Repository navigation
Expand file tree
/
Copy pathindex.go
More file actions
1580 lines (1402 loc) · 42.2 KB
/
Copy pathindex.go
File metadata and controls
1580 lines (1402 loc) · 42.2 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
// Copyright 2009 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// This file contains the infrastructure to create an
// identifier and full-text index for a set of Go files.
//
// Algorithm for identifier index:
// - traverse all .go files of the file tree specified by root
// - for each identifier (word) encountered, collect all occurrences (spots)
// into a list; this produces a list of spots for each word
// - reduce the lists: from a list of spots to a list of FileRuns,
// and from a list of FileRuns into a list of PakRuns
// - make a HitList from the PakRuns
//
// Details:
// - keep two lists per word: one containing package-level declarations
// that have snippets, and one containing all other spots
// - keep the snippets in a separate table indexed by snippet index
// and store the snippet index in place of the line number in a SpotInfo
// (the line number for spots with snippets is stored in the snippet)
// - at the end, create lists of alternative spellings for a given
// word
//
// Algorithm for full text index:
// - concatenate all source code in a byte buffer (in memory)
// - add the files to a file set in lockstep as they are added to the byte
// buffer such that a byte buffer offset corresponds to the Pos value for
// that file location
// - create a suffix array from the concatenated sources
//
// String lookup in full text index:
// - use the suffix array to lookup a string's offsets - the offsets
// correspond to the Pos values relative to the file set
// - translate the Pos values back into file and line information and
// sort the result
package godoc
import (
"bufio"
"bytes"
"encoding/gob"
"errors"
"fmt"
"go/ast"
"go/doc"
"go/parser"
"go/token"
"index/suffixarray"
"io"
"log"
"os"
pathpkg "path"
"path/filepath"
"regexp"
"runtime"
"sort"
"strconv"
"strings"
"sync"
"time"
"unicode"
"golang.org/x/tools/godoc/util"
"golang.org/x/tools/godoc/vfs"
)
// ----------------------------------------------------------------------------
// InterfaceSlice is a helper type for sorting interface
// slices according to some slice-specific sort criteria.
type comparer func(x, y interface{}) bool
type interfaceSlice struct {
slice []interface{}
less comparer
}
// ----------------------------------------------------------------------------
// RunList
// A RunList is a list of entries that can be sorted according to some
// criteria. A RunList may be compressed by grouping "runs" of entries
// which are equal (according to the sort criteria) into a new RunList of
// runs. For instance, a RunList containing pairs (x, y) may be compressed
// into a RunList containing pair runs (x, {y}) where each run consists of
// a list of y's with the same x.
type RunList []interface{}
func (h RunList) sort(less comparer) {
sort.Sort(&interfaceSlice{h, less})
}
func (p *interfaceSlice) Len() int { return len(p.slice) }
func (p *interfaceSlice) Less(i, j int) bool { return p.less(p.slice[i], p.slice[j]) }
func (p *interfaceSlice) Swap(i, j int) { p.slice[i], p.slice[j] = p.slice[j], p.slice[i] }
// Compress entries which are the same according to a sort criteria
// (specified by less) into "runs".
func (h RunList) reduce(less comparer, newRun func(h RunList) interface{}) RunList {
if len(h) == 0 {
return nil
}
// len(h) > 0
// create runs of entries with equal values
h.sort(less)
// for each run, make a new run object and collect them in a new RunList
var hh RunList
i, x := 0, h[0]
for j, y := range h {
if less(x, y) {
hh = append(hh, newRun(h[i:j]))
i, x = j, h[j] // start a new run
}
}
// add final run, if any
if i < len(h) {
hh = append(hh, newRun(h[i:]))
}
return hh
}
// ----------------------------------------------------------------------------
// KindRun
// Debugging support. Disable to see multiple entries per line.
const removeDuplicates = true
// A KindRun is a run of SpotInfos of the same kind in a given file.
