Vendor Update Go Libs (#13166)

* update github.com/alecthomas/chroma v0.8.0 -> v0.8.1

* github.com/blevesearch/bleve v1.0.10 -> v1.0.12

* editorconfig-core-go v2.1.1 -> v2.3.7

* github.com/gliderlabs/ssh v0.2.2 -> v0.3.1

* migrate editorconfig.ParseBytes to Parse

* github.com/shurcooL/vfsgen to 0d455de96546

* github.com/go-git/go-git/v5 v5.1.0 -> v5.2.0

* github.com/google/uuid v1.1.1 -> v1.1.2

* github.com/huandu/xstrings v1.3.0 -> v1.3.2

* github.com/klauspost/compress v1.10.11 -> v1.11.1

* github.com/markbates/goth v1.61.2 -> v1.65.0

* github.com/mattn/go-sqlite3 v1.14.0 -> v1.14.4

* github.com/mholt/archiver v3.3.0 -> v3.3.2

* github.com/microcosm-cc/bluemonday 4f7140c49acb -> v1.0.4

* github.com/minio/minio-go v7.0.4 -> v7.0.5

* github.com/olivere/elastic v7.0.9 -> v7.0.20

* github.com/urfave/cli v1.20.0 -> v1.22.4

* github.com/prometheus/client_golang v1.1.0 -> v1.8.0

* github.com/xanzy/go-gitlab v0.37.0 -> v0.38.1

* mvdan.cc/xurls v2.1.0 -> v2.2.0

Co-authored-by: Lauris BH <lauris@nix.lv>

vendor/github.com/klauspost/compress/flate/fast_encoder.go

@@ -127,7 +127,7 @@ func (e *fastGen) addBlock(src []byte) int32 {
 // hash4 returns the hash of u to fit in a hash table with h bits.
 // Preferably h should be a constant and should always be <32.
 func hash4u(u uint32, h uint8) uint32 {
-	return (u * prime4bytes) >> ((32 - h) & 31)
+	return (u * prime4bytes) >> ((32 - h) & reg8SizeMask32)
 }
 
 type tableEntryPrev struct {
@@ -138,25 +138,25 @@ type tableEntryPrev struct {
 // hash4x64 returns the hash of the lowest 4 bytes of u to fit in a hash table with h bits.
 // Preferably h should be a constant and should always be <32.
 func hash4x64(u uint64, h uint8) uint32 {
-	return (uint32(u) * prime4bytes) >> ((32 - h) & 31)
+	return (uint32(u) * prime4bytes) >> ((32 - h) & reg8SizeMask32)
 }
 
 // hash7 returns the hash of the lowest 7 bytes of u to fit in a hash table with h bits.
 // Preferably h should be a constant and should always be <64.
 func hash7(u uint64, h uint8) uint32 {
-	return uint32(((u << (64 - 56)) * prime7bytes) >> ((64 - h) & 63))
+	return uint32(((u << (64 - 56)) * prime7bytes) >> ((64 - h) & reg8SizeMask64))
 }
 
 // hash8 returns the hash of u to fit in a hash table with h bits.
 // Preferably h should be a constant and should always be <64.
 func hash8(u uint64, h uint8) uint32 {
-	return uint32((u * prime8bytes) >> ((64 - h) & 63))
+	return uint32((u * prime8bytes) >> ((64 - h) & reg8SizeMask64))
 }
 
 // hash6 returns the hash of the lowest 6 bytes of u to fit in a hash table with h bits.
 // Preferably h should be a constant and should always be <64.
 func hash6(u uint64, h uint8) uint32 {
-	return uint32(((u << (64 - 48)) * prime6bytes) >> ((64 - h) & 63))
+	return uint32(((u << (64 - 48)) * prime6bytes) >> ((64 - h) & reg8SizeMask64))
 }
 
 // matchlen will return the match length between offsets and t in src.

