want %q", f.Name(), pos, file)
+ }
+ }
+}
+
+// issue 4345. Test that concurrent use of FileSet.Pos does not trigger a
+// race in the FileSet position cache.
+func TestFileSetRace(t *testing.T) {
+ fset := NewFileSet()
+ for i := 0; i < 100; i++ {
+ fset.AddFile(fmt.Sprintf("file-%d", i), fset.Base(), 1031)
+ }
+ max := int32(fset.Base())
+ var stop sync.WaitGroup
+ r := rand.New(rand.NewSource(7))
+ for i := 0; i < 2; i++ {
+ r := rand.New(rand.NewSource(r.Int63()))
+ stop.Add(1)
+ go func() {
+ for i := 0; i < 1000; i++ {
+ fset.Position(Pos(r.Int31n(max)))
+ }
+ stop.Done()
+ }()
+ }
+ stop.Wait()
+}
+
+// issue 16548. Test that concurrent use of File.AddLine and FileSet.PositionFor
+// does not trigger a race in the FileSet position cache.
+func TestFileSetRace2(t *testing.T) {
+ const N = 1e3
+ var (
+ fset = NewFileSet()
+ file = fset.AddFile("", -1, N)
+ ch = make(chan int, 2)
+ )
+
+ go func() {
+ for i := 0; i < N; i++ {
+ file.AddLine(i)
+ }
+ ch <- 1
+ }()
+
+ go func() {
+ pos := file.Pos(0, 0)
+ for i := 0; i < N; i++ {
+ fset.PositionFor(pos, false)
+ }
+ ch <- 1
+ }()
+
+ <-ch
+ <-ch
+}
+
+func TestPositionFor(t *testing.T) {
+ src := []byte(`
+foo
+b
+ar
+//line :100
+foobar
+//line bar:3
+done
+`)
+
+ const filename = "foo"
+ fset := NewFileSet()
+ f := fset.AddFile(filename, fset.Base(), len(src))
+ f.SetLinesForContent(src)
+
+ // verify position info
+ for i, offs := range f.lines {
+ got1 := f.PositionFor(f.Pos(int(offs), 0), false)
+ got2 := f.PositionFor(f.Pos(int(offs), 0), true)
+ got3 := f.Position(f.Pos(int(offs), 0))
+ want := Position{filename, int(offs), i + 1, 1}
+ checkPos(t, "1. PositionFor unadjusted", got1, want)
+ checkPos(t, "1. PositionFor adjusted", got2, want)
+ checkPos(t, "1. Position", got3, want)
+ }
+
+ // manually add //line info on lines l1, l2
+ const l1, l2 = 5, 7
+ f.AddLineInfo(int(f.lines[l1-1]), "", 100)
+ f.AddLineInfo(int(f.lines[l2-1]), "bar", 3)
+
+ // unadjusted position info must remain unchanged
+ for i, offs := range f.lines {
+ got1 := f.PositionFor(f.Pos(int(offs), 0), false)
+ want := Position{filename, int(offs), i + 1, 1}
+ checkPos(t, "2. PositionFor unadjusted", got1, want)
+ }
+
+ // adjusted position info should have changed
+ for i, offs := range f.lines {
+ got2 := f.PositionFor(f.Pos(int(offs), 0), true)
+ got3 := f.Position(f.Pos(int(offs), 0))
+ want := Position{filename, int(offs), i + 1, 1}
+ // manually compute wanted filename and line
+ line := want.Line
+ if i+1 >= l1 {
+ want.Filename = ""
+ want.Line = line - l1 + 100
+ }
+ if i+1 >= l2 {
+ want.Filename = "bar"
+ want.Line = line - l2 + 3
+ }
+ checkPos(t, "3. PositionFor adjusted", got2, want)
+ checkPos(t, "3. Position", got3, want)
+ }
+}
diff --git a/cue/token/token.go b/cue/token/token.go
new file mode 100644
index 0000000..ef50c34
--- /dev/null
+++ b/cue/token/token.go
@@ -0,0 +1,263 @@
+// Copyright 2018 The CUE Authors
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+// http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Package token defines constants representing the lexical tokens of the Go
+// programming language and basic operations on tokens (printing, predicates).
+package token // import "cuelang.org/go/cue/token"
+
+import "strconv"
+
+// Token is the set of lexical tokens of the CUE configuration language.
+type Token int
+
+// The list of tokens.
+const (
+ // Special tokens
+ ILLEGAL Token = iota
+ EOF
+ COMMENT
+
+ literalBeg
+ // Identifiers and basic type literals
+ // (these tokens stand for classes of literals)
+ IDENT // main, _tmp
+ INT // 12_345Mi, 0700, 0xdeadbeef, 1.2M
+ FLOAT // 123.45,
+ // DURATION // 3m4s TODO
+ STRING // "abc"
+ INTERPOLATION // a part of a template string, e.g. `"age: \(`
+ BOTTOM // _|_
+
+ literalEnd
+
+ operatorBeg
+ // Operators and delimiters
+ ADD // +
+ SUB // -
+ MUL // *
+ POW // ^
+ QUO // /
+ REM // %
+
+ IQUO // quo
+ IREM // rem
+ IDIV // div
+ IMOD // mod
+
+ UNIFY // &
+ DISJUNCTION // |
+
+ LAND // &&
+ LOR // ||
+
+ BIND // =
+ EQL // ==
+ LSS // <
+ GTR // >
+ NOT // !
