1// Copyright 2010 The Go Authors. All rights reserved.
  2// Use of this source code is governed by a BSD-style
  3// license that can be found in the LICENSE file.
  4
  5// TODO(gri) consider making this a separate package outside the go directory.
  6
  7package token
  8
  9import (
 10	"fmt"
 11	"sort"
 12	"sync"
 13)
 14
 15// -----------------------------------------------------------------------------
 16// Positions
 17
 18// Position describes an arbitrary source position
 19// including the file, line, and column location.
 20// A Position is valid if the line number is > 0.
 21//
 22type Position struct {
 23	Filename string // filename, if any
 24	Offset   int    // offset, starting at 0
 25	Line     int    // line number, starting at 1
 26	Column   int    // column number, starting at 1 (character count)
 27}
 28
 29// IsValid returns true if the position is valid.
 30func (pos *Position) IsValid() bool { return pos.Line > 0 }
 31
 32// String returns a string in one of several forms:
 33//
 34//	file:line:column    valid position with file name
 35//	line:column         valid position without file name
 36//	file                invalid position with file name
 37//	-                   invalid position without file name
 38//
 39func (pos Position) String() string {
 40	s := pos.Filename
 41	if pos.IsValid() {
 42		if s != "" {
 43			s += ":"
 44		}
 45		s += fmt.Sprintf("%d:%d", pos.Line, pos.Column)
 46	}
 47	if s == "" {
 48		s = "-"
 49	}
 50	return s
 51}
 52
 53// Pos is a compact encoding of a source position within a file set.
 54// It can be converted into a Position for a more convenient, but much
 55// larger, representation.
 56//
 57// The Pos value for a given file is a number in the range [base, base+size],
 58// where base and size are specified when adding the file to the file set via
 59// AddFile.
 60//
 61// To create the Pos value for a specific source offset, first add
 62// the respective file to the current file set (via FileSet.AddFile)
 63// and then call File.Pos(offset) for that file. Given a Pos value p
 64// for a specific file set fset, the corresponding Position value is
 65// obtained by calling fset.Position(p).
 66//
 67// Pos values can be compared directly with the usual comparison operators:
 68// If two Pos values p and q are in the same file, comparing p and q is
 69// equivalent to comparing the respective source file offsets. If p and q
 70// are in different files, p < q is true if the file implied by p was added
 71// to the respective file set before the file implied by q.
 72//
 73type Pos int
 74
 75// The zero value for Pos is NoPos; there is no file and line information
 76// associated with it, and NoPos().IsValid() is false. NoPos is always
 77// smaller than any other Pos value. The corresponding Position value
 78// for NoPos is the zero value for Position.
 79//
 80const NoPos Pos = 0
 81
 82// IsValid returns true if the position is valid.
 83func (p Pos) IsValid() bool {
 84	return p != NoPos
 85}
 86
 87// -----------------------------------------------------------------------------
 88// File
 89
 90// A File is a handle for a file belonging to a FileSet.
 91// A File has a name, size, and line offset table.
 92//
 93type File struct {
 94	set  *FileSet
 95	name string // file name as provided to AddFile
 96	base int    // Pos value range for this file is [base...base+size]
 97	size int    // file size as provided to AddFile
 98
 99	// lines and infos are protected by set.mutex
100	lines []int
101	infos []lineInfo
102}
103
104// Name returns the file name of file f as registered with AddFile.
105func (f *File) Name() string {
106	return f.name
107}
108
109// Base returns the base offset of file f as registered with AddFile.
110func (f *File) Base() int {
111	return f.base
112}
113
114// Size returns the size of file f as registered with AddFile.
115func (f *File) Size() int {
116	return f.size
117}
118
119// LineCount returns the number of lines in file f.
120func (f *File) LineCount() int {
121	f.set.mutex.RLock()
122	n := len(f.lines)
123	f.set.mutex.RUnlock()
124	return n
125}
126
127// AddLine adds the line offset for a new line.
128// The line offset must be larger than the offset for the previous line
129// and smaller than the file size; otherwise the line offset is ignored.
130//
131func (f *File) AddLine(offset int) {
132	f.set.mutex.Lock()
133	if i := len(f.lines); (i == 0 || f.lines[i-1] < offset) && offset < f.size {
134		f.lines = append(f.lines, offset)
135	}
136	f.set.mutex.Unlock()
137}
138
139// SetLines sets the line offsets for a file and returns true if successful.
140// The line offsets are the offsets of the first character of each line;
141// for instance for the content "ab\nc\n" the line offsets are {0, 3}.
