-
Notifications
You must be signed in to change notification settings - Fork 2k
Expand file tree
/
Copy pathextractor.go
More file actions
2058 lines (1877 loc) · 67.6 KB
/
Copy pathextractor.go
File metadata and controls
2058 lines (1877 loc) · 67.6 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
package extractor
import (
"crypto/md5"
"encoding/hex"
"fmt"
"go/ast"
"go/constant"
"go/scanner"
"go/token"
"go/types"
"io"
"log"
"os"
"path/filepath"
"reflect"
"regexp"
"runtime"
"strconv"
"strings"
"sync"
"time"
"github.com/github/codeql-go/extractor/dbscheme"
"github.com/github/codeql-go/extractor/diagnostics"
"github.com/github/codeql-go/extractor/srcarchive"
"github.com/github/codeql-go/extractor/toolchain"
"github.com/github/codeql-go/extractor/trap"
"github.com/github/codeql-go/extractor/util"
"golang.org/x/tools/go/packages"
)
var MaxGoRoutines int
var typeParamParent map[*types.TypeParam]types.Object = make(map[*types.TypeParam]types.Object)
func init() {
// this sets the number of threads that the Go runtime will spawn; this is separate
// from the number of goroutines that the program spawns, which are scheduled into
// the system threads by the Go runtime scheduler
threads := os.Getenv("LGTM_THREADS")
if maxprocs, err := strconv.Atoi(threads); err == nil && maxprocs > 0 {
log.Printf("Max threads set to %d", maxprocs)
runtime.GOMAXPROCS(maxprocs)
} else if threads != "" {
log.Printf("Warning: LGTM_THREADS value %s is not valid, defaulting to using all available threads.", threads)
}
// if the value is empty or not set, use the Go default, which is the number of cores
// available since Go 1.5, but is subject to change
var err error
if MaxGoRoutines, err = strconv.Atoi(util.Getenv(
"CODEQL_EXTRACTOR_GO_MAX_GOROUTINES",
"SEMMLE_MAX_GOROUTINES",
)); err != nil {
MaxGoRoutines = 32
} else {
log.Printf("Max goroutines set to %d", MaxGoRoutines)
}
}
// Extract extracts the packages specified by the given patterns
func Extract(patterns []string) error {
return ExtractWithFlags(nil, patterns, false)
}
// ExtractWithFlags extracts the packages specified by the given patterns and build flags
func ExtractWithFlags(buildFlags []string, patterns []string, extractTests bool) error {
startTime := time.Now()
extraction := NewExtraction(buildFlags, patterns)
defer extraction.StatWriter.Close()
modEnabled := os.Getenv("GO111MODULE") != "off"
if !modEnabled {
log.Println("Go module mode disabled.")
}
modFlags := make([]string, 0, 1)
for _, flag := range buildFlags {
if strings.HasPrefix(flag, "-mod=") {
modFlags = append(modFlags, flag)
}
}
// If CODEQL_EXTRACTOR_GO_[OPTION_]EXTRACT_VENDOR_DIRS is "true", we extract `vendor` directories;
// otherwise (the default) is to exclude them from extraction
includeVendor, oldOptionUsed := util.IsVendorDirExtractionEnabled()
if oldOptionUsed {
log.Println("Warning: obsolete option \"CODEQL_EXTRACTOR_GO_EXTRACT_VENDOR_DIRS\" was set. Use \"CODEQL_EXTRACTOR_GO_OPTION_EXTRACT_VENDOR_DIRS\" or pass `--extractor-option extract_vendor_dirs=true` instead.")
}
modeNotifications := make([]string, 0, 2)
if extractTests {
modeNotifications = append(modeNotifications, "test extraction enabled")
}
if includeVendor {
modeNotifications = append(modeNotifications, "extracting vendor directories")
}
modeMessage := strings.Join(modeNotifications, ", ")
if modeMessage != "" {
modeMessage = " (" + modeMessage + ")"
}
log.Printf("Running packages.Load%s.", modeMessage)
// This includes test packages if either we're tracing a `go test` command,
// or if CODEQL_EXTRACTOR_GO_OPTION_EXTRACT_TESTS is set to "true".
cfg := &packages.Config{
Mode: packages.NeedName | packages.NeedFiles |
packages.NeedCompiledGoFiles |
packages.NeedImports | packages.NeedDeps |
packages.NeedTypes | packages.NeedTypesSizes |
packages.NeedTypesInfo | packages.NeedSyntax,
BuildFlags: buildFlags,
Tests: extractTests,
}
pkgs, err := packages.Load(cfg, patterns...)
if err != nil {
// the toolchain directive is only supported in Go 1.21 and above
if strings.Contains(err.Error(), "unknown directive: toolchain") {
diagnostics.EmitNewerSystemGoRequired("1.21.0")
}
return err
}
log.Println("Done running packages.Load.")
if len(pkgs) == 0 {
log.Println("No packages found.")
wd, err := os.Getwd()
if err != nil {
log.Printf("Warning: failed to get working directory: %s\n", err.Error())
} else if util.FindGoFiles(wd) {
diagnostics.EmitGoFilesFoundButNotProcessed()
}
}
log.Println("Extracting universe scope.")
extractUniverseScope()
log.Println("Done extracting universe scope.")