+ ARROW // <-
+ LAMBDA // ->
+
+ NEQ // !=
+ LEQ // <=
+ GEQ // >=
+
+ LPAREN // (
+ LBRACK // [
+ LBRACE // {
+ COMMA // ,
+ PERIOD // .
+ RANGE // ..
+ ELLIPSIS // ...
+
+ RPAREN // )
+ RBRACK // ]
+ RBRACE // }
+ SEMICOLON // :
+ COLON // :
+ operatorEnd
+
+ keywordBeg
+
+ IF
+ FOR
+ IN
+ LET
+
+ TRUE
+ FALSE
+ NULL
+
+ keywordEnd
+)
+
+var tokens = [...]string{
+ ILLEGAL: "ILLEGAL",
+
+ EOF: "EOF",
+ COMMENT: "COMMENT",
+
+ IDENT: "IDENT",
+ INT: "INT",
+ FLOAT: "FLOAT",
+ // DURATION: "DURATION", // TODO
+ STRING: "STRING",
+ INTERPOLATION: "INTERPOLATION",
+
+ ADD: "+",
+ SUB: "-",
+ MUL: "*",
+ POW: "^",
+ QUO: "/",
+ REM: "%",
+
+ IQUO: "quo",
+ IREM: "rem",
+ IDIV: "div",
+ IMOD: "mod",
+
+ UNIFY: "&",
+ DISJUNCTION: "|",
+
+ LAND: "&&",
+ LOR: "||",
+
+ BIND: "=",
+ EQL: "==",
+ LSS: "<",
+ GTR: ">",
+ NOT: "!",
+ ARROW: "<-",
+ LAMBDA: "->",
+
+ NEQ: "!=",
+ LEQ: "<=",
+ GEQ: ">=",
+
+ LPAREN: "(",
+ LBRACK: "[",
+ LBRACE: "{",
+ COMMA: ",",
+ PERIOD: ".",
+ RANGE: "..",
+ ELLIPSIS: "...",
+
+ RPAREN: ")",
+ RBRACK: "]",
+ RBRACE: "}",
+ SEMICOLON: ";",
+ COLON: ":",
+
+ BOTTOM: "_|_",
+
+ FALSE: "false",
+ TRUE: "true",
+ NULL: "null",
+
+ FOR: "for",
+ IF: "if",
+ IN: "in",
+ LET: "let",
+}
+
+// String returns the string corresponding to the token tok.
+// For operators, delimiters, and keywords the string is the actual
+// token character sequence (e.g., for the token ADD, the string is
+// "+"). For all other tokens the string corresponds to the token
+// constant name (e.g. for the token IDENT, the string is "IDENT").
+func (tok Token) String() string {
+ s := ""
+ if 0 <= tok && tok < Token(len(tokens)) {
+ s = tokens[tok]
+ }
+ if s == "" {
+ s = "token(" + strconv.Itoa(int(tok)) + ")"
+ }
+ return s
+}
+
+// A set of constants for precedence-based expression parsing.
+// Non-operators have lowest precedence, followed by operators
+// starting with precedence 1 up to unary operators. The highest
+// precedence serves as "catch-all" precedence for selector,
+// indexing, and other operator and delimiter tokens.
+const (
+ LowestPrec = lowestPrec
+ UnaryPrec = unaryPrec
+ HighestPrec = highestPrec
+)
+
+const (
+ lowestPrec = 0 // non-operators
+ unaryPrec = 8
+ highestPrec = 9
+)
+
+// Precedence returns the operator precedence of the binary
+// operator op. If op is not a binary operator, the result
+// is LowestPrecedence.
+//
+func (tok Token) Precedence() int {
+ switch tok {
+ case DISJUNCTION:
+ return 1
+ case UNIFY:
+ return 2
+ case LOR:
+ return 3
+ case LAND:
+ return 4
+ case EQL, NEQ, LSS, LEQ, GTR, GEQ:
+ return 5
+ case ADD, SUB:
+ return 6
+ case MUL, QUO, REM, IDIV, IMOD, IQUO, IREM:
+ return 7
+ case RANGE:
+ return 8
+ }
+ return lowestPrec
+}
+
+var keywords map[string]Token
+
+func init() {
+ keywords = make(map[string]Token)
+ for i := keywordBeg + 1; i < keywordEnd; i++ {
+ keywords[tokens[i]] = i
+ }
+}
+
+// Lookup maps an identifier to its keyword token or IDENT (if not a keyword).
+//
+func Lookup(ident string) Token {
+ if tok, isKeyword := keywords[ident]; isKeyword {
+ return tok
+ }
+ return IDENT
+}
+
+// Predicates
+
+// IsLiteral returns true for tokens corresponding to identifiers
+// and basic type literals; it returns false otherwise.
+func (tok Token) IsLiteral() bool { return literalBeg < tok && tok < literalEnd }
+
+// IsOperator returns true for tokens corresponding to operators and
+// delimiters; it returns false otherwise.
+func (tok Token) IsOperator() bool { return operatorBeg < tok && tok < operatorEnd }
+
+// IsKeyword returns true for tokens corresponding to keywords;
+// it returns false otherwise.
+func (tok Token) IsKeyword() bool { return keywordBeg < tok && tok < keywordEnd }