3 files changed, 202 insertions, 0 deletions
diff --git a/workhorse/internal/upload/exif/exif.go b/workhorse/internal/upload/exif/exif.go
new file mode 100644
index 00000000000..a9307b1ca90
--- /dev/null
+++ b/workhorse/internal/upload/exif/exif.go
@@ -0,0 +1,107 @@
+package exif
+
+import (
+	"bytes"
+	"context"
+	"errors"
+	"fmt"
+	"io"
+	"os/exec"
+	"regexp"
+
+	"gitlab.com/gitlab-org/labkit/log"
+)
+
+var ErrRemovingExif = errors.New("error while removing EXIF")
+
+type cleaner struct {
+	ctx    context.Context
+	cmd    *exec.Cmd
+	stdout io.Reader
+	stderr bytes.Buffer
+	eof    bool
+}
+
+func NewCleaner(ctx context.Context, stdin io.Reader) (io.ReadCloser, error) {
+	c := &cleaner{ctx: ctx}
+
+	if err := c.startProcessing(stdin); err != nil {
+		return nil, err
+	}
+
+	return c, nil
+}
+
+func (c *cleaner) Close() error {
+	if c.cmd == nil {
+		return nil
+	}
+
+	return c.cmd.Wait()
+}
+
+func (c *cleaner) Read(p []byte) (int, error) {
+	if c.eof {
+		return 0, io.EOF
+	}
+
+	n, err := c.stdout.Read(p)
+	if err == io.EOF {
+		if waitErr := c.cmd.Wait(); waitErr != nil {
+			log.WithContextFields(c.ctx, log.Fields{
+				"command": c.cmd.Args,
+				"stderr":  c.stderr.String(),
+				"error":   waitErr.Error(),
+			}).Print("exiftool command failed")
+
+			return n, ErrRemovingExif
+		}
+
+		c.eof = true
+	}
+
+	return n, err
+}
+
+func (c *cleaner) startProcessing(stdin io.Reader) error {
+	var err error
+
+	whitelisted_tags := []string{
+		"-ResolutionUnit",
+		"-XResolution",
+		"-YResolution",
+		"-YCbCrSubSampling",
+		"-YCbCrPositioning",
+		"-BitsPerSample",
+		"-ImageHeight",
+		"-ImageWidth",
+		"-ImageSize",
+		"-Copyright",
+		"-CopyrightNotice",
+		"-Orientation",
+	}
+
+	args := append([]string{"-all=", "--IPTC:all", "--XMP-iptcExt:all", "-tagsFromFile", "@"}, whitelisted_tags...)
+	args = append(args, "-")
+	c.cmd = exec.CommandContext(c.ctx, "exiftool", args...)
+
+	c.cmd.Stderr = &c.stderr
+	c.cmd.Stdin = stdin
+
+	c.stdout, err = c.cmd.StdoutPipe()
+	if err != nil {
+		return fmt.Errorf("failed to create stdout pipe: %v", err)
+	}
+
+	if err = c.cmd.Start(); err != nil {
+		return fmt.Errorf("start %v: %v", c.cmd.Args, err)
+	}
+
+	return nil
+}
+
+func IsExifFile(filename string) bool {
+	filenameMatch := regexp.MustCompile(`(?i)\.(jpg|jpeg|tiff)$`)
+
+	return filenameMatch.MatchString(filename)
+}
diff --git a/workhorse/internal/upload/exif/exif_test.go b/workhorse/internal/upload/exif/exif_test.go
new file mode 100644
index 00000000000..373d97f7fce
--- /dev/null
+++ b/workhorse/internal/upload/exif/exif_test.go
@@ -0,0 +1,95 @@
+package exif
+
+import (
+	"context"
+	"io"
+	"io/ioutil"
+	"os"
+	"strings"
+	"testing"
+
+	"github.com/stretchr/testify/require"
+)
+
+func TestIsExifFile(t *testing.T) {
+	tests := []struct {
+		name     string
+		expected bool
+	}{
+		{
+			name:     "/full/path.jpg",
+			expected: true,
+		},
+		{
+			name:     "path.jpeg",
+			expected: true,
+		},
+		{
+			name:     "path.tiff",
+			expected: true,
+		},
+		{
+			name:     "path.JPG",
+			expected: true,
+		},
+		{
+			name:     "path.tar",
+			expected: false,
+		},
+		{
+			name:     "path",
+			expected: false,
+		},
+	}
+	for _, test := range tests {
+		t.Run(test.name, func(t *testing.T) {
+			require.Equal(t, test.expected, IsExifFile(test.name))
+		})
+	}
+}
+
+func TestNewCleanerWithValidFile(t *testing.T) {
+	input, err := os.Open("testdata/sample_exif.jpg")
+	require.NoError(t, err)
+	ctx, cancel := context.WithCancel(context.Background())
+	defer cancel()
+
+	cleaner, err := NewCleaner(ctx, input)
+	require.NoError(t, err, "Expected no error when creating cleaner command")
+
+	size, err := io.Copy(ioutil.Discard, cleaner)
+	require.NoError(t, err, "Expected no error when reading output")
+
+	sizeAfterStrip := int64(25399)
+	require.Equal(t, sizeAfterStrip, size, "Different size of converted image")
+}
+
+func TestNewCleanerWithInvalidFile(t *testing.T) {
+	ctx, cancel := context.WithCancel(context.Background())
+	defer cancel()
+
+	cleaner, err := NewCleaner(ctx, strings.NewReader("invalid image"))
+	require.NoError(t, err, "Expected no error when creating cleaner command")
+
+	size, err := io.Copy(ioutil.Discard, cleaner)
+	require.Error(t, err, "Expected error when reading output")
+	require.Equal(t, int64(0), size, "Size of invalid image should be 0")
+}
+
+func TestNewCleanerReadingAfterEOF(t *testing.T) {
+	input, err := os.Open("testdata/sample_exif.jpg")
+	require.NoError(t, err)
+	ctx, cancel := context.WithCancel(context.Background())
+	defer cancel()
+
+	cleaner, err := NewCleaner(ctx, input)
+	require.NoError(t, err, "Expected no error when creating cleaner command")
+
+	_, err = io.Copy(ioutil.Discard, cleaner)
+	require.NoError(t, err, "Expected no error when reading output")
+
+	buf := make([]byte, 1)
+	size, err := cleaner.Read(buf)
+	require.Equal(t, 0, size, "The output was already consumed by previous reads")
+	require.Equal(t, io.EOF, err, "We return EOF")
+}
diff --git a/workhorse/internal/upload/exif/testdata/sample_exif.jpg b/workhorse/internal/upload/exif/testdata/sample_exif.jpg
new file mode 100644
index 00000000000..05eda3f7f95
--- /dev/null
+++ b/workhorse/internal/upload/exif/testdata/sample_exif.jpg