[bugfix] Update exif-terminator (fix png issue) (#2391)

* [bugfix] Update exif-terminator (fix png issue)

* bump exif terminator

* fix tests

3 files changed, 93 insertions, 116 deletions

diff --git a/vendor/github.com/superseriousbusiness/exif-terminator/logger.go b/vendor/github.com/superseriousbusiness/exif-terminator/logger.go
deleted file mode 100644
index e607c55dd..000000000
--- a/vendor/github.com/superseriousbusiness/exif-terminator/logger.go
+++ /dev/null
@@ -1,47 +0,0 @@
-/*
-   exif-terminator
-   Copyright (C) 2022 SuperSeriousBusiness admin@gotosocial.org
-
-   This program is free software: you can redistribute it and/or modify
-   it under the terms of the GNU Affero General Public License as published by
-   the Free Software Foundation, either version 3 of the License, or
-   (at your option) any later version.
-
-   This program is distributed in the hope that it will be useful,
-   but WITHOUT ANY WARRANTY; without even the implied warranty of
-   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-   GNU Affero General Public License for more details.
-
-   You should have received a copy of the GNU Affero General Public License
-   along with this program.  If not, see <http://www.gnu.org/licenses/>.
-*/
-
-package terminator
-
-import "fmt"
-
-var logger ErrorLogger
-
-func init() {
-	logger = &defaultErrorLogger{}
-}
-
-// ErrorLogger denotes a generic error logging function.
-type ErrorLogger interface {
-	Error(args ...interface{})
-}
-
-type defaultErrorLogger struct{}
-
-func (d *defaultErrorLogger) Error(args ...interface{}) {
-	fmt.Println(args...)
-}
-
-// SetErrorLogger allows a user of the exif-terminator library
-// to set the logger that will be used for error logging.
-//
-// If it is not set, the default error logger will be used, which
-// just prints errors to stdout.
-func SetErrorLogger(errorLogger ErrorLogger) {
-	logger = errorLogger
-}
diff --git a/vendor/github.com/superseriousbusiness/exif-terminator/png.go b/vendor/github.com/superseriousbusiness/exif-terminator/png.go
index 4a1ac5bf1..774ec0ed6 100644
--- a/vendor/github.com/superseriousbusiness/exif-terminator/png.go
+++ b/vendor/github.com/superseriousbusiness/exif-terminator/png.go
@@ -19,10 +19,9 @@
 package terminator
 
 import (
-	"encoding/binary"
 	"io"
 
-	pngstructure "github.com/dsoprea/go-png-image-structure/v2"
+	pngstructure "github.com/superseriousbusiness/go-png-image-structure/v2"
 )
 
 type pngVisitor struct {
@@ -45,49 +44,50 @@ func (v *pngVisitor) split(data []byte, atEOF bool) (int, []byte, error) {
 		}
 	}
 
-	// check if the splitter has any new chunks in it that we haven't written yet
-	chunkSlice := v.ps.Chunks()
+	// Check if the splitter now has
+	// any new chunks in it for us.
+	chunkSlice, err := v.ps.Chunks()
+	if err != nil {
+		return advance, token, err
+	}
+
+	// Write each chunk by passing it
+	// through our custom write func,
+	// which strips out exif and fixes
+	// the CRC of each chunk.
 	chunks := chunkSlice.Chunks()
 	for i, chunk := range chunks {
-		// look through all the chunks in the splitter
-		if i > v.lastWrittenChunk {
-			// we've got a chunk we haven't written yet! write it...
-			if err := v.writeChunk(chunk); err != nil {
-				return advance, token, err
-			}
-			// then remove the data
-			chunk.Data = chunk.Data[:0]
-			// and update
-			v.lastWrittenChunk = i
+		if i <= v.lastWrittenChunk {
+			// Skip already
+			// written chunks.
+			continue
 		}
+
+		// Write this new chunk.
+		if err := v.writeChunk(chunk); err != nil {
+			return advance, token, err
+		}
+		v.lastWrittenChunk = i
+
+		// Zero data; here you
+		// go garbage collector.
+		chunk.Data = nil
 	}
 