vendor/github.com/klauspost/compress/flate/gen_inflate.go

@@ -85,7 +85,7 @@ readLiteral:
 						return
 					}
 					f.roffset++
-					b |= uint32(c) << (nb & 31)
+					b |= uint32(c) << (nb & regSizeMaskUint32)
 					nb += 8
 				}
 				chunk := f.hl.chunks[b&(huffmanNumChunks-1)]
@@ -104,7 +104,7 @@ readLiteral:
 						f.err = CorruptInputError(f.roffset)
 						return
 					}
-					f.b = b >> (n & 31)
+					f.b = b >> (n & regSizeMaskUint32)
 					f.nb = nb - n
 					v = int(chunk >> huffmanValueShift)
 					break
@@ -167,15 +167,15 @@ readLiteral:
 					return
 				}
 			}
-			length += int(f.b & uint32(1<<n-1))
-			f.b >>= n
+			length += int(f.b & uint32(1<<(n&regSizeMaskUint32)-1))
+			f.b >>= n & regSizeMaskUint32
 			f.nb -= n
 		}
 
-		var dist int
+		var dist uint32
 		if f.hd == nil {
 			for f.nb < 5 {
-				if err = moreBits(); err != nil {
+				if err = f.moreBits(); err != nil {
 					if debugDecode {
 						fmt.Println("morebits f.nb<5:", err)
 					}
@@ -183,17 +183,19 @@ readLiteral:
 					return
 				}
 			}
-			dist = int(bits.Reverse8(uint8(f.b & 0x1F << 3)))
+			dist = uint32(bits.Reverse8(uint8(f.b & 0x1F << 3)))
 			f.b >>= 5
 			f.nb -= 5
 		} else {
-			if dist, err = f.huffSym(f.hd); err != nil {
+			sym, err := f.huffSym(f.hd)
+			if err != nil {
 				if debugDecode {
 					fmt.Println("huffsym:", err)
 				}
 				f.err = err
 				return
 			}
+			dist = uint32(sym)
 		}
 
 		switch {
@@ -202,9 +204,9 @@ readLiteral:
 		case dist < maxNumDist:
 			nb := uint(dist-2) >> 1
 			// have 1 bit in bottom of dist, need nb more.
-			extra := (dist & 1) << nb
+			extra := (dist & 1) << (nb & regSizeMaskUint32)
 			for f.nb < nb {
-				if err = moreBits(); err != nil {
+				if err = f.moreBits(); err != nil {
 					if debugDecode {
 						fmt.Println("morebits f.nb<nb:", err)
 					}
@@ -212,10 +214,10 @@ readLiteral:
 					return
 				}
 			}
-			extra |= int(f.b & uint32(1<<nb-1))
-			f.b >>= nb
+			extra |= f.b & uint32(1<<(nb&regSizeMaskUint32)-1)
+			f.b >>= nb & regSizeMaskUint32
 			f.nb -= nb
-			dist = 1<<(nb+1) + 1 + extra
+			dist = 1<<((nb+1)&regSizeMaskUint32) + 1 + extra
 		default:
 			if debugDecode {
 				fmt.Println("dist too big:", dist, maxNumDist)
@@ -225,7 +227,7 @@ readLiteral:
 		}
 
 		// No check on length; encoding can be prescient.
-		if dist > f.dict.histSize() {
+		if dist > uint32(f.dict.histSize()) {
 			if debugDecode {
 				fmt.Println("dist > f.dict.histSize():", dist, f.dict.histSize())
 			}
@@ -233,7 +235,7 @@ readLiteral:
 			return
 		}
 
-		f.copyLen, f.copyDist = length, dist
+		f.copyLen, f.copyDist = length, int(dist)
 		goto copyHistory
 	}
 

