From: Saint Xavier <wget@sxav.eu>
Date: Thu, 19 Jun 2008 20:33:38 +0000 (+0200)
Subject: Add "content-type" meta tag parsing for retrieving HTML page encoding.
X-Git-Tag: v1.13~338^2~7^2~6^2~13^2~26
X-Git-Url: http://sjero.net/git/?p=wget;a=commitdiff_plain;h=13fec855660ee55c43f64fe47fbc284f35ca6e6e

Add "content-type" meta tag parsing for retrieving HTML page encoding.
---

diff --git a/src/ChangeLog b/src/ChangeLog
index ac27e15a..e30990b0 100644
--- a/src/ChangeLog
+++ b/src/ChangeLog
@@ -1,3 +1,11 @@
+2008-06-19  Xavier Saint  <wget@sxav.eu>
+
+	* html-url.c : Add "content-type" meta tag parsing for
+	retrieving page encoding.
+
+	* iri.h : Make no-op version of parse_charset() return
+	NULL.
+
 2008-06-14  Xavier Saint  <wget@sxav.eu>
 
  	* iri.c, iri.h : New files.
diff --git a/src/html-url.c b/src/html-url.c
index 0eb66506..9b515432 100644
--- a/src/html-url.c
+++ b/src/html-url.c
@@ -42,6 +42,7 @@ as that of the covered work.  */
 #include "hash.h"
 #include "convert.h"
 #include "recur.h"              /* declaration of get_urls_html */
+#include "iri.h"
 
 struct map_context;
 
@@ -534,6 +535,25 @@ tag_handle_meta (int tagid, struct taginfo *tag, struct map_context *ctx)
           entry->link_expect_html = 1;
         }
     }
+  else if (http_equiv && 0 == strcasecmp (http_equiv, "content-type"))
+    {
+      /* Handle stuff like:
+         <meta http-equiv="Content-Type" content="text/html; charset=CHARSET"> */
+
+      char *mcharset;
+      char *content = find_attr (tag, "content", NULL);
+      if (!content)
+        return;
+
+      mcharset = parse_charset (content);
+      if (!mcharset)
+        return;
+
+      logprintf (LOG_VERBOSE, "Meta tag charset : %s\n", quote (mcharset));
+
+      /* sXXXav: Not used yet */
+      xfree (mcharset);
+    }
   else if (name && 0 == strcasecmp (name, "robots"))
     {
       /* Handle stuff like:
diff --git a/src/iri.h b/src/iri.h
index d135e868..2ac7d5e7 100644
--- a/src/iri.h
+++ b/src/iri.h
@@ -37,7 +37,7 @@ char *parse_charset (char *str);
 
 #else /* ENABLE_IRI */
 
-#define parse_charset(str)	/* no-op */
+#define parse_charset(str)	NULL
 
 #endif /* ENABLE_IRI */
 #endif /* IRI_H */