 	return advance, token, err
 }
 
 func (v *pngVisitor) writeChunk(chunk *pngstructure.Chunk) error {
-	if err := binary.Write(v.writer, binary.BigEndian, chunk.Length); err != nil {
-		return err
-	}
-
-	if _, err := v.writer.Write([]byte(chunk.Type)); err != nil {
-		return err
-	}
-
 	if chunk.Type == pngstructure.EXifChunkType {
-		blank := make([]byte, len(chunk.Data))
-		if _, err := v.writer.Write(blank); err != nil {
-			return err
-		}
-	} else {
-		if _, err := v.writer.Write(chunk.Data); err != nil {
-			return err
-		}
+		// Replace exif data
+		// with zero bytes.
+		clear(chunk.Data)
 	}
 
-	if err := binary.Write(v.writer, binary.BigEndian, chunk.Crc); err != nil {
-		return err
-	}
+	// Fix CRC of each chunk.
+	chunk.UpdateCrc32()
 
-	return nil
+	// finally, write chunk to writer.
+	_, err := chunk.WriteTo(v.writer)
+	return err
 }
diff --git a/vendor/github.com/superseriousbusiness/exif-terminator/terminator.go b/vendor/github.com/superseriousbusiness/exif-terminator/terminator.go
index 9d9e6e743..7dd3d9ad7 100644
--- a/vendor/github.com/superseriousbusiness/exif-terminator/terminator.go
+++ b/vendor/github.com/superseriousbusiness/exif-terminator/terminator.go
@@ -25,29 +25,34 @@ import (
 	"fmt"
 	"io"
 
-	pngstructure "github.com/dsoprea/go-png-image-structure/v2"
 	jpegstructure "github.com/superseriousbusiness/go-jpeg-image-structure/v2"
+	pngstructure "github.com/superseriousbusiness/go-png-image-structure/v2"
 )
 
 func Terminate(in io.Reader, fileSize int, mediaType string) (io.Reader, error) {
-	// to avoid keeping too much stuff in memory we want to pipe data directly
+	// To avoid keeping too much stuff
+	// in memory we want to pipe data
+	// directly to the reader.
 	pipeReader, pipeWriter := io.Pipe()
 
-	// we don't know ahead of time how long segments might be: they could be as large as
-	// the file itself, so unfortunately we need to allocate a buffer here that'scanner as large
-	// as the file
+	// We don't know ahead of time how long
+	// segments might be: they could be as
+	// large as the file itself, so we need
+	// a buffer with generous overhead.
 	scanner := bufio.NewScanner(in)
 	scanner.Buffer([]byte{}, fileSize)
-	var err error
 
+	var err error
 	switch mediaType {
 	case "image/jpeg", "jpeg", "jpg":
 		err = terminateJpeg(scanner, pipeWriter, fileSize)
+
 	case "image/webp", "webp":
 		err = terminateWebp(scanner, pipeWriter)
+
 	case "image/png", "png":
-		// for pngs we need to skip the header bytes, so read them in
-		// and check we're really dealing with a png here
+		// For pngs we need to skip the header bytes, so read
+		// them in and check we're really dealing with a png.
 		header := make([]byte, len(pngstructure.PngSignature))
 		if _, headerError := in.Read(header); headerError != nil {
 			err = headerError
@@ -67,68 +72,87 @@ func Terminate(in io.Reader, fileSize int, mediaType string) (io.Reader, error)
 	return pipeReader, err
 }
 
-func terminateJpeg(scanner *bufio.Scanner, writer io.WriteCloser, expectedFileSize int) error {
-	// jpeg visitor is where the spicy hack of streaming the de-exifed data is contained
+func terminateJpeg(scanner *bufio.Scanner, writer *io.PipeWriter, expectedFileSize int) error {
 	v := &jpegVisitor{
 		writer:           writer,
 		expectedFileSize: expectedFileSize,
 	}
 