vendor/github.com/klauspost/compress/flate/huffman_bit_writer.go

@@ -206,7 +206,7 @@ func (w *huffmanBitWriter) write(b []byte) {
 }
 
 func (w *huffmanBitWriter) writeBits(b int32, nb uint16) {
-	w.bits |= uint64(b) << (w.nbits & 63)
+	w.bits |= uint64(b) << (w.nbits & reg16SizeMask64)
 	w.nbits += nb
 	if w.nbits >= 48 {
 		w.writeOutBits()
@@ -759,7 +759,7 @@ func (w *huffmanBitWriter) writeTokens(tokens []token, leCodes, oeCodes []hcode)
 		} else {
 			// inlined
 			c := lengths[lengthCode&31]
-			w.bits |= uint64(c.code) << (w.nbits & 63)
+			w.bits |= uint64(c.code) << (w.nbits & reg16SizeMask64)
 			w.nbits += c.len
 			if w.nbits >= 48 {
 				w.writeOutBits()
@@ -779,7 +779,7 @@ func (w *huffmanBitWriter) writeTokens(tokens []token, leCodes, oeCodes []hcode)
 		} else {
 			// inlined
 			c := offs[offsetCode&31]
-			w.bits |= uint64(c.code) << (w.nbits & 63)
+			w.bits |= uint64(c.code) << (w.nbits & reg16SizeMask64)
 			w.nbits += c.len
 			if w.nbits >= 48 {
 				w.writeOutBits()
@@ -878,7 +878,7 @@ func (w *huffmanBitWriter) writeBlockHuff(eof bool, input []byte, sync bool) {
 	for _, t := range input {
 		// Bitwriting inlined, ~30% speedup
 		c := encoding[t]
-		w.bits |= uint64(c.code) << ((w.nbits) & 63)
+		w.bits |= uint64(c.code) << ((w.nbits) & reg16SizeMask64)
 		w.nbits += c.len
 		if w.nbits >= 48 {
 			bits := w.bits

vendor/github.com/klauspost/compress/flate/inflate.go

@@ -522,8 +522,8 @@ func (f *decompressor) readHuffman() error {
 				return err
 			}
 		}
-		rep += int(f.b & uint32(1<<nb-1))
-		f.b >>= nb
+		rep += int(f.b & uint32(1<<(nb&regSizeMaskUint32)-1))
+		f.b >>= nb & regSizeMaskUint32
 		f.nb -= nb
 		if i+rep > n {
 			if debugDecode {
@@ -603,7 +603,7 @@ readLiteral:
 						return
 					}
 					f.roffset++
-					b |= uint32(c) << (nb & 31)
+					b |= uint32(c) << (nb & regSizeMaskUint32)
 					nb += 8
 				}
 				chunk := f.hl.chunks[b&(huffmanNumChunks-1)]
@@ -622,7 +622,7 @@ readLiteral:
 						f.err = CorruptInputError(f.roffset)
 						return
 					}
-					f.b = b >> (n & 31)
+					f.b = b >> (n & regSizeMaskUint32)
 					f.nb = nb - n
 					v = int(chunk >> huffmanValueShift)
 					break
@@ -685,12 +685,12 @@ readLiteral:
 					return
 				}
 			}
-			length += int(f.b & uint32(1<<n-1))
-			f.b >>= n
+			length += int(f.b & uint32(1<<(n&regSizeMaskUint32)-1))
+			f.b >>= n & regSizeMaskUint32
 			f.nb -= n
 		}
 
-		var dist int
+		var dist uint32
 		if f.hd == nil {
 			for f.nb < 5 {
 				if err = f.moreBits(); err != nil {
@@ -701,17 +701,19 @@ readLiteral:
 					return
 				}
 			}
-			dist = int(bits.Reverse8(uint8(f.b & 0x1F << 3)))
+			dist = uint32(bits.Reverse8(uint8(f.b & 0x1F << 3)))
 			f.b >>= 5
 			f.nb -= 5
 		} else {
-			if dist, err = f.huffSym(f.hd); err != nil {
+			sym, err := f.huffSym(f.hd)
+			if err != nil {
 				if debugDecode {
 					fmt.Println("huffsym:", err)
 				}
 				f.err = err
 				return
 			}
+			dist = uint32(sym)
 		}
 
 		switch {
@@ -720,7 +722,7 @@ readLiteral:
 		case dist < maxNumDist:
 			nb := uint(dist-2) >> 1
 			// have 1 bit in bottom of dist, need nb more.
-			extra := (dist & 1) << nb
+			extra := (dist & 1) << (nb & regSizeMaskUint32)
 			for f.nb < nb {
 				if err = f.moreBits(); err != nil {
 					if debugDecode {
@@ -730,10 +732,10 @@ readLiteral:
 					return
 				}
 			}
-			extra |= int(f.b & uint32(1<<nb-1))
-			f.b >>= nb
+			extra |= f.b & uint32(1<<(nb&regSizeMaskUint32)-1)
+			f.b >>= nb & regSizeMaskUint32
 			f.nb -= nb
-			dist = 1<<(nb+1) + 1 + extra
+			dist = 1<<((nb+1)&regSizeMaskUint32) + 1 + extra
 		default:
 			if debugDecode {
 				fmt.Println("dist too big:", dist, maxNumDist)
@@ -743,7 +745,7 @@ readLiteral:
 		}
 