// The kind (3 bits) is stored in each SpotInfo element; to find the
// kind of a KindRun, look at any of its elements.
type KindRun []SpotInfo
// KindRuns are sorted by line number or index. Since the isIndex bit
// is always the same for all infos in one list we can compare lori's.
func (k KindRun) Len() int { return len(k) }
func (k KindRun) Less(i, j int) bool { return k[i].Lori() < k[j].Lori() }
func (k KindRun) Swap(i, j int) { k[i], k[j] = k[j], k[i] }
// FileRun contents are sorted by Kind for the reduction into KindRuns.
func lessKind(x, y interface{}) bool { return x.(SpotInfo).Kind() < y.(SpotInfo).Kind() }
// newKindRun allocates a new KindRun from the SpotInfo run h.
func newKindRun(h RunList) interface{} {
run := make(KindRun, len(h))
for i, x := range h {
run[i] = x.(SpotInfo)
}
// Spots were sorted by file and kind to create this run.
// Within this run, sort them by line number or index.
sort.Sort(run)
if removeDuplicates {
// Since both the lori and kind field must be
// same for duplicates, and since the isIndex
// bit is always the same for all infos in one
// list we can simply compare the entire info.
k := 0
prev := SpotInfo(1<<32 - 1) // an unlikely value
for _, x := range run {
if x != prev {
run[k] = x
k++
prev = x
}
}
run = run[0:k]
}
return run
}
// ----------------------------------------------------------------------------
// FileRun
// A Pak describes a Go package.
type Pak struct {
Path string // path of directory containing the package
Name string // package name as declared by package clause
}
// Paks are sorted by name (primary key) and by import path (secondary key).
func (p *Pak) less(q *Pak) bool {
return p.Name < q.Name || p.Name == q.Name && p.Path < q.Path
}
// A File describes a Go file.
type File struct {
Name string // directory-local file name
Pak *Pak // the package to which the file belongs
}
// Path returns the file path of f.
func (f *File) Path() string {
return pathpkg.Join(f.Pak.Path, f.Name)
}
// A Spot describes a single occurrence of a word.
type Spot struct {
File *File
Info SpotInfo
}
// A FileRun is a list of KindRuns belonging to the same file.
type FileRun struct {
File *File
Groups []KindRun
}
// Spots are sorted by file path for the reduction into FileRuns.
func lessSpot(x, y interface{}) bool {
fx := x.(Spot).File
fy := y.(Spot).File
// same as "return fx.Path() < fy.Path()" but w/o computing the file path first
px := fx.Pak.Path
py := fy.Pak.Path
return px < py || px == py && fx.Name < fy.Name
}
// newFileRun allocates a new FileRun from the Spot run h.
func newFileRun(h RunList) interface{} {
file := h[0].(Spot).File
// reduce the list of Spots into a list of KindRuns
h1 := make(RunList, len(h))
for i, x := range h {
h1[i] = x.(Spot).Info
}
h2 := h1.reduce(lessKind, newKindRun)
// create the FileRun
groups := make([]KindRun, len(h2))
for i, x := range h2 {
groups[i] = x.(KindRun)
}
return &FileRun{file, groups}
}
// ----------------------------------------------------------------------------
// PakRun
// A PakRun describes a run of *FileRuns of a package.
type PakRun struct {
Pak *Pak
Files []*FileRun
}
// Sorting support for files within a PakRun.
func (p *PakRun) Len() int { return len(p.Files) }
func (p *PakRun) Less(i, j int) bool { return p.Files[i].File.Name < p.Files[j].File.Name }
func (p *PakRun) Swap(i, j int) { p.Files[i], p.Files[j] = p.Files[j], p.Files[i] }
// FileRuns are sorted by package for the reduction into PakRuns.
func lessFileRun(x, y interface{}) bool {
return x.(*FileRun).File.Pak.less(y.(*FileRun).File.Pak)
}
// newPakRun allocates a new PakRun from the *FileRun run h.
func newPakRun(h RunList) interface{} {
pak := h[0].(*FileRun).File.Pak
files := make([]*FileRun, len(h))
for i, x := range h {
files[i] = x.(*FileRun)
}
run := &PakRun{pak, files}
sort.Sort(run) // files were sorted by package; sort them by file now
return run
}
// ----------------------------------------------------------------------------
// HitList
// A HitList describes a list of PakRuns.