142// An empty file has an empty line offset table.
143// Each line offset must be larger than the offset for the previous line
144// and smaller than the file size; otherwise SetLines fails and returns
145// false.
146//
147func (f *File) SetLines(lines []int) bool {
148	// verify validity of lines table
149	size := f.size
150	for i, offset := range lines {
151		if i > 0 && offset <= lines[i-1] || size <= offset {
152			return false
153		}
154	}
155
156	// set lines table
157	f.set.mutex.Lock()
158	f.lines = lines
159	f.set.mutex.Unlock()
160	return true
161}
162
163// SetLinesForContent sets the line offsets for the given file content.
164func (f *File) SetLinesForContent(content []byte) {
165	var lines []int
166	line := 0
167	for offset, b := range content {
168		if line >= 0 {
169			lines = append(lines, line)
170		}
171		line = -1
172		if b == '\n' {
173			line = offset + 1
174		}
175	}
176
177	// set lines table
178	f.set.mutex.Lock()
179	f.lines = lines
180	f.set.mutex.Unlock()
181}
182
183// A lineInfo object describes alternative file and line number
184// information (such as provided via a //line comment in a .go
185// file) for a given file offset.
186type lineInfo struct {
187	// fields are exported to make them accessible to gob
188	Offset   int
189	Filename string
190	Line     int
191}
192
193// AddLineInfo adds alternative file and line number information for
194// a given file offset. The offset must be larger than the offset for
195// the previously added alternative line info and smaller than the
196// file size; otherwise the information is ignored.
197//
198// AddLineInfo is typically used to register alternative position
199// information for //line filename:line comments in source files.
200//
201func (f *File) AddLineInfo(offset int, filename string, line int) {
202	f.set.mutex.Lock()
203	if i := len(f.infos); i == 0 || f.infos[i-1].Offset < offset && offset < f.size {
204		f.infos = append(f.infos, lineInfo{offset, filename, line})
205	}
206	f.set.mutex.Unlock()
207}
208
209// Pos returns the Pos value for the given file offset;
210// the offset must be <= f.Size().
211// f.Pos(f.Offset(p)) == p.
212//
213func (f *File) Pos(offset int) Pos {
214	if offset > f.size {
215		panic("illegal file offset")
216	}
217	return Pos(f.base + offset)
218}
219
220// Offset returns the offset for the given file position p;
221// p must be a valid Pos value in that file.
222// f.Offset(f.Pos(offset)) == offset.
223//
224func (f *File) Offset(p Pos) int {
225	if int(p) < f.base || int(p) > f.base+f.size {
226		panic("illegal Pos value")
227	}
228	return int(p) - f.base
229}
230
231// Line returns the line number for the given file position p;
232// p must be a Pos value in that file or NoPos.
233//
234func (f *File) Line(p Pos) int {
235	// TODO(gri) this can be implemented much more efficiently
236	return f.Position(p).Line
237}
238
239func searchLineInfos(a []lineInfo, x int) int {
240	return sort.Search(len(a), func(i int) bool { return a[i].Offset > x }) - 1
241}
242
243// info returns the file name, line, and column number for a file offset.
244func (f *File) info(offset int) (filename string, line, column int) {
245	filename = f.name
246	if i := searchInts(f.lines, offset); i >= 0 {
247		line, column = i+1, offset-f.lines[i]+1
248	}
249	if len(f.infos) > 0 {
250		// almost no files have extra line infos
251		if i := searchLineInfos(f.infos, offset); i >= 0 {
252			alt := &f.infos[i]
253			filename = alt.Filename
254			if i := searchInts(f.lines, alt.Offset); i >= 0 {
255				line += alt.Line - i - 1
256			}
257		}
258	}
259	return
260}
261
262func (f *File) position(p Pos) (pos Position) {
263	offset := int(p) - f.base
264	pos.Offset = offset
265	pos.Filename, pos.Line, pos.Column = f.info(offset)
266	return
267}
268
269// Position returns the Position value for the given file position p;
270// p must be a Pos value in that file or NoPos.
271//
272func (f *File) Position(p Pos) (pos Position) {
273	if p != NoPos {
274		if int(p) < f.base || int(p) > f.base+f.size {
275			panic("illegal Pos value")
276		}
277		pos = f.position(p)
278	}
279	return
280}
281
282// -----------------------------------------------------------------------------
283// FileSet
284
285// A FileSet represents a set of source files.
286// Methods of file sets are synchronized; multiple goroutines
287// may invoke them concurrently.