 		// No check on length; encoding can be prescient.
-		if dist > f.dict.histSize() {
+		if dist > uint32(f.dict.histSize()) {
 			if debugDecode {
 				fmt.Println("dist > f.dict.histSize():", dist, f.dict.histSize())
 			}
@@ -751,7 +753,7 @@ readLiteral:
 			return
 		}
 
-		f.copyLen, f.copyDist = length, dist
+		f.copyLen, f.copyDist = length, int(dist)
 		goto copyHistory
 	}
 
@@ -869,7 +871,7 @@ func (f *decompressor) moreBits() error {
 		return noEOF(err)
 	}
 	f.roffset++
-	f.b |= uint32(c) << f.nb
+	f.b |= uint32(c) << (f.nb & regSizeMaskUint32)
 	f.nb += 8
 	return nil
 }
@@ -894,7 +896,7 @@ func (f *decompressor) huffSym(h *huffmanDecoder) (int, error) {
 				return 0, noEOF(err)
 			}
 			f.roffset++
-			b |= uint32(c) << (nb & 31)
+			b |= uint32(c) << (nb & regSizeMaskUint32)
 			nb += 8
 		}
 		chunk := h.chunks[b&(huffmanNumChunks-1)]
@@ -913,7 +915,7 @@ func (f *decompressor) huffSym(h *huffmanDecoder) (int, error) {
 				f.err = CorruptInputError(f.roffset)
 				return 0, f.err
 			}
-			f.b = b >> (n & 31)
+			f.b = b >> (n & regSizeMaskUint32)
 			f.nb = nb - n
 			return int(chunk >> huffmanValueShift), nil
 		}

vendor/github.com/klauspost/compress/flate/inflate_gen.go

@@ -63,7 +63,7 @@ readLiteral:
 						return
 					}
 					f.roffset++
-					b |= uint32(c) << (nb & 31)
+					b |= uint32(c) << (nb & regSizeMaskUint32)
 					nb += 8
 				}
 				chunk := f.hl.chunks[b&(huffmanNumChunks-1)]
@@ -82,7 +82,7 @@ readLiteral:
 						f.err = CorruptInputError(f.roffset)
 						return
 					}
-					f.b = b >> (n & 31)
+					f.b = b >> (n & regSizeMaskUint32)
 					f.nb = nb - n
 					v = int(chunk >> huffmanValueShift)
 					break
@@ -145,15 +145,15 @@ readLiteral:
 					return
 				}
 			}
-			length += int(f.b & uint32(1<<n-1))
-			f.b >>= n
+			length += int(f.b & uint32(1<<(n&regSizeMaskUint32)-1))
+			f.b >>= n & regSizeMaskUint32
 			f.nb -= n
 		}
 
-		var dist int
+		var dist uint32
 		if f.hd == nil {
 			for f.nb < 5 {
-				if err = moreBits(); err != nil {
+				if err = f.moreBits(); err != nil {
 					if debugDecode {
 						fmt.Println("morebits f.nb<5:", err)
 					}
@@ -161,17 +161,19 @@ readLiteral:
 					return
 				}
 			}
-			dist = int(bits.Reverse8(uint8(f.b & 0x1F << 3)))
+			dist = uint32(bits.Reverse8(uint8(f.b & 0x1F << 3)))
 			f.b >>= 5
 			f.nb -= 5
 		} else {
-			if dist, err = f.huffSym(f.hd); err != nil {
+			sym, err := f.huffSym(f.hd)
+			if err != nil {
 				if debugDecode {
 					fmt.Println("huffsym:", err)
 				}
 				f.err = err
 				return
 			}
+			dist = uint32(sym)
 		}
 