// a map of package path to source directory and module root directory
pkgInfos := make(map[string]toolchain.PkgInfo)
// root directories of packages that we want to extract
wantedRoots := make(map[string]bool)
if os.Getenv("CODEQL_EXTRACTOR_GO_FAST_PACKAGE_INFO") != "false" {
log.Printf("Running go list to resolve package and module directories.")
// get all packages information
pkgInfos, err = toolchain.GetPkgsInfo(patterns, true, extractTests, modFlags...)
if err != nil {
log.Fatalf("Error getting dependency package or module directories: %v.", err)
}
log.Printf("Done running go list deps: resolved %d packages.", len(pkgInfos))
}
pkgsNotFound := make([]string, 0, len(pkgs))
// Build a map from package paths to their longest IDs--
// in the context of a `go test -c` compilation, we will see the same package more than
// once, with IDs like "abc.com/pkgname [abc.com/pkgname.test]" to distinguish the version
// that contains and is used by test code.
// For our purposes it is simplest to just ignore the non-test version, since the test
// version seems to be a superset of it.
longestPackageIds := make(map[string]string)
packages.Visit(pkgs, nil, func(pkg *packages.Package) {
if longestIDSoFar, present := longestPackageIds[pkg.PkgPath]; present {
if len(pkg.ID) > len(longestIDSoFar) {
longestPackageIds[pkg.PkgPath] = pkg.ID
}
} else {
longestPackageIds[pkg.PkgPath] = pkg.ID
}
})
// Do a post-order traversal and extract the package scope of each package
packages.Visit(pkgs, nil, func(pkg *packages.Package) {
// Note that if test extraction is enabled, we will encounter a package twice here:
// once as the main package, and once as the test package (with a package ID like
// "abc.com/pkgname [abc.com/pkgname.test]").
//
// We will extract it both times however, because we need to visit the packages
// in the right order in order to visit used types before their users, and the
// ordering determined by packages.Visit for the main and the test package may differ.
//
// This should only cause some wasted time and not inconsistency because the names for
// objects seen in this process should be the same each time.
log.Printf("Processing package %s.", pkg.PkgPath)
if _, ok := pkgInfos[pkg.PkgPath]; !ok {
pkgInfos[pkg.PkgPath] = toolchain.GetPkgInfo(pkg.PkgPath, modFlags...)
}
log.Printf("Extracting types for package %s.", pkg.PkgPath)
tw, err := trap.NewWriter(pkg.PkgPath, pkg)
if err != nil {
log.Fatal(err)
}
defer tw.Close()
scope := extractPackageScope(tw, pkg)
extractObjectTypes(tw)
lbl := tw.Labeler.GlobalID(util.EscapeTrapSpecialChars(pkg.PkgPath) + ";pkg")
dbscheme.PackagesTable.Emit(tw, lbl, pkg.Name, pkg.PkgPath, scope)
if len(pkg.Errors) != 0 {
log.Printf("Warning: encountered errors extracting package `%s`:", pkg.PkgPath)
for i, err := range pkg.Errors {
errString := err.Error()
log.Printf(" %s", errString)
if strings.Contains(errString, "build constraints exclude all Go files in ") {
// `err` is a NoGoError from the package cmd/go/internal/load, which we cannot access as it is internal
diagnostics.EmitPackageDifferentOSArchitecture(pkg.PkgPath)
} else if strings.Contains(errString, "cannot find package") ||
strings.Contains(errString, "no required module provides package") {
pkgsNotFound = append(pkgsNotFound, pkg.PkgPath)
}
extraction.extractError(tw, err, lbl, i)
}
}
log.Printf("Done extracting types for package %s.", pkg.PkgPath)
})
if len(pkgsNotFound) > 0 {
diagnostics.EmitCannotFindPackages(pkgsNotFound)
}
for _, pkg := range pkgs {
pkgInfo, ok := pkgInfos[pkg.PkgPath]
if !ok || pkgInfo.PkgDir == "" {
log.Fatalf("Unable to get a source directory for input package %s.", pkg.PkgPath)
}
wantedRoots[pkgInfo.PkgDir] = true
if pkgInfo.ModDir != "" {
wantedRoots[pkgInfo.ModDir] = true
}
}
log.Println("Done processing dependencies.")