type HitList []*PakRun
// PakRuns are sorted by package.
func lessPakRun(x, y interface{}) bool { return x.(*PakRun).Pak.less(y.(*PakRun).Pak) }
func reduce(h0 RunList) HitList {
// reduce a list of Spots into a list of FileRuns
h1 := h0.reduce(lessSpot, newFileRun)
// reduce a list of FileRuns into a list of PakRuns
h2 := h1.reduce(lessFileRun, newPakRun)
// sort the list of PakRuns by package
h2.sort(lessPakRun)
// create a HitList
h := make(HitList, len(h2))
for i, p := range h2 {
h[i] = p.(*PakRun)
}
return h
}
// filter returns a new HitList created by filtering
// all PakRuns from h that have a matching pakname.
func (h HitList) filter(pakname string) HitList {
var hh HitList
for _, p := range h {
if p.Pak.Name == pakname {
hh = append(hh, p)
}
}
return hh
}
// ----------------------------------------------------------------------------
// AltWords
type wordPair struct {
canon string // canonical word spelling (all lowercase)
alt string // alternative spelling
}
// An AltWords describes a list of alternative spellings for a
// canonical (all lowercase) spelling of a word.
type AltWords struct {
Canon string // canonical word spelling (all lowercase)
Alts []string // alternative spelling for the same word
}
// wordPairs are sorted by their canonical spelling.
func lessWordPair(x, y interface{}) bool { return x.(*wordPair).canon < y.(*wordPair).canon }
// newAltWords allocates a new AltWords from the *wordPair run h.
func newAltWords(h RunList) interface{} {
canon := h[0].(*wordPair).canon
alts := make([]string, len(h))
for i, x := range h {
alts[i] = x.(*wordPair).alt
}
return &AltWords{canon, alts}
}
func (a *AltWords) filter(s string) *AltWords {
var alts []string
for _, w := range a.Alts {
if w != s {
alts = append(alts, w)
}
}
if len(alts) > 0 {
return &AltWords{a.Canon, alts}
}
return nil
}
// Ident stores information about external identifiers in order to create
// links to package documentation.
type Ident struct {
Path string // e.g. "net/http"
Package string // e.g. "http"
Name string // e.g. "NewRequest"
Doc string // e.g. "NewRequest returns a new Request..."
}
// byImportCount sorts the given slice of Idents by the import
// counts of the packages to which they belong.
type byImportCount struct {
Idents []Ident
ImportCount map[string]int
}
func (ic byImportCount) Len() int {
return len(ic.Idents)
}
func (ic byImportCount) Less(i, j int) bool {
ri := ic.ImportCount[ic.Idents[i].Path]
rj := ic.ImportCount[ic.Idents[j].Path]
if ri == rj {
return ic.Idents[i].Path < ic.Idents[j].Path
}
return ri > rj
}
func (ic byImportCount) Swap(i, j int) {
ic.Idents[i], ic.Idents[j] = ic.Idents[j], ic.Idents[i]
}
func (ic byImportCount) String() string {
buf := bytes.NewBuffer([]byte("["))
for _, v := range ic.Idents {
buf.WriteString(fmt.Sprintf("\n\t%s, %s (%d)", v.Path, v.Name, ic.ImportCount[v.Path]))
}
buf.WriteString("\n]")
return buf.String()
}
// filter creates a new Ident list where the results match the given
// package name.
func (ic byImportCount) filter(pakname string) []Ident {
if ic.Idents == nil {
return nil
}
var res []Ident
for _, i := range ic.Idents {
if i.Package == pakname {
res = append(res, i)
}
}
return res
}
// top returns the top n identifiers.
func (ic byImportCount) top(n int) []Ident {
if len(ic.Idents) > n {
return ic.Idents[:n]
}
return ic.Idents
}
// ----------------------------------------------------------------------------
// Indexer
type IndexResult struct {
Decls RunList // package-level declarations (with snippets)
Others RunList // all other occurrences
}
// Statistics provides statistics information for an index.