 		switch {
@@ -180,9 +182,9 @@ readLiteral:
 		case dist < maxNumDist:
 			nb := uint(dist-2) >> 1
 			// have 1 bit in bottom of dist, need nb more.
-			extra := (dist & 1) << nb
+			extra := (dist & 1) << (nb & regSizeMaskUint32)
 			for f.nb < nb {
-				if err = moreBits(); err != nil {
+				if err = f.moreBits(); err != nil {
 					if debugDecode {
 						fmt.Println("morebits f.nb<nb:", err)
 					}
@@ -190,10 +192,10 @@ readLiteral:
 					return
 				}
 			}
-			extra |= int(f.b & uint32(1<<nb-1))
-			f.b >>= nb
+			extra |= f.b & uint32(1<<(nb&regSizeMaskUint32)-1)
+			f.b >>= nb & regSizeMaskUint32
 			f.nb -= nb
-			dist = 1<<(nb+1) + 1 + extra
+			dist = 1<<((nb+1)&regSizeMaskUint32) + 1 + extra
 		default:
 			if debugDecode {
 				fmt.Println("dist too big:", dist, maxNumDist)
@@ -203,7 +205,7 @@ readLiteral:
 		}
 
 		// No check on length; encoding can be prescient.
-		if dist > f.dict.histSize() {
+		if dist > uint32(f.dict.histSize()) {
 			if debugDecode {
 				fmt.Println("dist > f.dict.histSize():", dist, f.dict.histSize())
 			}
@@ -211,7 +213,7 @@ readLiteral:
 			return
 		}
 
-		f.copyLen, f.copyDist = length, dist
+		f.copyLen, f.copyDist = length, int(dist)
 		goto copyHistory
 	}
 
@@ -287,7 +289,7 @@ readLiteral:
 						return
 					}
 					f.roffset++
-					b |= uint32(c) << (nb & 31)
+					b |= uint32(c) << (nb & regSizeMaskUint32)
 					nb += 8
 				}
 				chunk := f.hl.chunks[b&(huffmanNumChunks-1)]
@@ -306,7 +308,7 @@ readLiteral:
 						f.err = CorruptInputError(f.roffset)
 						return
 					}
-					f.b = b >> (n & 31)
+					f.b = b >> (n & regSizeMaskUint32)
 					f.nb = nb - n
 					v = int(chunk >> huffmanValueShift)
 					break
@@ -369,15 +371,15 @@ readLiteral:
 					return
 				}
 			}
-			length += int(f.b & uint32(1<<n-1))
-			f.b >>= n
+			length += int(f.b & uint32(1<<(n&regSizeMaskUint32)-1))
+			f.b >>= n & regSizeMaskUint32
 			f.nb -= n
 		}
 
-		var dist int
+		var dist uint32
 		if f.hd == nil {
 			for f.nb < 5 {
-				if err = moreBits(); err != nil {
+				if err = f.moreBits(); err != nil {
 					if debugDecode {
 						fmt.Println("morebits f.nb<5:", err)
 					}
@@ -385,17 +387,19 @@ readLiteral:
 					return
 				}
 			}
-			dist = int(bits.Reverse8(uint8(f.b & 0x1F << 3)))
+			dist = uint32(bits.Reverse8(uint8(f.b & 0x1F << 3)))
 			f.b >>= 5
 			f.nb -= 5
 		} else {
-			if dist, err = f.huffSym(f.hd); err != nil {
+			sym, err := f.huffSym(f.hd)
+			if err != nil {
 				if debugDecode {
 					fmt.Println("huffsym:", err)
 				}
 				f.err = err
 				return
 			}
+			dist = uint32(sym)
 		}
 