log.Println("Starting to extract packages.")
sep := regexp.QuoteMeta(string(filepath.Separator))
// Construct a list of directory segments to exclude from extraction, starting with ".."
excludedDirs := []string{`\.\.`}
if !includeVendor {
excludedDirs = append(excludedDirs, "vendor")
}
// If a path matches this regexp, we don't extract this package. It checks whether the path
// contains one of the `excludedDirs`.
noExtractRe := regexp.MustCompile(`.*(^|` + sep + `)(` + strings.Join(excludedDirs, "|") + `)($|` + sep + `).*`)
// extract AST information for all packages
packages.Visit(pkgs, nil, func(pkg *packages.Package) {
// If this is a variant of a package that also occurs with a longer ID, skip it;
// otherwise we would extract the same file more than once including extracting the
// body of methods twice, causing database inconsistencies.
//
// We prefer the version with the longest ID because that is (so far as I know) always
// the version that defines more entities -- the only case I'm aware of being a test
// variant of a package, which includes test-only functions in addition to the complete
// contents of the main variant.
if pkg.ID != longestPackageIds[pkg.PkgPath] {
return
}
for root := range wantedRoots {
pkgInfo := pkgInfos[pkg.PkgPath]
relDir, err := filepath.Rel(root, pkgInfo.PkgDir)
if err != nil || noExtractRe.MatchString(relDir) {
// if the path can't be made relative or matches the noExtract regexp skip it
continue
}
extraction.extractPackage(pkg)
modDir := pkgInfo.ModDir
if modDir == "" {
modDir = pkgInfo.PkgDir
}
if modDir != "" {
modPath := filepath.Join(modDir, "go.mod")
if util.FileExists(modPath) {
log.Printf("Extracting %s", modPath)
start := time.Now()
err := extraction.extractGoMod(modPath)
if err != nil {
log.Printf("Failed to extract go.mod: %s", err.Error())
}
end := time.Since(start)
log.Printf("Done extracting %s (%dms)", modPath, end.Nanoseconds()/1000000)
}
}
return
}
log.Printf("Skipping dependency package %s.", pkg.PkgPath)
})
extraction.WaitGroup.Wait()
log.Println("Done extracting packages.")
t := time.Now()
elapsed := t.Sub(startTime)
dbscheme.CompilationFinishedTable.Emit(extraction.StatWriter, extraction.Label, 0.0, elapsed.Seconds())
return nil
}
type Extraction struct {
// A lock for preventing concurrent writes to maps and the stat trap writer, as they are not
// thread-safe
Lock sync.Mutex
LabelKey string
Label trap.Label
StatWriter *trap.Writer
WaitGroup sync.WaitGroup
GoroutineSem *semaphore
FdSem *semaphore
NextFileId int
FileInfo map[string]*FileInfo
SeenGoMods map[string]bool
}
type FileInfo struct {
Idx int
NextErr int
}
func (extraction *Extraction) SeenFile(path string) bool {
_, ok := extraction.FileInfo[path]
return ok
}
func (extraction *Extraction) GetFileInfo(path string) *FileInfo {
if fileInfo, ok := extraction.FileInfo[path]; ok {
return fileInfo
}
extraction.FileInfo[path] = &FileInfo{extraction.NextFileId, 0}
extraction.NextFileId += 1
return extraction.FileInfo[path]
}
func (extraction *Extraction) GetFileIdx(path string) int {
return extraction.GetFileInfo(path).Idx
}
func (extraction *Extraction) GetNextErr(path string) int {
finfo := extraction.GetFileInfo(path)
res := finfo.NextErr
finfo.NextErr += 1
return res
}
func NewExtraction(buildFlags []string, patterns []string) *Extraction {
hash := md5.New()
io.WriteString(hash, "go")
for _, buildFlag := range buildFlags {
io.WriteString(hash, " "+buildFlag)
}
io.WriteString(hash, " --")
for _, pattern := range patterns {
io.WriteString(hash, " "+pattern)
}
sum := hash.Sum(nil)
i := 0
var path string
// split compilation files into directories to avoid filling a single directory with too many files
pathFmt := fmt.Sprintf("compilations/%s/%s_%%d", hex.EncodeToString(sum[:1]), hex.EncodeToString(sum[1:]))
for {
path = fmt.Sprintf(pathFmt, i)
file, err := trap.FileFor(path)
if err != nil {
log.Fatalf("Error creating trap file: %s\n", err.Error())
}
i++
if !util.FileExists(file) {
break
}
}
statWriter, err := trap.NewWriter(path, nil)
if err != nil {
log.Fatal(err)
}
lblKey := fmt.Sprintf("%s_%d;compilation", hex.EncodeToString(sum), i)
lbl := statWriter.Labeler.GlobalID(lblKey)
wd, err := os.Getwd()
if err != nil {
log.Fatalf("Unable to determine current directory: %s\n", err.Error())