type Statistics struct {
Bytes int // total size of indexed source files
Files int // number of indexed source files
Lines int // number of lines (all files)
Words int // number of different identifiers
Spots int // number of identifier occurrences
}
// An Indexer maintains the data structures and provides the machinery
// for indexing .go files under a file tree. It implements the path.Visitor
// interface for walking file trees, and the ast.Visitor interface for
// walking Go ASTs.
type Indexer struct {
c *Corpus
fset *token.FileSet // file set for all indexed files
fsOpenGate chan bool // send pre fs.Open; receive on close
mu sync.Mutex // guards all the following
sources bytes.Buffer // concatenated sources
strings map[string]string // interned string
packages map[Pak]*Pak // interned *Paks
words map[string]*IndexResult // RunLists of Spots
snippets []*Snippet // indices are stored in SpotInfos
current *token.File // last file added to file set
file *File // AST for current file
decl ast.Decl // AST for current decl
stats Statistics
throttle *util.Throttle
importCount map[string]int // package path ("net/http") => count
packagePath map[string]map[string]bool // "template" => "text/template" => true
exports map[string]map[string]SpotKind // "net/http" => "ListenAndServe" => FuncDecl
curPkgExports map[string]SpotKind
idents map[SpotKind]map[string][]Ident // kind => name => list of Idents
}
func (x *Indexer) intern(s string) string {
if s, ok := x.strings[s]; ok {
return s
}
x.strings[s] = s
return s
}
func (x *Indexer) lookupPackage(path, name string) *Pak {
// In the source directory tree, more than one package may
// live in the same directory. For the packages map, construct
// a key that includes both the directory path and the package
// name.
key := Pak{Path: x.intern(path), Name: x.intern(name)}
pak := x.packages[key]
if pak == nil {
pak = &key
x.packages[key] = pak
}
return pak
}
func (x *Indexer) addSnippet(s *Snippet) int {
index := len(x.snippets)
x.snippets = append(x.snippets, s)
return index
}
func (x *Indexer) visitIdent(kind SpotKind, id *ast.Ident) {
if id == nil {
return
}
name := x.intern(id.Name)
switch kind {
case TypeDecl, FuncDecl, ConstDecl, VarDecl:
x.curPkgExports[name] = kind
}
lists, found := x.words[name]
if !found {
lists = new(IndexResult)
x.words[name] = lists
}
if kind == Use || x.decl == nil {
if x.c.IndexGoCode {
// not a declaration or no snippet required
info := makeSpotInfo(kind, x.current.Line(id.Pos()), false)
lists.Others = append(lists.Others, Spot{x.file, info})
}
} else {
// a declaration with snippet
index := x.addSnippet(NewSnippet(x.fset, x.decl, id))
info := makeSpotInfo(kind, index, true)
lists.Decls = append(lists.Decls, Spot{x.file, info})
}
x.stats.Spots++
}
func (x *Indexer) visitFieldList(kind SpotKind, flist *ast.FieldList) {
for _, f := range flist.List {
x.decl = nil // no snippets for fields
for _, name := range f.Names {
x.visitIdent(kind, name)
}
ast.Walk(x, f.Type)
// ignore tag - not indexed at the moment
}
}
func (x *Indexer) visitSpec(kind SpotKind, spec ast.Spec) {
switch n := spec.(type) {
case *ast.ImportSpec:
x.visitIdent(ImportDecl, n.Name)
if n.Path != nil {
if imp, err := strconv.Unquote(n.Path.Value); err == nil {
x.importCount[x.intern(imp)]++
}
}
case *ast.ValueSpec:
for _, n := range n.Names {
x.visitIdent(kind, n)
}
ast.Walk(x, n.Type)
for _, v := range n.Values {
ast.Walk(x, v)
}
case *ast.TypeSpec:
x.visitIdent(TypeDecl, n.Name)
ast.Walk(x, n.Type)
}
}
func (x *Indexer) visitGenDecl(decl *ast.GenDecl) {
kind := VarDecl
if decl.Tok == token.CONST {
kind = ConstDecl
}
x.decl = decl
for _, s := range decl.Specs {
x.visitSpec(kind, s)
}
}
func (x *Indexer) Visit(node ast.Node) ast.Visitor {
switch n := node.(type) {
