Better detection if a text starts with html or not

2 files changed, 13 insertions, 4 deletions

diff --git a/go.mod b/go.mod
index 0ff1224..1d1041d 100644
--- a/go.mod
+++ b/go.mod
@@ -11,6 +11,6 @@ require (
 	github.com/google/go-cmp v0.4.0
 	github.com/google/uuid v1.1.1
 	github.com/mmcdole/gofeed v1.0.0-beta2.0.20200331235650-4298e4366be3
-	golang.org/x/net v0.0.0-20200501053045-e0ff5e5a1de5 // indirect
+	golang.org/x/net v0.0.0-20200501053045-e0ff5e5a1de5
 	gopkg.in/yaml.v3 v3.0.0-20200313102051-9f266ea9e77c
 )
diff --git a/internal/feed/mail.go b/internal/feed/mail.go
index 0a8bd10..f77987e 100644
--- a/internal/feed/mail.go
+++ b/internal/feed/mail.go
@@ -16,6 +16,7 @@ import (
 	"github.com/emersion/go-message"
 	"github.com/emersion/go-message/mail"
 	"github.com/gabriel-vasile/mimetype"
+	"golang.org/x/net/html"
 
 	"github.com/Necoro/feed2imap-go/internal/feed/template"
 	"github.com/Necoro/feed2imap-go/internal/http"
@@ -240,6 +241,13 @@ func getBody(content, description string, bodyCfg config.Body) string {
 	}
 }
 
+func startsWithText(str string) bool {
+	reader := strings.NewReader(str)
+	tokenizer := html.NewTokenizerFragment(reader, "")
+
+	return tokenizer.Next() == html.TextToken
+}
+
 func (item *item) buildBody() {
 	feed := item.feed
 	feedUrl, err := url.Parse(feed.Url)
@@ -248,6 +256,9 @@ func (item *item) buildBody() {
 	}
 
 	body := getBody(item.Content, item.Description, feed.Body)
+	if body != "" && startsWithText(body) {
+		body = "<br />" + body
+	}
 
 	if !feed.InclImages {
 		item.Body = body
@@ -257,9 +268,7 @@ func (item *item) buildBody() {
 	doc, err := goquery.NewDocumentFromReader(strings.NewReader(body))
 	if err != nil {
 		log.Errorf("Feed %s: Item %s: Error while parsing html content: %s", feed.Name, item.Link, err)
-		if body != "" {
-			item.Body = "<br />" + body
-		}
+		item.Body = body
 		return
 	}