internal/zstd: configure window size for single segment frames

Set window size to frame content size when single segment flag is set.

For #62513

Change-Id: I2a60c33123aca4f6a631e6d625f4582ff31a63cb
GitHub-Last-Rev: 9bafe01
GitHub-Pull-Request: #63224
Reviewed-on: https://go-review.googlesource.com/c/go/+/531075
Auto-Submit: Ian Lance Taylor <[email protected]>
LUCI-TryBot-Result: Go LUCI <[email protected]>
Reviewed-by: Ian Lance Taylor <[email protected]>
Reviewed-by: Bryan Mills <[email protected]>

Author: AlexanderYastrebov

src/internal/zstd/testdata/1890a371.gettysburg.txt-100x.zst

@@ -0,0 +1,10 @@
+This directory holds files for testing zstd.NewReader.
+
+Each one is a Zstandard compressed file named as hash.arbitrary-name.zst,
+where hash is the first eight hexadecimal digits of the SHA256 hash
+of the expected uncompressed content:
+
+	zstd -d < 1890a371.gettysburg.txt-100x.zst | sha256sum | head -c 8
+	1890a371
+
+The test uses hash value to verify decompression result.

src/internal/zstd/testdata/README

@@ -235,10 +235,7 @@ retry:
 	// Figure out the maximum amount of data we need to retain
 	// for backreferences.
 	var windowSize int
-	if singleSegment {
-		// No window required, as all the data is in a single buffer.
-		windowSize = 0
-	} else {
+	if !singleSegment {
 		// Window descriptor. RFC 3.1.1.1.2.
 		windowDescriptor := r.scratch[0]
 		exponent := uint64(windowDescriptor >> 3)
@@ -252,11 +249,6 @@ retry:
 		if fuzzing && (windowLog > 31 || windowSize > 1<<27) {
 			return r.makeError(relativeOffset, "windowSize too large")
 		}
-
-		// RFC 8878 permits us to set an 8M max on window size.
-		if windowSize > 8<<20 {
-			windowSize = 8 << 20
-		}
 	}
 
 	// Frame_Content_Size. RFC 3.1.1.4.
@@ -278,6 +270,18 @@ retry:
 		panic("unreachable")
 	}
 
+	// RFC 3.1.1.1.2.
+	// When Single_Segment_Flag is set, Window_Descriptor is not present.
+	// In this case, Window_Size is Frame_Content_Size.
+	if singleSegment {
+		windowSize = int(r.remainingFrameSize)
+	}
+
+	// RFC 8878 3.1.1.1.1.2. permits us to set an 8M max on window size.
+	if windowSize > 8<<20 {
+		windowSize = 8 << 20
+	}
+
 	relativeOffset += headerSize
 
 	r.sawFrameHeader = true

src/internal/zstd/zstd.go

@@ -6,12 +6,14 @@ package zstd
 
 import (
 	"bytes"
+	"crypto/sha256"
 	"fmt"
 	"internal/race"
 	"internal/testenv"
 	"io"
 	"os"
 	"os/exec"
+	"path/filepath"
 	"strings"
 	"sync"
 	"testing"
@@ -237,6 +239,39 @@ func TestAlloc(t *testing.T) {
 	}
 }
 
+func TestFileSamples(t *testing.T) {
+	samples, err := os.ReadDir("testdata")
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	for _, sample := range samples {
+		name := sample.Name()
+		if !strings.HasSuffix(name, ".zst") {
+			continue
+		}
+
+		t.Run(name, func(t *testing.T) {
+			f, err := os.Open(filepath.Join("testdata", name))
+			if err != nil {
+				t.Fatal(err)
+			}
+
+			r := NewReader(f)
+			h := sha256.New()
+			if _, err := io.Copy(h, r); err != nil {
+				t.Fatal(err)
+			}
+			got := fmt.Sprintf("%x", h.Sum(nil))[:8]
+
+			want, _, _ := strings.Cut(name, ".")
+			if got != want {
+				t.Errorf("Wrong uncompressed content hash: got %s, want %s", got, want)
+			}
+		})
+	}
+}
+
 func BenchmarkLarge(b *testing.B) {
 	b.StopTimer()
 	b.ReportAllocs()