288//
289type FileSet struct {
290	mutex sync.RWMutex // protects the file set
291	base  int          // base offset for the next file
292	files []*File      // list of files in the order added to the set
293	last  *File        // cache of last file looked up
294}
295
296// NewFileSet creates a new file set.
297func NewFileSet() *FileSet {
298	s := new(FileSet)
299	s.base = 1 // 0 == NoPos
300	return s
301}
302
303// Base returns the minimum base offset that must be provided to
304// AddFile when adding the next file.
305//
306func (s *FileSet) Base() int {
307	s.mutex.RLock()
308	b := s.base
309	s.mutex.RUnlock()
310	return b
311
312}
313
314// AddFile adds a new file with a given filename, base offset, and file size
315// to the file set s and returns the file. Multiple files may have the same
316// name. The base offset must not be smaller than the FileSet's Base(), and
317// size must not be negative.
318//
319// Adding the file will set the file set's Base() value to base + size + 1
320// as the minimum base value for the next file. The following relationship
321// exists between a Pos value p for a given file offset offs:
322//
323//	int(p) = base + offs
324//
325// with offs in the range [0, size] and thus p in the range [base, base+size].
326// For convenience, File.Pos may be used to create file-specific position
327// values from a file offset.
328//
329func (s *FileSet) AddFile(filename string, base, size int) *File {
330	s.mutex.Lock()
331	defer s.mutex.Unlock()
332	if base < s.base || size < 0 {
333		panic("illegal base or size")
334	}
335	// base >= s.base && size >= 0
336	f := &File{s, filename, base, size, []int{0}, nil}
337	base += size + 1 // +1 because EOF also has a position
338	if base < 0 {
339		panic("token.Pos offset overflow (> 2G of source code in file set)")
340	}
341	// add the file to the file set
342	s.base = base
343	s.files = append(s.files, f)
344	s.last = f
345	return f
346}
347
348// Iterate calls f for the files in the file set in the order they were added
349// until f returns false.
350//
351func (s *FileSet) Iterate(f func(*File) bool) {
352	for i := 0; ; i++ {
353		var file *File
354		s.mutex.RLock()
355		if i < len(s.files) {
356			file = s.files[i]
357		}
358		s.mutex.RUnlock()
359		if file == nil || !f(file) {
360			break
361		}
362	}
363}
364
365func searchFiles(a []*File, x int) int {
366	return sort.Search(len(a), func(i int) bool { return a[i].base > x }) - 1
367}
368
369func (s *FileSet) file(p Pos) *File {
370	// common case: p is in last file
371	if f := s.last; f != nil && f.base <= int(p) && int(p) <= f.base+f.size {
372		return f
373	}
374	// p is not in last file - search all files
375	if i := searchFiles(s.files, int(p)); i >= 0 {
376		f := s.files[i]
377		// f.base <= int(p) by definition of searchFiles
378		if int(p) <= f.base+f.size {
379			s.last = f
380			return f
381		}
382	}
383	return nil
384}
385
386// File returns the file that contains the position p.
387// If no such file is found (for instance for p == NoPos),
388// the result is nil.
389//
390func (s *FileSet) File(p Pos) (f *File) {
391	if p != NoPos {
392		s.mutex.RLock()
393		f = s.file(p)
394		s.mutex.RUnlock()
395	}
396	return
397}
398
399// Position converts a Pos in the fileset into a general Position.
400func (s *FileSet) Position(p Pos) (pos Position) {
401	if p != NoPos {
402		s.mutex.RLock()
403		if f := s.file(p); f != nil {
404			pos = f.position(p)
405		}
406		s.mutex.RUnlock()
407	}
408	return
409}
410
411// -----------------------------------------------------------------------------
412// Helper functions
413
414func searchInts(a []int, x int) int {
415	// This function body is a manually inlined version of:
416	//
417	//   return sort.Search(len(a), func(i int) bool { return a[i] > x }) - 1
418	//
419	// With better compiler optimizations, this may not be needed in the
420	// future, but at the moment this change improves the go/printer
421	// benchmark performance by ~30%. This has a direct impact on the
422	// speed of gofmt and thus seems worthwhile (2011-04-29).
423	// TODO(gri): Remove this when compilers have caught up.
424	i, j := 0, len(a)
425	for i < j {
426		h := i + (j-i)/2 // avoid overflow when computing h
427		// i ≤ h < j
428		if a[h] <= x {
429			i = h + 1
430		} else {
431			j = h
432		}
433	}
434	return i - 1
435}