}
dbscheme.CompilationsTable.Emit(statWriter, lbl, wd)
i = 0
extractorPath, err := util.GetExtractorPath()
if err != nil {
log.Fatalf("Unable to get extractor path: %s\n", err.Error())
}
dbscheme.CompilationArgsTable.Emit(statWriter, lbl, 0, extractorPath)
i++
for _, flag := range buildFlags {
dbscheme.CompilationArgsTable.Emit(statWriter, lbl, i, flag)
i++
}
// emit a fake "--" argument to make it clear that what comes after it are patterns
dbscheme.CompilationArgsTable.Emit(statWriter, lbl, i, "--")
i++
for _, pattern := range patterns {
dbscheme.CompilationArgsTable.Emit(statWriter, lbl, i, pattern)
i++
}
return &Extraction{
LabelKey: lblKey,
Label: lbl,
StatWriter: statWriter,
// this semaphore is used to limit the number of files that are open at once;
// this is to prevent the extractor from running into issues with caps on the
// number of open files that can be held by one process
FdSem: newSemaphore(100),
// this semaphore is used to limit the number of goroutines spawned, so we
// don't run into memory issues
GoroutineSem: newSemaphore(MaxGoRoutines),
NextFileId: 0,
FileInfo: make(map[string]*FileInfo),
SeenGoMods: make(map[string]bool),
}
}
// extractUniverseScope extracts symbol table information for the universe scope
func extractUniverseScope() {
tw, err := trap.NewWriter("universe", nil)
if err != nil {
log.Fatal(err)
}
defer tw.Close()
lbl := tw.Labeler.ScopeID(types.Universe, nil)
dbscheme.ScopesTable.Emit(tw, lbl, dbscheme.UniverseScopeType.Index())
extractObjects(tw, types.Universe, lbl)
// Always extract an empty interface type
extractType(tw, types.NewInterfaceType([]*types.Func{}, []types.Type{}))
}
// extractObjects extracts all objects declared in the given scope
// For more information on objects, see:
// https://github.com/golang/example/blob/master/gotypes/README.md#objects
func extractObjects(tw *trap.Writer, scope *types.Scope, scopeLabel trap.Label) {
for _, name := range scope.Names() {
obj := scope.Lookup(name)
lbl, exists := tw.Labeler.ScopedObjectID(obj, func() trap.Label { return extractType(tw, obj.Type()) })
if !exists {
// Populate type parameter parents for functions. Note that methods
// do not appear as objects in any scope, so they have to be dealt
// with separately in extractMethods.
if funcObj, ok := obj.(*types.Func); ok {
populateTypeParamParents(funcObj.Type().(*types.Signature).TypeParams(), obj)
populateTypeParamParents(funcObj.Type().(*types.Signature).RecvTypeParams(), obj)
}
// Populate type parameter parents for defined types and alias types.
if typeNameObj, ok := obj.(*types.TypeName); ok {
// `types.TypeName` represents a type with a name: a defined
// type, an alias type, a type parameter, or a predeclared
// type such as `int` or `error`. We can distinguish these
// using `typeNameObj.Type()`, except that we need to be
// careful with alias types because before Go 1.24 they would
// return the underlying type.
if tp, ok := typeNameObj.Type().(*types.Named); ok && !typeNameObj.IsAlias() {
populateTypeParamParents(tp.TypeParams(), obj)
} else if tp, ok := typeNameObj.Type().(*types.Alias); ok {
populateTypeParamParents(tp.TypeParams(), obj)
}
}
extractObject(tw, obj, lbl)
}
if obj.Parent() != scope {
// this can happen if a scope is embedded into another with a `.` import.
continue
}
dbscheme.ObjectScopesTable.Emit(tw, lbl, scopeLabel)
}
}
// extractMethod extracts a method `meth` and emits it to the objects table, then returns its label
func extractMethod(tw *trap.Writer, meth *types.Func) trap.Label {
// get the receiver type of the method
recvtyp := meth.Type().(*types.Signature).Recv().Type()
// ensure receiver type has been extracted
recvtyplbl := extractType(tw, recvtyp)
// if the method label does not exist, extract it
methlbl, exists := tw.Labeler.MethodID(meth, recvtyplbl)
if !exists {
// Populate type parameter parents for methods. They do not appear as
// objects in any scope, so they have to be dealt with separately here.
populateTypeParamParents(meth.Type().(*types.Signature).TypeParams(), meth)
populateTypeParamParents(meth.Type().(*types.Signature).RecvTypeParams(), meth)
extractObject(tw, meth, methlbl)
}
return methlbl
}
// extractObject extracts a single object and emits it to the objects table.