 		switch {
@@ -404,9 +408,9 @@ readLiteral:
 		case dist < maxNumDist:
 			nb := uint(dist-2) >> 1
 			// have 1 bit in bottom of dist, need nb more.
-			extra := (dist & 1) << nb
+			extra := (dist & 1) << (nb & regSizeMaskUint32)
 			for f.nb < nb {
-				if err = moreBits(); err != nil {
+				if err = f.moreBits(); err != nil {
 					if debugDecode {
 						fmt.Println("morebits f.nb<nb:", err)
 					}
@@ -414,10 +418,10 @@ readLiteral:
 					return
 				}
 			}
-			extra |= int(f.b & uint32(1<<nb-1))
-			f.b >>= nb
+			extra |= f.b & uint32(1<<(nb&regSizeMaskUint32)-1)
+			f.b >>= nb & regSizeMaskUint32
 			f.nb -= nb
-			dist = 1<<(nb+1) + 1 + extra
+			dist = 1<<((nb+1)&regSizeMaskUint32) + 1 + extra
 		default:
 			if debugDecode {
 				fmt.Println("dist too big:", dist, maxNumDist)
@@ -427,7 +431,7 @@ readLiteral:
 		}
 
 		// No check on length; encoding can be prescient.
-		if dist > f.dict.histSize() {
+		if dist > uint32(f.dict.histSize()) {
 			if debugDecode {
 				fmt.Println("dist > f.dict.histSize():", dist, f.dict.histSize())
 			}
@@ -435,7 +439,7 @@ readLiteral:
 			return
 		}
 
-		f.copyLen, f.copyDist = length, dist
+		f.copyLen, f.copyDist = length, int(dist)
 		goto copyHistory
 	}
 
@@ -511,7 +515,7 @@ readLiteral:
 						return
 					}
 					f.roffset++
-					b |= uint32(c) << (nb & 31)
+					b |= uint32(c) << (nb & regSizeMaskUint32)
 					nb += 8
 				}
 				chunk := f.hl.chunks[b&(huffmanNumChunks-1)]
@@ -530,7 +534,7 @@ readLiteral:
 						f.err = CorruptInputError(f.roffset)
 						return
 					}
-					f.b = b >> (n & 31)
+					f.b = b >> (n & regSizeMaskUint32)
 					f.nb = nb - n
 					v = int(chunk >> huffmanValueShift)
 					break
@@ -593,15 +597,15 @@ readLiteral:
 					return
 				}
 			}
-			length += int(f.b & uint32(1<<n-1))
-			f.b >>= n
+			length += int(f.b & uint32(1<<(n&regSizeMaskUint32)-1))
+			f.b >>= n & regSizeMaskUint32
 			f.nb -= n
 		}
 
-		var dist int
+		var dist uint32
 		if f.hd == nil {
 			for f.nb < 5 {
-				if err = moreBits(); err != nil {
+				if err = f.moreBits(); err != nil {
 					if debugDecode {
 						fmt.Println("morebits f.nb<5:", err)
 					}
@@ -609,17 +613,19 @@ readLiteral:
 					return
 				}
 			}
-			dist = int(bits.Reverse8(uint8(f.b & 0x1F << 3)))
+			dist = uint32(bits.Reverse8(uint8(f.b & 0x1F << 3)))
 			f.b >>= 5
 			f.nb -= 5
 		} else {
-			if dist, err = f.huffSym(f.hd); err != nil {
+			sym, err := f.huffSym(f.hd)
+			if err != nil {
 				if debugDecode {
 					fmt.Println("huffsym:", err)
 				}
 				f.err = err
 				return
 			}
+			dist = uint32(sym)
 		}
 
 		switch {
@@ -628,9 +634,9 @@ readLiteral:
 		case dist < maxNumDist:
 			nb := uint(dist-2) >> 1
 			// have 1 bit in bottom of dist, need nb more.
-			extra := (dist & 1) << nb
+			extra := (dist & 1) << (nb & regSizeMaskUint32)
 			for f.nb < nb {
-				if err = moreBits(); err != nil {
+				if err = f.moreBits(); err != nil {
 					if debugDecode {
 						fmt.Println("morebits f.nb<nb:", err)
 					}
@@ -638,10 +644,10 @@ readLiteral:
 					return
 				}
 			}
-			extra |= int(f.b & uint32(1<<nb-1))
-			f.b >>= nb
+			extra |= f.b & uint32(1<<(nb&regSizeMaskUint32)-1)
+			f.b >>= nb & regSizeMaskUint32
 			f.nb -= nb
-			dist = 1<<(nb+1) + 1 + extra
+			dist = 1<<((nb+1)&regSizeMaskUint32) + 1 + extra
 		default:
 			if debugDecode {
 				fmt.Println("dist too big:", dist, maxNumDist)
@@ -651,7 +657,7 @@ readLiteral:
 		}
 