case nil:
// nothing to do
case *ast.Ident:
x.visitIdent(Use, n)
case *ast.FieldList:
x.visitFieldList(VarDecl, n)
case *ast.InterfaceType:
x.visitFieldList(MethodDecl, n.Methods)
case *ast.DeclStmt:
// local declarations should only be *ast.GenDecls;
// ignore incorrect ASTs
if decl, ok := n.Decl.(*ast.GenDecl); ok {
x.decl = nil // no snippets for local declarations
x.visitGenDecl(decl)
}
case *ast.GenDecl:
x.decl = n
x.visitGenDecl(n)
case *ast.FuncDecl:
kind := FuncDecl
if n.Recv != nil {
kind = MethodDecl
ast.Walk(x, n.Recv)
}
x.decl = n
x.visitIdent(kind, n.Name)
ast.Walk(x, n.Type)
if n.Body != nil {
ast.Walk(x, n.Body)
}
case *ast.File:
x.decl = nil
x.visitIdent(PackageClause, n.Name)
for _, d := range n.Decls {
ast.Walk(x, d)
}
default:
return x
}
return nil
}
// addFile adds a file to the index if possible and returns the file set file
// and the file's AST if it was successfully parsed as a Go file. If addFile
// failed (that is, if the file was not added), it returns file == nil.
func (x *Indexer) addFile(f vfs.ReadSeekCloser, filename string, goFile bool) (file *token.File, ast *ast.File) {
defer f.Close()
// The file set's base offset and x.sources size must be in lock-step;
// this permits the direct mapping of suffix array lookup results to
// to corresponding Pos values.
//
// When a file is added to the file set, its offset base increases by
// the size of the file + 1; and the initial base offset is 1. Add an
// extra byte to the sources here.
x.sources.WriteByte(0)
// If the sources length doesn't match the file set base at this point
// the file set implementation changed or we have another error.
base := x.fset.Base()
if x.sources.Len() != base {
panic("internal error: file base incorrect")
}
// append file contents (src) to x.sources
if _, err := x.sources.ReadFrom(f); err == nil {
src := x.sources.Bytes()[base:]
if goFile {
// parse the file and in the process add it to the file set
if ast, err = parser.ParseFile(x.fset, filename, src, parser.ParseComments); err == nil {
file = x.fset.File(ast.Pos()) // ast.Pos() is inside the file
return
}
// file has parse errors, and the AST may be incorrect -
// set lines information explicitly and index as ordinary
// text file (cannot fall through to the text case below
// because the file has already been added to the file set
// by the parser)
file = x.fset.File(token.Pos(base)) // token.Pos(base) is inside the file
file.SetLinesForContent(src)
ast = nil
return
}
if util.IsText(src) {
// only add the file to the file set (for the full text index)
file = x.fset.AddFile(filename, x.fset.Base(), len(src))
file.SetLinesForContent(src)
return
}
}
// discard possibly added data
x.sources.Truncate(base - 1) // -1 to remove added byte 0 since no file was added
return
}
// Design note: Using an explicit white list of permitted files for indexing
// makes sure that the important files are included and massively reduces the
// number of files to index. The advantage over a blacklist is that unexpected
// (non-blacklisted) files won't suddenly explode the index.
// Files are whitelisted if they have a file name or extension
// present as key in whitelisted.
var whitelisted = map[string]bool{
".bash": true,
".c": true,
".cc": true,
".cpp": true,
".cxx": true,
".css": true,
".go": true,
".goc": true,
".h": true,
".hh": true,
".hpp": true,
".hxx": true,
".html": true,
".js": true,
".out": true,
".py": true,
".s": true,
".sh": true,
".txt": true,
".xml": true,
"AUTHORS": true,
"CONTRIBUTORS": true,
"LICENSE": true,
"Makefile": true,
"PATENTS": true,
"README": true,
}
// isWhitelisted returns true if a file is on the list
// of "permitted" files for indexing. The filename must
// be the directory-local name of the file.