// For more information on objects, see:
// https://github.com/golang/example/blob/master/gotypes/README.md#objects
func extractObject(tw *trap.Writer, obj types.Object, lbl trap.Label) {
checkObjectNotSpecialized(obj)
name := obj.Name()
isBuiltin := obj.Parent() == types.Universe
var kind int
switch obj.(type) {
case *types.PkgName:
kind = dbscheme.PkgObjectType.Index()
case *types.TypeName:
if isBuiltin {
kind = dbscheme.BuiltinTypeObjectType.Index()
} else {
kind = dbscheme.DeclTypeObjectType.Index()
}
case *types.Const:
if isBuiltin {
kind = dbscheme.BuiltinConstObjectType.Index()
} else {
kind = dbscheme.DeclConstObjectType.Index()
}
case *types.Nil:
kind = dbscheme.BuiltinConstObjectType.Index()
case *types.Var:
kind = dbscheme.DeclVarObjectType.Index()
case *types.Builtin:
kind = dbscheme.BuiltinFuncObjectType.Index()
case *types.Func:
kind = dbscheme.DeclFuncObjectType.Index()
case *types.Label:
kind = dbscheme.LabelObjectType.Index()
default:
log.Fatalf("unknown object of type %T", obj)
}
dbscheme.ObjectsTable.Emit(tw, lbl, kind, name)
// for methods, additionally extract information about the receiver
if sig, ok := obj.Type().(*types.Signature); ok {
if recv := sig.Recv(); recv != nil {
recvlbl, exists := tw.Labeler.ReceiverObjectID(recv, lbl)
if !exists {
extractObject(tw, recv, recvlbl)
}
dbscheme.MethodReceiversTable.Emit(tw, lbl, recvlbl)
}
}
}
// extractObjectTypes extracts type and receiver information for all objects
// For more information on objects, see:
// https://github.com/golang/example/blob/master/gotypes/README.md#objects
func extractObjectTypes(tw *trap.Writer) {
// calling `extractType` on a defined type will extract all methods defined
// on it, which will add new objects. Therefore we need to do this first
// before we loop over all objects and emit them.
changed := true
for changed {
changed = tw.ForEachObject(extractObjectType)
}
changed = tw.ForEachObject(emitObjectType)
if changed {
log.Printf("Warning: more objects were labeled while emitting object types")
}
}
// extractObjectType extracts type and receiver information for a given object
// For more information on objects, see:
// https://github.com/golang/example/blob/master/gotypes/README.md#objects
func extractObjectType(tw *trap.Writer, obj types.Object, lbl trap.Label) {
if tp := obj.Type(); tp != nil {
extractType(tw, tp)
}
}
// emitObjectType emits the type information for a given object
func emitObjectType(tw *trap.Writer, obj types.Object, lbl trap.Label) {
if tp := obj.Type(); tp != nil {
dbscheme.ObjectTypesTable.Emit(tw, lbl, extractType(tw, tp))
}
}
var (
// file:line:col
threePartPos = regexp.MustCompile(`^(.+):(\d+):(\d+)$`)
// file:line
twoPartPos = regexp.MustCompile(`^(.+):(\d+)$`)
)
// extractError extracts the message and location of a frontend error
func (extraction *Extraction) extractError(tw *trap.Writer, err packages.Error, pkglbl trap.Label, idx int) {
var (
lbl = tw.Labeler.FreshID()
tag = dbscheme.ErrorTags[err.Kind]
kind = dbscheme.ErrorTypes[err.Kind].Index()
pos = err.Pos
file = ""
line, col int
e error
)
if pos == "" || pos == "-" {
// extract a dummy file
wd, e := os.Getwd()
if e != nil {
wd = "."
