diff options
Diffstat (limited to 'packfile')
| -rw-r--r-- | packfile/delta/apply/apply.go | 160 | ||||
| -rw-r--r-- | packfile/delta/apply/header.go | 47 | ||||
| -rw-r--r-- | packfile/delta/doc.go | 2 | ||||
| -rw-r--r-- | packfile/ingest/delta_header.go | 2 | ||||
| -rw-r--r-- | packfile/ingest/record_delta.go | 2 |
5 files changed, 211 insertions, 2 deletions
diff --git a/packfile/delta/apply/apply.go b/packfile/delta/apply/apply.go new file mode 100644 index 00000000..f5006e3c --- /dev/null +++ b/packfile/delta/apply/apply.go @@ -0,0 +1,160 @@ +// Package apply applies Git delta instruction streams. +package apply + +import "fmt" + +// Apply applies one Git delta instruction stream to base. +func Apply(base, delta []byte) ([]byte, error) { + pos := 0 + + srcSize, err := readVarint(delta, &pos) + if err != nil { + return nil, err + } + + dstSize, err := readVarint(delta, &pos) + if err != nil { + return nil, err + } + + if srcSize != len(base) { + return nil, fmt.Errorf("delta/apply: delta source size mismatch: got %d want %d", srcSize, len(base)) + } + + out := make([]byte, dstSize) + outPos := 0 + + for pos < len(delta) { + op := delta[pos] + pos++ + + //nolint:nestif + if op&0x80 != 0 { + off := 0 + + if op&0x01 != 0 { + if pos >= len(delta) { + return nil, fmt.Errorf("delta/apply: malformed delta copy offset") + } + + off |= int(delta[pos]) + pos++ + } + + if op&0x02 != 0 { + if pos >= len(delta) { + return nil, fmt.Errorf("delta/apply: malformed delta copy offset") + } + + off |= int(delta[pos]) << 8 + pos++ + } + + if op&0x04 != 0 { + if pos >= len(delta) { + return nil, fmt.Errorf("delta/apply: malformed delta copy offset") + } + + off |= int(delta[pos]) << 16 + pos++ + } + + if op&0x08 != 0 { + if pos >= len(delta) { + return nil, fmt.Errorf("delta/apply: malformed delta copy offset") + } + + off |= int(delta[pos]) << 24 + pos++ + } + + n := 0 + + if op&0x10 != 0 { + if pos >= len(delta) { + return nil, fmt.Errorf("delta/apply: malformed delta copy size") + } + + n |= int(delta[pos]) + pos++ + } + + if op&0x20 != 0 { + if pos >= len(delta) { + return nil, fmt.Errorf("delta/apply: malformed delta copy size") + } + + n |= int(delta[pos]) << 8 + pos++ + } + + if op&0x40 != 0 { + if pos >= len(delta) { + return nil, fmt.Errorf("delta/apply: malformed delta copy size") + } + + n |= int(delta[pos]) << 16 + pos++ + } + + if n == 0 { + n = 0x10000 + } + + if off < 0 || n < 0 || off+n > len(base) || outPos+n > len(out) { + return nil, fmt.Errorf("delta/apply: delta copy out of bounds") + } + + copy(out[outPos:outPos+n], base[off:off+n]) + outPos += n + + continue + } + + if op == 0 { + return nil, fmt.Errorf("delta/apply: invalid delta opcode 0") + } + + n := int(op) + if pos+n > len(delta) || outPos+n > len(out) { + return nil, fmt.Errorf("delta/apply: delta insert out of bounds") + } + + copy(out[outPos:outPos+n], delta[pos:pos+n]) + outPos += n + pos += n + } + + if outPos != len(out) { + return nil, fmt.Errorf("delta/apply: delta output size mismatch: got %d want %d", outPos, len(out)) + } + + return out, nil +} + +// readVarint parses one Git delta varint and advances pos. +func readVarint(buf []byte, pos *int) (int, error) { + value := 0 + shift := uint(0) + + for { + if *pos >= len(buf) { + return 0, fmt.Errorf("delta/apply: malformed delta varint") + } + + b := buf[*pos] + *pos++ + + value |= int(b&0x7f) << shift + if b&0x80 == 0 { + break + } + + shift += 7 + if shift > 63 { + return 0, fmt.Errorf("delta/apply: delta varint overflow") + } + } + + return value, nil +} diff --git a/packfile/delta/apply/header.go b/packfile/delta/apply/header.go new file mode 100644 index 00000000..69c9659a --- /dev/null +++ b/packfile/delta/apply/header.go @@ -0,0 +1,47 @@ +package apply + +import ( + "fmt" + "io" +) + +// ReadHeaderSizes reads the first two varints in one inflated delta stream. +// +// Callers that continue reading the same stream should pass their own buffered +// byte reader and keep using that same reader afterwards. +func ReadHeaderSizes(reader io.ByteReader) (int, int, error) { + srcSize, err := readVarintFromByteReader(reader) + if err != nil { + return 0, 0, err + } + + dstSize, err := readVarintFromByteReader(reader) + if err != nil { + return 0, 0, err + } + + return srcSize, dstSize, nil +} + +// readVarintFromByteReader parses one Git delta varint from reader. +func readVarintFromByteReader(reader io.ByteReader) (int, error) { + value := 0 + shift := uint(0) + + for { + b, err := reader.ReadByte() + if err != nil { + return 0, fmt.Errorf("delta/apply: malformed delta varint: %w", err) + } + + value |= int(b&0x7f) << shift + if b&0x80 == 0 { + return value, nil + } + + shift += 7 + if shift > 63 { + return 0, fmt.Errorf("delta/apply: delta varint overflow") + } + } +} diff --git a/packfile/delta/doc.go b/packfile/delta/doc.go new file mode 100644 index 00000000..f63c96a8 --- /dev/null +++ b/packfile/delta/doc.go @@ -0,0 +1,2 @@ +// Package delta provides various routines to handle Git delta compression. +package delta diff --git a/packfile/ingest/delta_header.go b/packfile/ingest/delta_header.go index 76e90172..63fda066 100644 --- a/packfile/ingest/delta_header.go +++ b/packfile/ingest/delta_header.go @@ -1,6 +1,6 @@ package ingest -import deltaapply "codeberg.org/lindenii/furgit/delta/apply" +import deltaapply "codeberg.org/lindenii/furgit/packfile/delta/apply" // finalizeStreamPackHash consumes trailer bytes and verifies stream integrity. // readDeltaHeaderSizes reads source and destination sizes from one delta payload. diff --git a/packfile/ingest/record_delta.go b/packfile/ingest/record_delta.go index f1637f5c..69cb8524 100644 --- a/packfile/ingest/record_delta.go +++ b/packfile/ingest/record_delta.go @@ -3,8 +3,8 @@ package ingest import ( "fmt" - deltaapply "codeberg.org/lindenii/furgit/delta/apply" objecttype "codeberg.org/lindenii/furgit/object/type" + deltaapply "codeberg.org/lindenii/furgit/packfile/delta/apply" ) // applyDeltaRecord applies one delta record onto base content. |