func isWhitelisted(filename string) bool {
key := pathpkg.Ext(filename)
if key == "" {
// file has no extension - use entire filename
key = filename
}
return whitelisted[key]
}
func (x *Indexer) indexDocs(dirname string, filename string, astFile *ast.File) {
pkgName := x.intern(astFile.Name.Name)
if pkgName == "main" {
return
}
pkgPath := x.intern(strings.TrimPrefix(strings.TrimPrefix(dirname, "/src/"), "pkg/"))
astPkg := ast.Package{
Name: pkgName,
Files: map[string]*ast.File{
filename: astFile,
},
}
var m doc.Mode
docPkg := doc.New(&astPkg, dirname, m)
addIdent := func(sk SpotKind, name string, docstr string) {
if x.idents[sk] == nil {
x.idents[sk] = make(map[string][]Ident)
}
name = x.intern(name)
x.idents[sk][name] = append(x.idents[sk][name], Ident{
Path: pkgPath,
Package: pkgName,
Name: name,
Doc: doc.Synopsis(docstr),
})
}
if x.idents[PackageClause] == nil {
x.idents[PackageClause] = make(map[string][]Ident)
}
// List of words under which the package identifier will be stored.
// This includes the package name and the components of the directory
// in which it resides.
words := strings.Split(pathpkg.Dir(pkgPath), "/")
if words[0] == "." {
words = []string{}
}
name := x.intern(docPkg.Name)
synopsis := doc.Synopsis(docPkg.Doc)
words = append(words, name)
pkgIdent := Ident{
Path: pkgPath,
Package: pkgName,
Name: name,
Doc: synopsis,
}
for _, word := range words {
word = x.intern(word)
found := false
pkgs := x.idents[PackageClause][word]
for i, p := range pkgs {
if p.Path == pkgPath {
if docPkg.Doc != "" {
p.Doc = synopsis
pkgs[i] = p
}
found = true
break
}
}
if !found {
x.idents[PackageClause][word] = append(x.idents[PackageClause][word], pkgIdent)
}
}
for _, c := range docPkg.Consts {
for _, name := range c.Names {
addIdent(ConstDecl, name, c.Doc)
}
}
for _, t := range docPkg.Types {
addIdent(TypeDecl, t.Name, t.Doc)
for _, c := range t.Consts {
for _, name := range c.Names {
addIdent(ConstDecl, name, c.Doc)
}
}
for _, v := range t.Vars {
for _, name := range v.Names {
addIdent(VarDecl, name, v.Doc)
}
}
for _, f := range t.Funcs {
addIdent(FuncDecl, f.Name, f.Doc)
}
for _, f := range t.Methods {
addIdent(MethodDecl, f.Name, f.Doc)
// Change the name of methods to be "<typename>.<methodname>".
// They will still be indexed as <methodname>.
idents := x.idents[MethodDecl][f.Name]
idents[len(idents)-1].Name = x.intern(t.Name + "." + f.Name)
}
}
for _, v := range docPkg.Vars {
for _, name := range v.Names {
addIdent(VarDecl, name, v.Doc)
}
}
for _, f := range docPkg.Funcs {
addIdent(FuncDecl, f.Name, f.Doc)
}
}
func (x *Indexer) indexGoFile(dirname string, filename string, file *token.File, astFile *ast.File) {
pkgName := astFile.Name.Name
if x.c.IndexGoCode {
x.current = file
pak := x.lookupPackage(dirname, pkgName)
x.file = &File{filename, pak}
ast.Walk(x, astFile)
}
if x.c.IndexDocs {
// Test files are already filtered out in visitFile if IndexGoCode and
// IndexFullText are false. Otherwise, check here.
isTestFile := (x.c.IndexGoCode || x.c.IndexFullText) &&
(strings.HasSuffix(filename, "_test.go") || strings.HasPrefix(dirname, "/test/"))
if !isTestFile {
x.indexDocs(dirname, filename, astFile)
}
}
ppKey := x.intern(pkgName)
if _, ok := x.packagePath[ppKey]; !ok {
x.packagePath[ppKey] = make(map[string]bool)
}
pkgPath := x.intern(strings.TrimPrefix(strings.TrimPrefix(dirname, "/src/"), "pkg/"))
x.packagePath[ppKey][pkgPath] = true
// Merge in exported symbols found walking this file into
// the map for that package.