log.Printf("Warning: failed to get working directory")
}
ewd, e := filepath.EvalSymlinks(wd)
if e != nil {
ewd = wd
log.Printf("Warning: failed to evaluate symlinks for %s", wd)
}
file = filepath.Join(ewd, "-")
extraction.extractFileInfo(tw, file, true)
} else {
var rawfile string
if parts := threePartPos.FindStringSubmatch(pos); parts != nil {
// "file:line:col"
col, e = strconv.Atoi(parts[3])
if e != nil {
log.Printf("Warning: malformed column number `%s`: %v", parts[3], e)
}
line, e = strconv.Atoi(parts[2])
if e != nil {
log.Printf("Warning: malformed line number `%s`: %v", parts[2], e)
}
rawfile = parts[1]
} else if parts := twoPartPos.FindStringSubmatch(pos); parts != nil {
// "file:line"
line, e = strconv.Atoi(parts[2])
if e != nil {
log.Printf("Warning: malformed line number `%s`: %v", parts[2], e)
}
rawfile = parts[1]
} else if pos != "" && pos != "-" {
log.Printf("Warning: malformed error position `%s`", pos)
}
afile, e := filepath.Abs(rawfile)
if e != nil {
log.Printf("Warning: failed to get absolute path for for %s", file)
afile = file
}
file, e = filepath.EvalSymlinks(afile)
if e != nil {
log.Printf("Warning: failed to evaluate symlinks for %s", afile)
file = afile
}
extraction.extractFileInfo(tw, file, false)
}
extraction.Lock.Lock()
flbl := extraction.StatWriter.Labeler.FileLabelFor(file)
diagLbl := extraction.StatWriter.Labeler.FreshID()
dbscheme.DiagnosticsTable.Emit(
extraction.StatWriter, diagLbl, 1, tag, err.Msg, err.Msg,
emitLocation(extraction.StatWriter, flbl, line, col, line, col))
dbscheme.DiagnosticForTable.Emit(extraction.StatWriter, diagLbl, extraction.Label, extraction.GetFileIdx(file), extraction.GetNextErr(file))
extraction.Lock.Unlock()
transformed := filepath.ToSlash(srcarchive.TransformPath(file))
dbscheme.ErrorsTable.Emit(tw, lbl, kind, err.Msg, pos, transformed, line, col, pkglbl, idx)
}
// extractPackage extracts AST information for all files in the given package
func (extraction *Extraction) extractPackage(pkg *packages.Package) {
for _, astFile := range pkg.Syntax {
extraction.WaitGroup.Add(1)
extraction.GoroutineSem.acquire(1)
go func(astFile *ast.File) {
err := extraction.extractFile(astFile, pkg)
if err != nil {
log.Fatal(err)
}
extraction.GoroutineSem.release(1)
extraction.WaitGroup.Done()
}(astFile)
}
}
// normalizedPath computes the normalized path (with symlinks resolved) for the given file
func normalizedPath(ast *ast.File, fset *token.FileSet) string {
file := fset.File(ast.Package).Name()
path, err := filepath.EvalSymlinks(file)
if err != nil {
return file
}
return path
}
// extractFile extracts AST information for the given file
func (extraction *Extraction) extractFile(ast *ast.File, pkg *packages.Package) error {
fset := pkg.Fset
if ast.Package == token.NoPos {
log.Printf("Skipping extracting a file without a 'package' declaration")
return nil
}
path := normalizedPath(ast, fset)
extraction.FdSem.acquire(3)
log.Printf("Extracting %s", path)
start := time.Now()
defer extraction.FdSem.release(1)
tw, err := trap.NewWriter(path, pkg)
if err != nil {
extraction.FdSem.release(2)
return err
}
defer tw.Close()
err = srcarchive.Add(path)
extraction.FdSem.release(2)
if err != nil {
return err
}
extraction.extractFileInfo(tw, path, false)
extractScopes(tw, ast, pkg)
extractFileNode(tw, ast)
extractObjectTypes(tw)
extractNumLines(tw, path, ast)
end := time.Since(start)
log.Printf("Done extracting %s (%dms)", path, end.Nanoseconds()/1000000)
return nil
}
// extractFileInfo extracts file-system level information for the given file, populating
// the `files` and `containerparent` tables
func (extraction *Extraction) extractFileInfo(tw *trap.Writer, file string, isDummy bool) {
// We may visit the same file twice because `extractError` calls this function to describe files containing
// compilation errors. It is also called for user source files being extracted.
extraction.Lock.Lock()
if extraction.SeenFile(file) {
extraction.Lock.Unlock()
return
}
extraction.Lock.Unlock()
path := filepath.ToSlash(srcarchive.TransformPath(file))
components := strings.Split(path, "/")
parentPath := ""
var parentLbl trap.Label
for i, component := range components {
// displayPath is the same as rawPath except for root directories: if
// rawPath is "" then displayPath is "/"; if rawPath is "C:" then
// displayPath is "C:/".