 		// No check on length; encoding can be prescient.
-		if dist > f.dict.histSize() {
+		if dist > uint32(f.dict.histSize()) {
 			if debugDecode {
 				fmt.Println("dist > f.dict.histSize():", dist, f.dict.histSize())
 			}
@@ -659,7 +665,7 @@ readLiteral:
 			return
 		}
 
-		f.copyLen, f.copyDist = length, dist
+		f.copyLen, f.copyDist = length, int(dist)
 		goto copyHistory
 	}
 
@@ -735,7 +741,7 @@ readLiteral:
 						return
 					}
 					f.roffset++
-					b |= uint32(c) << (nb & 31)
+					b |= uint32(c) << (nb & regSizeMaskUint32)
 					nb += 8
 				}
 				chunk := f.hl.chunks[b&(huffmanNumChunks-1)]
@@ -754,7 +760,7 @@ readLiteral:
 						f.err = CorruptInputError(f.roffset)
 						return
 					}
-					f.b = b >> (n & 31)
+					f.b = b >> (n & regSizeMaskUint32)
 					f.nb = nb - n
 					v = int(chunk >> huffmanValueShift)
 					break
@@ -817,15 +823,15 @@ readLiteral:
 					return
 				}
 			}
-			length += int(f.b & uint32(1<<n-1))
-			f.b >>= n
+			length += int(f.b & uint32(1<<(n&regSizeMaskUint32)-1))
+			f.b >>= n & regSizeMaskUint32
 			f.nb -= n
 		}
 
-		var dist int
+		var dist uint32
 		if f.hd == nil {
 			for f.nb < 5 {
-				if err = moreBits(); err != nil {
+				if err = f.moreBits(); err != nil {
 					if debugDecode {
 						fmt.Println("morebits f.nb<5:", err)
 					}
@@ -833,17 +839,19 @@ readLiteral:
 					return
 				}
 			}
-			dist = int(bits.Reverse8(uint8(f.b & 0x1F << 3)))
+			dist = uint32(bits.Reverse8(uint8(f.b & 0x1F << 3)))
 			f.b >>= 5
 			f.nb -= 5
 		} else {
-			if dist, err = f.huffSym(f.hd); err != nil {
+			sym, err := f.huffSym(f.hd)
+			if err != nil {
 				if debugDecode {
 					fmt.Println("huffsym:", err)
 				}
 				f.err = err
 				return
 			}
+			dist = uint32(sym)
 		}
 
 		switch {
@@ -852,9 +860,9 @@ readLiteral:
 		case dist < maxNumDist:
 			nb := uint(dist-2) >> 1
 			// have 1 bit in bottom of dist, need nb more.
-			extra := (dist & 1) << nb
+			extra := (dist & 1) << (nb & regSizeMaskUint32)
 			for f.nb < nb {
-				if err = moreBits(); err != nil {
+				if err = f.moreBits(); err != nil {
 					if debugDecode {
 						fmt.Println("morebits f.nb<nb:", err)
 					}
@@ -862,10 +870,10 @@ readLiteral:
 					return
 				}
 			}
-			extra |= int(f.b & uint32(1<<nb-1))
-			f.b >>= nb
+			extra |= f.b & uint32(1<<(nb&regSizeMaskUint32)-1)
+			f.b >>= nb & regSizeMaskUint32
 			f.nb -= nb
-			dist = 1<<(nb+1) + 1 + extra
+			dist = 1<<((nb+1)&regSizeMaskUint32) + 1 + extra
 		default:
 			if debugDecode {
 				fmt.Println("dist too big:", dist, maxNumDist)
@@ -875,7 +883,7 @@ readLiteral:
 		}
 