-	// provide the visitor to the splitter so that it triggers on every section scan
+	// Provide the visitor to the splitter so
+	// that it triggers on every section scan.
 	js := jpegstructure.NewJpegSplitter(v)
 
-	// the visitor also needs to read back the list of segments: for this it needs
-	// to know what jpeg splitter it's attached to, so give it a pointer to the splitter
+	// The visitor also needs to read back the
+	// list of segments: for this it needs to
+	// know what jpeg splitter it's attached to,
+	// so give it a pointer to the splitter.
 	v.js = js
 
-	// use the jpeg splitters 'split' function, which satisfies the bufio.SplitFunc interface
+	// Jpeg visitor's 'split' function
+	// satisfies bufio.SplitFunc{}.
 	scanner.Split(js.Split)
 
-	scanAndClose(scanner, writer)
+	go scanAndClose(scanner, writer)
 	return nil
 }
 
-func terminateWebp(scanner *bufio.Scanner, writer io.WriteCloser) error {
+func terminateWebp(scanner *bufio.Scanner, writer *io.PipeWriter) error {
 	v := &webpVisitor{
 		writer: writer,
 	}
 
-	// use the webp visitor's 'split' function, which satisfies the bufio.SplitFunc interface
+	// Webp visitor's 'split' function
+	// satisfies bufio.SplitFunc{}.
 	scanner.Split(v.split)
 
-	scanAndClose(scanner, writer)
+	go scanAndClose(scanner, writer)
 	return nil
 }
 
-func terminatePng(scanner *bufio.Scanner, writer io.WriteCloser) error {
+func terminatePng(scanner *bufio.Scanner, writer *io.PipeWriter) error {
 	ps := pngstructure.NewPngSplitter()
 
+	// Don't bother checking CRC;
+	// we're overwriting it anyway.
+	ps.DoCheckCrc(false)
+
 	v := &pngVisitor{
 		ps:               ps,
 		writer:           writer,
 		lastWrittenChunk: -1,
 	}
 
-	// use the png visitor's 'split' function, which satisfies the bufio.SplitFunc interface
+	// Png visitor's 'split' function
+	// satisfies bufio.SplitFunc{}.
 	scanner.Split(v.split)
 
-	scanAndClose(scanner, writer)
+	go scanAndClose(scanner, writer)
 	return nil
 }
 
-func scanAndClose(scanner *bufio.Scanner, writer io.WriteCloser) {
-	// scan asynchronously until there's nothing left to scan, and then close the writer
-	// so that the reader on the other side knows that we're done
-	//
-	// due to the nature of io.Pipe, writing won't actually work
-	// until the pipeReader starts being read by the caller, which
-	// is why we do this asynchronously
-	go func() {
-		defer writer.Close()
-		for scanner.Scan() {
-		}
-		if scanner.Err() != nil {
-			logger.Error(scanner.Err())
-		}
+// scanAndClose scans through the given scanner until there's
+// nothing left to scan, and then closes the writer so that the
+// reader on the other side of the pipe knows that we're done.
+//
+// Any error encountered when scanning will be logged by terminator.
+//
+// Due to the nature of io.Pipe, writing won't actually work
+// until the pipeReader starts being read by the caller, which
+// is why this function should always be called asynchronously.
+func scanAndClose(scanner *bufio.Scanner, writer *io.PipeWriter) {
+	var err error
+
+	defer func() {
+		// Always close writer, using returned
+		// scanner error (if any). If err is nil
+		// then the standard io.EOF will be used.
+		// (this will not overwrite existing).
+		writer.CloseWithError(err)
 	}()
+
+	for scanner.Scan() {
+	}
+
+	// Set error on return.
+	err = scanner.Err()
 }