var rawPath, displayPath string
if i == 0 {
rawPath = component
if component == "" || regexp.MustCompile(`^[A-Za-z]:$`).MatchString(component) {
// Handle linux root and Windows drive letters by appending "/"
displayPath = rawPath + "/"
} else {
displayPath = rawPath
}
} else {
rawPath = parentPath + "/" + component
displayPath = rawPath
}
if i == len(components)-1 {
lbl := tw.Labeler.FileLabelFor(file)
dbscheme.FilesTable.Emit(tw, lbl, displayPath)
dbscheme.ContainerParentTable.Emit(tw, parentLbl, lbl)
dbscheme.HasLocationTable.Emit(tw, lbl, emitLocation(tw, lbl, 0, 0, 0, 0))
extraction.Lock.Lock()
slbl := extraction.StatWriter.Labeler.FileLabelFor(file)
if !isDummy {
dbscheme.CompilationCompilingFilesTable.Emit(extraction.StatWriter, extraction.Label, extraction.GetFileIdx(file), slbl)
}
extraction.Lock.Unlock()
break
}
lbl := tw.Labeler.GlobalID(util.EscapeTrapSpecialChars(displayPath) + ";folder")
dbscheme.FoldersTable.Emit(tw, lbl, displayPath)
if i > 0 {
dbscheme.ContainerParentTable.Emit(tw, parentLbl, lbl)
}
parentPath = rawPath
parentLbl = lbl
}
}
// extractLocation emits a location entity for the given entity
func extractLocation(tw *trap.Writer, entity trap.Label, sl int, sc int, el int, ec int) {
filelbl := tw.Labeler.FileLabel()
dbscheme.HasLocationTable.Emit(tw, entity, emitLocation(tw, filelbl, sl, sc, el, ec))
}
// emitLocation emits a location entity
func emitLocation(tw *trap.Writer, filelbl trap.Label, sl int, sc int, el int, ec int) trap.Label {
locLbl := tw.Labeler.GlobalID(fmt.Sprintf("loc,{%s},%d,%d,%d,%d", filelbl, sl, sc, el, ec))
dbscheme.LocationsDefaultTable.Emit(tw, locLbl, filelbl, sl, sc, el, ec)
return locLbl
}
// extractNodeLocation extracts location information for the given node
func extractNodeLocation(tw *trap.Writer, nd ast.Node, lbl trap.Label) {
if nd == nil {
return
}
fset := tw.Package.Fset
start, end := fset.Position(nd.Pos()), fset.Position(nd.End())
extractLocation(tw, lbl, start.Line, start.Column, end.Line, end.Column-1)
}
// extractPackageScope extracts symbol table information for the given package
func extractPackageScope(tw *trap.Writer, pkg *packages.Package) trap.Label {
pkgScope := pkg.Types.Scope()
pkgScopeLabel := tw.Labeler.ScopeID(pkgScope, pkg.Types)
dbscheme.ScopesTable.Emit(tw, pkgScopeLabel, dbscheme.PackageScopeType.Index())
dbscheme.ScopeNestingTable.Emit(tw, pkgScopeLabel, tw.Labeler.ScopeID(types.Universe, nil))
extractObjects(tw, pkgScope, pkgScopeLabel)
return pkgScopeLabel
}
// extractScopeLocation extracts location information for the given scope
func extractScopeLocation(tw *trap.Writer, scope *types.Scope, lbl trap.Label) {
fset := tw.Package.Fset
start, end := fset.Position(scope.Pos()), fset.Position(scope.End())
extractLocation(tw, lbl, start.Line, start.Column, end.Line, end.Column-1)
}
// extractScopes extracts symbol table information for the package scope and all local scopes
// of the given package. Note that this will not encounter methods or struct fields as
// they do not have a parent scope.
