[PATCH] Add id3v2 encoding detect via enca

[Heiher]

---
 configure.ac               |   13 +++++++++++++
 gst/id3demux/Makefile.am   |    4 ++--
 gst/id3demux/id3v2frames.c |   43 +++++++++++++++++++++++++++++++++++++++++--
 3 files changed, 56 insertions(+), 4 deletions(-)

diff --git a/configure.ac b/configure.ac
index c7757bb..71108f2 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1004,6 +1004,18 @@ AG_GST_CHECK_FEATURE(ZLIB, [zlib support for
id3demux/qtdemux/matroska],, [
   ])
 ])

+dnl *** id3demux prefer to have enca ***
+translit(dnm, m, l) AM_CONDITIONAL(USE_ENCA, true)
+AG_GST_CHECK_FEATURE(ENCA, [enca support for id3demux],, [
+  AG_GST_CHECK_LIBHEADER(ENCA,
+    enca, enca_analyser_alloc,, enca.h, [
+    HAVE_ENCA="yes"
+    ENCA_CFLAGS=""
+    ENCA_LIBS="-lenca"
+    AC_SUBST(ENCA_LIBS)
+  ])
+])
+
 dnl *** matroska prefers to have bz2 ***
 translit(dnm, m, l) AM_CONDITIONAL(USE_BZ2, true)
 AG_GST_CHECK_FEATURE(BZ2, [bz2 library for matroska ],, [
@@ -1049,6 +1061,7 @@ AM_CONDITIONAL(USE_X, false)
 AM_CONDITIONAL(USE_XSHM, false)
 AM_CONDITIONAL(USE_XVIDEO, false)
 AM_CONDITIONAL(USE_ZLIB, false)
+AM_CONDITIONAL(USE_ENCA, false)

 fi dnl of EXT plugins

diff --git a/gst/id3demux/Makefile.am b/gst/id3demux/Makefile.am
index 3f00443..f02f684 100644
--- a/gst/id3demux/Makefile.am
+++ b/gst/id3demux/Makefile.am
@@ -1,9 +1,9 @@
 plugin_LTLIBRARIES = libgstid3demux.la

 libgstid3demux_la_SOURCES = gstid3demux.c id3tags.c id3v2frames.c
-libgstid3demux_la_CFLAGS = $(GST_PLUGINS_BASE_CFLAGS) $(GST_BASE_CFLAGS)
$(GST_CFLAGS) 
+libgstid3demux_la_CFLAGS = $(GST_PLUGINS_BASE_CFLAGS) $(GST_BASE_CFLAGS)
$(GST_CFLAGS) $(ENCA_CFLAGS)
 libgstid3demux_la_LIBADD = $(GST_PLUGINS_BASE_LIBS) -lgsttag- at GST_MAJORMINOR@
\
-    -lgstpbutils- at GST_MAJORMINOR@ $(GST_BASE_LIBS) $(ZLIB_LIBS)
+    -lgstpbutils- at GST_MAJORMINOR@ $(GST_BASE_LIBS) $(ZLIB_LIBS) $(ENCA_LIBS)
 libgstid3demux_la_LDFLAGS = $(GST_PLUGIN_LDFLAGS)
 libgstid3demux_la_LIBTOOLFLAGS = --tag=disable-static

diff --git a/gst/id3demux/id3v2frames.c b/gst/id3demux/id3v2frames.c
index e51bbb7..24b242a 100644
--- a/gst/id3demux/id3v2frames.c
+++ b/gst/id3demux/id3v2frames.c
@@ -33,6 +33,10 @@
 #include <zlib.h>
 #endif

+#ifdef HAVE_ENCA
+#include <enca.h>
+#endif
+
 #include "id3tags.h"

 GST_DEBUG_CATEGORY_EXTERN (id3demux_debug);
@@ -962,6 +966,10 @@ string_utf8_dup (const gchar * start, const guint size)
   const gchar *env;
   gsize bytes_read;
   gchar *utf8;
+#ifdef HAVE_ENCA
+  EncaAnalyser eanalyser;
+  gchar langname[3];
+#endif

   /* Should we try the charsets specified
    * via environment variables FIRST ? */
@@ -995,6 +1003,8 @@ string_utf8_dup (const gchar * start, const guint size)
       }
     }
   }
+
+#ifndef HAVE_ENCA
   /* Try current locale (if not UTF-8) */
   if (!g_get_charset (&env)) {
     if ((utf8 = g_locale_to_utf8 (start, size, &bytes_read, NULL, NULL))) {
@@ -1015,11 +1025,40 @@ string_utf8_dup (const gchar * start, const guint size)

   g_free (utf8);
   return NULL;
+#else /* HAVE_ENCA */
+  env = g_getenv("LANG");
+  if(NULL == env) {
+      return NULL;
+  }
+  langname[0] = env[0];
+  langname[1] = env[1];
+  langname[2] = 0;
+
+  eanalyser = enca_analyser_alloc(langname);
+  if(eanalyser) {
+      EncaEncoding encoding;
+      const gchar * charset;
+
+      encoding = enca_analyse(eanalyser, (unsigned char*)start, size);
+      if(ENCA_CS_UNKNOWN != encoding.charset) {
+          charset = enca_charset_name(encoding.charset,
ENCA_NAME_STYLE_ICONV);
+
+          utf8 = g_convert (start, size, "UTF-8", charset, &bytes_read, NULL,
NULL);
+          if (utf8 != NULL && bytes_read == size) {
+              enca_analyser_free(eanalyser);
+              goto beach;
+          }
+      }
+
+      enca_analyser_free(eanalyser);
+  }

-beach:
+  g_free(utf8);
+  return NULL;
+#endif

+beach:
   g_strchomp (utf8);
-
   return (utf8);
 }

-- 
1.7.4.4

-- 
Configure bugmail: https://bugzilla.gnome.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.
You are the assignee for the bug.