if len(x.curPkgExports) > 0 {
dest, ok := x.exports[pkgPath]
if !ok {
dest = make(map[string]SpotKind)
x.exports[pkgPath] = dest
}
for k, v := range x.curPkgExports {
dest[k] = v
}
}
}
func (x *Indexer) visitFile(dirname string, fi os.FileInfo) {
if fi.IsDir() || !x.c.IndexEnabled {
return
}
filename := pathpkg.Join(dirname, fi.Name())
goFile := isGoFile(fi)
switch {
case x.c.IndexFullText:
if !isWhitelisted(fi.Name()) {
return
}
case x.c.IndexGoCode:
if !goFile {
return
}
case x.c.IndexDocs:
if !goFile ||
strings.HasSuffix(fi.Name(), "_test.go") ||
strings.HasPrefix(dirname, "/test/") {
return
}
default:
// No indexing turned on.
return
}
x.fsOpenGate <- true
defer func() { <-x.fsOpenGate }()
// open file
f, err := x.c.fs.Open(filename)
if err != nil {
return
}
x.mu.Lock()
defer x.mu.Unlock()
x.throttle.Throttle()
x.curPkgExports = make(map[string]SpotKind)
file, fast := x.addFile(f, filename, goFile)
if file == nil {
return // addFile failed
}
if fast != nil {
x.indexGoFile(dirname, fi.Name(), file, fast)
}
// update statistics
x.stats.Bytes += file.Size()
x.stats.Files++
x.stats.Lines += file.LineCount()
}
// indexOptions contains information that affects the contents of an index.
type indexOptions struct {
// Docs provides documentation search results.
// It is only consulted if IndexEnabled is true.
// The default values is true.
Docs bool
// GoCode provides Go source code search results.
// It is only consulted if IndexEnabled is true.
// The default values is true.
GoCode bool
// FullText provides search results from all files.
// It is only consulted if IndexEnabled is true.
// The default values is true.
FullText bool
// MaxResults optionally specifies the maximum results for indexing.
// The default is 1000.
MaxResults int
}
// ----------------------------------------------------------------------------
// Index
type LookupResult struct {
Decls HitList // package-level declarations (with snippets)
Others HitList // all other occurrences
}
type Index struct {
fset *token.FileSet // file set used during indexing; nil if no textindex
suffixes *suffixarray.Index // suffixes for concatenated sources; nil if no textindex
words map[string]*LookupResult // maps words to hit lists
alts map[string]*AltWords // maps canonical(words) to lists of alternative spellings
snippets []*Snippet // all snippets, indexed by snippet index
stats Statistics
importCount map[string]int // package path ("net/http") => count
packagePath map[string]map[string]bool // "template" => "text/template" => true
exports map[string]map[string]SpotKind // "net/http" => "ListenAndServe" => FuncDecl
idents map[SpotKind]map[string][]Ident
opts indexOptions
}
func canonical(w string) string { return strings.ToLower(w) }
// Somewhat arbitrary, but I figure low enough to not hurt disk-based filesystems
// consuming file descriptors, where some systems have low 256 or 512 limits.
// Go should have a built-in way to cap fd usage under the ulimit.
const (
maxOpenFiles = 200
maxOpenDirs = 50
)
func (c *Corpus) throttle() float64 {
if c.IndexThrottle <= 0 {
return 0.9
}
if c.IndexThrottle > 1.0 {
return 1.0
}
return c.IndexThrottle
}
// NewIndex creates a new index for the .go files provided by the corpus.
func (c *Corpus) NewIndex() *Index {
// initialize Indexer
// (use some reasonably sized maps to start)
x := &Indexer{
c: c,
fset: token.NewFileSet(),
fsOpenGate: make(chan bool, maxOpenFiles),
strings: make(map[string]string),