 		// No check on length; encoding can be prescient.
-		if dist > f.dict.histSize() {
+		if dist > uint32(f.dict.histSize()) {
 			if debugDecode {
 				fmt.Println("dist > f.dict.histSize():", dist, f.dict.histSize())
 			}
@@ -883,7 +891,7 @@ readLiteral:
 			return
 		}
 
-		f.copyLen, f.copyDist = length, dist
+		f.copyLen, f.copyDist = length, int(dist)
 		goto copyHistory
 	}
 

vendor/github.com/klauspost/compress/flate/regmask_amd64.go

@@ -0,0 +1,37 @@
+package flate
+
+const (
+	// Masks for shifts with register sizes of the shift value.
+	// This can be used to work around the x86 design of shifting by mod register size.
+	// It can be used when a variable shift is always smaller than the register size.
+
+	// reg8SizeMaskX - shift value is 8 bits, shifted is X
+	reg8SizeMask8  = 7
+	reg8SizeMask16 = 15
+	reg8SizeMask32 = 31
+	reg8SizeMask64 = 63
+
+	// reg16SizeMaskX - shift value is 16 bits, shifted is X
+	reg16SizeMask8  = reg8SizeMask8
+	reg16SizeMask16 = reg8SizeMask16
+	reg16SizeMask32 = reg8SizeMask32
+	reg16SizeMask64 = reg8SizeMask64
+
+	// reg32SizeMaskX - shift value is 32 bits, shifted is X
+	reg32SizeMask8  = reg8SizeMask8
+	reg32SizeMask16 = reg8SizeMask16
+	reg32SizeMask32 = reg8SizeMask32
+	reg32SizeMask64 = reg8SizeMask64
+
+	// reg64SizeMaskX - shift value is 64 bits, shifted is X
+	reg64SizeMask8  = reg8SizeMask8
+	reg64SizeMask16 = reg8SizeMask16
+	reg64SizeMask32 = reg8SizeMask32
+	reg64SizeMask64 = reg8SizeMask64
+
+	// regSizeMaskUintX - shift value is uint, shifted is X
+	regSizeMaskUint8  = reg8SizeMask8
+	regSizeMaskUint16 = reg8SizeMask16
+	regSizeMaskUint32 = reg8SizeMask32
+	regSizeMaskUint64 = reg8SizeMask64
+)

vendor/github.com/klauspost/compress/flate/regmask_other.go

@@ -0,0 +1,39 @@
+//+build !amd64
+
+package flate
+
+const (
+	// Masks for shifts with register sizes of the shift value.
+	// This can be used to work around the x86 design of shifting by mod register size.
+	// It can be used when a variable shift is always smaller than the register size.
+
+	// reg8SizeMaskX - shift value is 8 bits, shifted is X
+	reg8SizeMask8  = 0xff
+	reg8SizeMask16 = 0xff
+	reg8SizeMask32 = 0xff
+	reg8SizeMask64 = 0xff
+
+	// reg16SizeMaskX - shift value is 16 bits, shifted is X
+	reg16SizeMask8  = 0xffff
+	reg16SizeMask16 = 0xffff
+	reg16SizeMask32 = 0xffff
+	reg16SizeMask64 = 0xffff
+
+	// reg32SizeMaskX - shift value is 32 bits, shifted is X
+	reg32SizeMask8  = 0xffffffff
+	reg32SizeMask16 = 0xffffffff
+	reg32SizeMask32 = 0xffffffff
+	reg32SizeMask64 = 0xffffffff
+
+	// reg64SizeMaskX - shift value is 64 bits, shifted is X
+	reg64SizeMask8  = 0xffffffffffffffff
+	reg64SizeMask16 = 0xffffffffffffffff
+	reg64SizeMask32 = 0xffffffffffffffff
+	reg64SizeMask64 = 0xffffffffffffffff
+
+	// regSizeMaskUintX - shift value is uint, shifted is X
+	regSizeMaskUint8  = ^uint(0)
+	regSizeMaskUint16 = ^uint(0)
+	regSizeMaskUint32 = ^uint(0)
+	regSizeMaskUint64 = ^uint(0)
+)