Merge remote-tracking branch 'origin/main' into dev

internal/channels/telegram/format.go

@@ -61,6 +61,17 @@ func markdownToTelegramHTML(text string) string {
 	inlineCodes := extractInlineCodes(text)
 	text = inlineCodes.text
 
+
+	// Extract and protect bare URLs from italic parsing.
+	// URLs with underscores (e.g. syngas_dailymail_2026_ai) get broken by
+	// the italic regex which matches _text_ patterns inside URLs.
+	var urlPlaceholders []string
+	reURL := regexp.MustCompile(`https?://[^\s<>\)\]]+`)
+	text = reURL.ReplaceAllStringFunc(text, func(s string) string {
+		idx := len(urlPlaceholders)
+		urlPlaceholders = append(urlPlaceholders, s)
+		return fmt.Sprintf("\x00URL%d\x00", idx)
+	})
 	// Strip markdown headers
 	text = regexp.MustCompile(`(?m)^#{1,6}\s+(.+)$`).ReplaceAllString(text, "$1")
 
@@ -115,6 +126,12 @@ func markdownToTelegramHTML(text string) string {
 	// List items
 	text = regexp.MustCompile(`(?m)^[-*]\s+`).ReplaceAllString(text, "• ")
 
+	// Restore bare URLs (protected from italic parsing above).
+	for i, u := range urlPlaceholders {
+		escaped := escapeHTML(u)
+		text = strings.ReplaceAll(text, fmt.Sprintf("\x00URL%d\x00", i), escaped)
+	}
+
 	// Restore inline code
 	for i, code := range inlineCodes.codes {
 		escaped := escapeHTML(code)

internal/channels/telegram/format_test.go

@@ -182,3 +182,42 @@ func TestChunkHTML(t *testing.T) {
 		})
 	}
 }
+
+func TestMarkdownToTelegramHTML_URLsWithUnderscores(t *testing.T) {
+	tests := []struct {
+		name string
+		input string
+		want  string
+		deny  string
+	}{
+		{
+			name:  "bare URL with underscores not broken by italic",
+			input: "Check https://pre.glomotra.dev/uk/syngas_dailymail_2026_ai/?fname=James here",
+			want:  "https://pre.glomotra.dev/uk/syngas_dailymail_2026_ai/?fname=James",
+			deny:  "<i>",
+		},
+		{
+			name:  "URL without underscores unchanged",
+			input: "Visit https://example.com/path today",
+			want:  "https://example.com/path",
+		},
+		{
+			name:  "markdown link with underscored URL preserved",
+			input: "[Click](https://example.com/a_b_c)",
+			want:  `href="https://example.com/a_b_c"`,
+			deny:  "<i>",
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			got := markdownToTelegramHTML(tt.input)
+			if !strings.Contains(got, tt.want) {
+				t.Errorf("expected %q in output, got: %s", tt.want, got)
+			}
+			if tt.deny != "" && strings.Contains(got, tt.deny) {
+				t.Errorf("unexpected %q in output, got: %s", tt.deny, got)
+			}
+		})
+	}
+}