func extractScopes(tw *trap.Writer, nd *ast.File, pkg *packages.Package) {
pkgScopeLabel := extractPackageScope(tw, pkg)
fileScope := pkg.TypesInfo.Scopes[nd]
if fileScope != nil {
extractLocalScope(tw, fileScope, pkgScopeLabel)
}
}
// extractLocalScope extracts symbol table information for the given scope and all its nested scopes
func extractLocalScope(tw *trap.Writer, scope *types.Scope, parentScopeLabel trap.Label) {
scopeLabel := tw.Labeler.ScopeID(scope, nil)
dbscheme.ScopesTable.Emit(tw, scopeLabel, dbscheme.LocalScopeType.Index())
extractScopeLocation(tw, scope, scopeLabel)
dbscheme.ScopeNestingTable.Emit(tw, scopeLabel, parentScopeLabel)
for i := 0; i < scope.NumChildren(); i++ {
childScope := scope.Child(i)
extractLocalScope(tw, childScope, scopeLabel)
}
extractObjects(tw, scope, scopeLabel)
}
// extractFileNode extracts AST information for the given file and all nodes contained in it
func extractFileNode(tw *trap.Writer, nd *ast.File) {
lbl := tw.Labeler.FileLabel()
extractExpr(tw, nd.Name, lbl, 0, false)
for i, decl := range nd.Decls {
extractDecl(tw, decl, lbl, i)
}
for i, cg := range nd.Comments {
extractCommentGroup(tw, cg, lbl, i)
}
extractDoc(tw, nd.Doc, lbl)
emitScopeNodeInfo(tw, nd, lbl)
}
// extractDoc extracts information about a doc comment group associated with a given element
func extractDoc(tw *trap.Writer, doc *ast.CommentGroup, elt trap.Label) {
if doc != nil {
dbscheme.DocCommentsTable.Emit(tw, elt, tw.Labeler.LocalID(doc))
}
}
// extractCommentGroup extracts information about a doc comment group
func extractCommentGroup(tw *trap.Writer, cg *ast.CommentGroup, parent trap.Label, idx int) {
lbl := tw.Labeler.LocalID(cg)
dbscheme.CommentGroupsTable.Emit(tw, lbl, parent, idx)
extractNodeLocation(tw, cg, lbl)
for i, c := range cg.List {
extractComment(tw, c, lbl, i)
}
}
// extractComment extracts information about a given comment
func extractComment(tw *trap.Writer, c *ast.Comment, parent trap.Label, idx int) {
lbl := tw.Labeler.LocalID(c)
rawText := c.Text
var kind int
var text string
if rawText[:2] == "//" {
kind = dbscheme.SlashSlashComment.Index()
text = rawText[2:]
} else {
kind = dbscheme.SlashStarComment.Index()
text = rawText[2 : len(rawText)-2]
}
dbscheme.CommentsTable.Emit(tw, lbl, kind, parent, idx, text)
extractNodeLocation(tw, c, lbl)
}
// emitScopeNodeInfo associates an AST node with its induced scope, if any
func emitScopeNodeInfo(tw *trap.Writer, nd ast.Node, lbl trap.Label) {
scope, exists := tw.Package.TypesInfo.Scopes[nd]
if exists {
dbscheme.ScopeNodesTable.Emit(tw, lbl, tw.Labeler.ScopeID(scope, tw.Package.Types))
}
}
// extractExpr extracts AST information for the given expression and all its subexpressions
func extractExpr(tw *trap.Writer, expr ast.Expr, parent trap.Label, idx int, skipExtractingValue bool) {
if expr == nil || reflect.ValueOf(expr).IsNil() {
return
}
lbl := tw.Labeler.LocalID(expr)
extractTypeOf(tw, expr, lbl)
var kind int
switch expr := expr.(type) {
case *ast.BadExpr:
kind = dbscheme.BadExpr.Index()
case *ast.Ident:
kind = dbscheme.IdentExpr.Index()
dbscheme.LiteralsTable.Emit(tw, lbl, expr.Name, expr.Name)
def := tw.Package.TypesInfo.Defs[expr]
// Note that there are some cases where `expr` is in the map but `def`
// is nil. The docs for `tw.Package.TypesInfo.Defs` give the following
// examples: the package name in package clauses, or symbolic variables
// `t` in `t := x.(type)` of type switch headers.
if def != nil {
defTyp := extractType(tw, def.Type())
objlbl, exists := tw.Labeler.LookupObjectID(def, defTyp)
if objlbl == trap.InvalidLabel {
log.Printf("Omitting def binding to unknown object %v", def)
} else {
if !exists {
extractObject(tw, def, objlbl)
}
dbscheme.DefsTable.Emit(tw, lbl, objlbl)
}
}
use := getObjectBeingUsed(tw, expr)
if use != nil {
useTyp := extractType(tw, use.Type())
objlbl, exists := tw.Labeler.LookupObjectID(use, useTyp)
if objlbl == trap.InvalidLabel {
log.Printf("Omitting use binding to unknown object %v", use)
} else {
if !exists {
extractObject(tw, use, objlbl)
}
dbscheme.UsesTable.Emit(tw, lbl, objlbl)
}
}
case *ast.Ellipsis:
kind = dbscheme.EllipsisExpr.Index()
extractExpr(tw, expr.Elt, lbl, 0, false)
case *ast.BasicLit:
value := ""
switch expr.Kind {
case token.INT:
ival, _ := strconv.ParseInt(expr.Value, 0, 64)
value = strconv.FormatInt(ival, 10)
kind = dbscheme.IntLitExpr.Index()
case token.FLOAT:
value = expr.Value