From 134a1ac3372fe1eae6bc5c6acd12666c17e82696 Mon Sep 17 00:00:00 2001
From: Hiltjo Posthuma <hiltjo@codemadness.org>
Date: Sat, 31 Oct 2020 19:51:17 +0100
Subject: sfeed_web: improve parsing a <link> if it has no type attribute
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

This happens because the previous link type is not reset when a <link> tag
starts again, but it is reset when a type attribute starts.

Found on the spanish newspaper site: elpais.com

Input:

	<link rel="alternate" href="https://feeds.elpais.com/mrss-s/pages/ep/site/elpais.com/portada" type="application/rss+xml" title="RSS de la portada de El País"/>
	<link rel="canonical" href="https://elpais.com"/>

Would print (second line is incorrect).

	https://feeds.elpais.com/mrss-s/pages/ep/site/elpais.com/portada	application/rss+xml
	https://elpais.com/	application/rss+xml

Now prints:

	https://feeds.elpais.com/mrss-s/pages/ep/site/elpais.com/portada	application/rss+xml

Fix: reset it also at the start of a <link> tag in this case (for <base href />
it is still not wanted).
---
 sfeed_web.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/sfeed_web.c b/sfeed_web.c
index e0ab874..6d547a7 100644
--- a/sfeed_web.c
+++ b/sfeed_web.c
@@ -32,6 +32,7 @@ xmltagstart(XMLParser *p, const char *t, size_t tl)
 	} else if (!strcasecmp(t, "link")) {
 		islinktag = 1;
 		linkhref[0] = '\0';
+		linktype[0] = '\0';
 	}
 }
 
-- 
cgit v1.2.3