From fb19f59f93146405e766e26cce58d9a16eaf0d6b Mon Sep 17 00:00:00 2001
From: PyYoshi <yohihiro.m@gmail.com>
Date: Sat, 7 Jul 2012 11:31:10 +0900
Subject: [PATCH] fixed "sometimes, to output invalid confidence value" bug.

---
 src/cchardet/cchardet.pyx                                  | 3 +--
 .../universalchardet/src/base/nsUniversalDetector.cpp      | 7 ++++++-
 2 files changed, 7 insertions(+), 3 deletions(-)

diff --git a/src/cchardet/cchardet.pyx b/src/cchardet/cchardet.pyx
index 02cb15b..c506133 100644
--- a/src/cchardet/cchardet.pyx
+++ b/src/cchardet/cchardet.pyx
@@ -33,11 +33,10 @@ def detect(char *msg):
         return csd_close(csd)
 
 def detect_with_confidence(char *msg):
-    # TODO: fix "sometimes, to output invalid confidence value" bug.
     cdef csd_t csd = csd_open()
     cdef int length = strlen(msg)
     cdef int result = csd_consider(csd, msg, length)
-    cdef float confidence
+    cdef float confidence = 0.0
     cdef const_char_ptr detected_charset
     # ref: charsetdetect.cpp
     if result == -1: # Error, signal with a negative number
diff --git a/src/ext/libcharsetdetect/mozilla/extensions/universalchardet/src/base/nsUniversalDetector.cpp b/src/ext/libcharsetdetect/mozilla/extensions/universalchardet/src/base/nsUniversalDetector.cpp
index d9e6bc4..8993aae 100644
--- a/src/ext/libcharsetdetect/mozilla/extensions/universalchardet/src/base/nsUniversalDetector.cpp
+++ b/src/ext/libcharsetdetect/mozilla/extensions/universalchardet/src/base/nsUniversalDetector.cpp
@@ -45,6 +45,8 @@
 #include "nsEscCharsetProber.h"
 #include "nsLatin1Prober.h"
 
+#include <iostream>
+
 nsUniversalDetector::nsUniversalDetector(PRUint32 aLanguageFilter)
 {
   mDone = PR_FALSE;
@@ -292,6 +294,8 @@ void nsUniversalDetector::DataEnd2(float *confidence)
   {
     mDone = PR_TRUE;
     Report(mDetectedCharset);
+    // no confidence
+    *confidence = (float)0.99;
     return;
   }
 
@@ -316,9 +320,10 @@ void nsUniversalDetector::DataEnd2(float *confidence)
         }
       }
       //do not report anything because we are not confident of it, that's in fact a negative answer
-      if (maxProberConfidence > MINIMUM_THRESHOLD)
+      if (maxProberConfidence > MINIMUM_THRESHOLD){
         Report(mCharSetProbers[maxProber]->GetCharSetName());
         *confidence = maxProberConfidence;
+      }
     }
     break;
   case eEscAscii: