Add WithIfModifiedSince to enable use of caching

The HTTP Getter can leverage the use of RFC7234 4.3.2 by setting the
HTTP Header 'If-Modified-Since'. This will enable callers to avoid
downloading index files when they haven't been modified since the
last get operation.

Signed-off-by: Paulo Gomes <pjbgf@linux.com>

pkg/getter/getter.go

@@ -45,6 +45,7 @@ type options struct {
 	registryClient        *registry.Client
 	timeout               time.Duration
 	transport             *http.Transport
+	ifModifiedSince       *time.Time
 }
 
 // Option allows specifying various settings configurable by the user for overriding the defaults
@@ -121,6 +122,20 @@ func WithUntar() Option {
 	}
 }
 
+// WithIfModifiedSince leverages the HTTP caching mechanism
+// defined on RFC7234 4.3.2. It prevents the HTTP getter from
+// downloading index that haven't been modified based on `since`.
+// If the file was not modified, getter will return a
+// `304 Not Modified`, as per RFC.
+//
+// The caller is responsible for caching previously downloaded
+// index files and handling the returning error.
+func WithIfModifiedSince(since time.Time) Option {
+	return func(opts *options) {
+		opts.ifModifiedSince = &since
+	}
+}
+
 // WithTransport sets the http.Transport to allow overwriting the HTTPGetter default.
 func WithTransport(transport *http.Transport) Option {
 	return func(opts *options) {

pkg/getter/httpgetter.go

@@ -22,6 +22,7 @@ import (
 	"net/http"
 	"net/url"
 	"sync"
+	"time"
 
 	"github.com/pkg/errors"
 
@@ -58,6 +59,10 @@ func (g *HTTPGetter) get(href string) (*bytes.Buffer, error) {
 		req.Header.Set("User-Agent", g.opts.userAgent)
 	}
 
+	if g.opts.ifModifiedSince != nil {
+		req.Header.Set("If-Modified-Since", g.opts.ifModifiedSince.Format(time.RFC1123))
+	}
+
 	// Before setting the basic auth credentials, make sure the URL associated
 	// with the basic auth is the one being fetched.
 	u1, err := url.Parse(g.opts.url)

pkg/getter/httpgetter_test.go

@@ -530,3 +530,62 @@ func TestHTTPTransportOption(t *testing.T) {
 		t.Fatal("transport.TLSClientConfig should not be set")
 	}
 }
+
+func TestHTTPIfModifiedSinceOption(t *testing.T) {
+	sinceNow := time.Now()
+	sinceNowExpected := sinceNow.Format(time.RFC1123)
+
+	tests := []struct {
+		name          string
+		since         *time.Time
+		sinceExpected string
+	}{
+		{
+			name:          "If-Modified-Since option set",
+			since:         &sinceNow,
+			sinceExpected: sinceNowExpected,
+		},
+		{
+			name: "If-Modified-Since option not set",
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			var g Getter
+			var err error
+
+			if tt.since == nil {
+				g, err = NewHTTPGetter()
+			} else {
+				g, err = NewHTTPGetter(
+					WithIfModifiedSince(*tt.since),
+				)
+			}
+			if err != nil {
+				t.Fatal(err)
+			}
+
+			srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+				gotSince, found := r.Header["If-Modified-Since"]
+				if tt.since != nil && !found {
+					t.Errorf("Header If-Modified-Since was expected but was not set")
+				}
+
+				if tt.since != nil && gotSince[0] != tt.sinceExpected {
+					t.Errorf("Expected '%s', got '%s'", tt.sinceExpected, gotSince)
+				}
+
+				if tt.since == nil && found {
+					t.Errorf("Header If-Modified-Since was not expected but was set")
+				}
+			}))
+			defer srv.Close()
+
+			_, err = g.Get(srv.URL, WithURL(srv.URL))
+			if err != nil {
+				t.Fatal(err)
+			}
+		})
+	}
+}