From e62906e63dee06492e4ed5f659f251394ba2a1c6 Mon Sep 17 00:00:00 2001
From: nkanaev <nkanaev@live.com>
Date: Mon, 14 Nov 2022 15:11:05 +0000
Subject: [PATCH] fix readability edge case

---
 doc/changelog.txt                      | 1 +
 src/content/readability/readability.go | 4 ++++
 src/server/routes.go                   | 5 +++--
 3 files changed, 8 insertions(+), 2 deletions(-)

diff --git a/doc/changelog.txt b/doc/changelog.txt
index 1552378..800baec 100644
--- a/doc/changelog.txt
+++ b/doc/changelog.txt
@@ -6,6 +6,7 @@
 - (fix) favicon visibility in dark mode (thanks to @caycaycarly for the report)
 - (fix) autoloading more articles not working in certain edge cases (thanks to @fenuks for the report)
 - (fix) handle Google URL redirects in "Read Here" (thanks to @cubbei for discovery)
+- (fix) handle failures to extract content in "Read Here" (thanks to @grigio for the report)
 
 # v2.3 (2022-05-03)
 
diff --git a/src/content/readability/readability.go b/src/content/readability/readability.go
index 324d7f3..86e9086 100644
--- a/src/content/readability/readability.go
+++ b/src/content/readability/readability.go
@@ -6,6 +6,7 @@ package readability
 
 import (
 	"bytes"
+	"errors"
 	"fmt"
 	"io"
 	"math"
@@ -59,6 +60,9 @@ func ExtractContent(page io.Reader) (string, error) {
 			best = body
 			break
 		}
+        if best == nil {
+            return "", errors.New("failed to extract content")
+        }
 	}
 	//log.Printf("[Readability] TopCandidate: %v", topCandidate)
 
diff --git a/src/server/routes.go b/src/server/routes.go
index 00991e2..d9fa61a 100644
--- a/src/server/routes.go
+++ b/src/server/routes.go
@@ -474,8 +474,9 @@ func (s *Server) handlePageCrawl(c *router.Context) {
 	}
 	content, err := readability.ExtractContent(strings.NewReader(body))
 	if err != nil {
-		log.Print(err)
-		c.Out.WriteHeader(http.StatusNoContent)
+        c.JSON(http.StatusOK, map[string]string{
+            "content": "error: " + err.Error(),
+        })
 		return
 	}
 	content = sanitizer.Sanitize(url, content)