Properly handle UTF-8 characters in commodity strings.

9 files changed, 112 insertions, 100 deletions

diff --git a/src/amount.cc b/src/amount.cc
index 3fb8ddaf..ffeaeea6 100644
--- a/src/amount.cc
+++ b/src/amount.cc
@@ -31,6 +31,7 @@
 
 #include "amount.h"
 #include "commodity.h"
+#include "unistring.h"
 
 namespace ledger {
 
@@ -1001,6 +1002,20 @@ void amount_t::print(std::ostream& _out) const
   _out << out.str();
 }
 
+void amount_t::right_justify(std::ostream& out, int width) const
+{
+  std::ostringstream buf;
+  buf << *this;
+
+  unistring temp(buf.str());
+
+  int spacing = width - int(temp.length());
+  while (spacing-- > 0)
+    out << ' ';
+
+  out << temp.extract();
+}
+
 bool amount_t::valid() const
 {
   if (quantity) {
diff --git a/src/amount.h b/src/amount.h
index 0943a4c0..a9077a4b 100644
--- a/src/amount.h
+++ b/src/amount.h
@@ -640,6 +640,7 @@ public:
       of its commodity's display precision.
   */
   void print(std::ostream& out) const;
+  void right_justify(std::ostream& out, int width) const;
 
   /*@}*/
 
diff --git a/src/balance.cc b/src/balance.cc
index c9ca148d..7f8ed572 100644
--- a/src/balance.cc
+++ b/src/balance.cc
@@ -240,17 +240,11 @@ void balance_t::print(std::ostream& out,
       first = false;
       width = first_width;
     }
-
-    out.width(width);
-    out.fill(' ');
-    out << std::right << *amount;
+    amount->right_justify(out, width);
   }
 
-  if (first) {
-    out.width(first_width);
-    out.fill(' ');
-    out << std::right << "0";
-  }
+  if (first)
+    amount_t(0L).right_justify(out, first_width);
 }
 
 } // namespace ledger
diff --git a/src/commodity.cc b/src/commodity.cc
index 5169e3af..2dfba880 100644
--- a/src/commodity.cc
+++ b/src/commodity.cc
@@ -571,7 +571,55 @@ void commodity_t::parse_symbol(std::istream& in, string& symbol)
     else
       throw_(amount_error, "Quoted commodity symbol lacks closing quote");
   } else {
-    READ_INTO(in, buf, 255, c, ! invalid_chars[static_cast<unsigned char>(c)]);
+    char * _p = buf;
+    c = in.peek();
+    while (_p - buf < 255 && in.good() && ! in.eof() && c != '\n') {
+      int bytes = 0;
+      int size  = _p - buf;
+
+      unsigned char d = c;
+
+      // Check for the start of a UTF-8 multi-byte encoded string
+      if (d >= 192 && d <= 223 && size < 254)
+	bytes = 2;
+      else if (d >= 224 && d <= 239 && size < 253)
+	bytes = 3;
+      else if (d >= 240 && d <= 247 && size < 252)
+	bytes = 4;
+      else if (d >= 248 && d <= 251 && size < 251)
+	bytes = 5;
+      else if (d >= 252 && d <= 253 && size < 250)
+	bytes = 6;
+      else if (d >= 254) // UTF-8 encoding error
+	break;
+
+      if (bytes > 0) {		// we're looking at a UTF-8 encoding
+	for (int i = 0; i < bytes; i++) {
+	  in.get(c);
+	  if (in.bad() || in.eof())
+	    break;
+	  *_p++ = c;
+	}
+      }
+      else if (invalid_chars[static_cast<unsigned char>(c)]) {
+	break;
+      }
+      else {
+	in.get(c);
+	if (in.eof())
+	  break;
+	if (c == '\\') {
+	  in.get(c);
+	  if (in.eof())
+	    break;
+	}
+	*_p++ = c;
+      }
+
+      c = in.peek();
+    }
+    *_p = '\0';
+
     if (is_reserved_token(buf))
       buf[0] = '\0';
   }
diff --git a/src/format.h b/src/format.h
index 49350104..d1716dbf 100644
--- a/src/format.h
+++ b/src/format.h
@@ -47,58 +47,12 @@
 #define _FORMAT_H
 
 #include "expr.h"
+#include "unistring.h"
 
 namespace ledger {
 
 DECLARE_EXCEPTION(format_error, std::runtime_error);
 
-#if defined(SUPPORT_UNICODE)
-
-/**
- * @class unistring
- *
- * @brief Abstract working with UTF-32 encoded Unicode strings
- *
- * The input to the string is a UTF8 encoded ledger::string, which can
- * then have its true length be taken, or characters extracted.
- */
-class unistring
-{
-  std::vector<uint32_t> utf32chars;
-
-public:
-  unistring(const string& input)
-  {
-    TRACE_CTOR(unistring, "");
-
-    const char * p   = input.c_str();
-    std::size_t	 len = input.length();
-
-    VERIFY(utf8::is_valid(p, p + len));
-
-    utf8::utf8to32(p, p + len, std::back_inserter(utf32chars));
-  }
-  ~unistring() {
-    TRACE_DTOR(unistring);
-  }
-
-  std::size_t length() const {
-    return utf32chars.size();
-  }
-
-  string extract(const std::size_t begin = 0,
-		 const std::size_t len   = 0) const
-  {
-    string utf8result;
-    utf8::utf32to8(utf32chars.begin() + begin,
-		   utf32chars.begin() + begin + (len ? len : length()),
-		   std::back_inserter(utf8result));
-    return utf8result;
-  }
-};
-
-#endif // SUPPORT_UNICODE
-
 class report_t;
 
 /**
diff --git a/src/report.cc b/src/report.cc
index e319c8aa..7d8353d9 100644
--- a/src/report.cc
+++ b/src/report.cc
@@ -68,7 +68,8 @@ report_t::report_t(session_t& _session) : session(_session)
     "  %12(amount)%(comment | \"\")\n%/\n");
 
   HANDLER(balance_format_).on(
-    "%20(strip(display_total))  %(depth_spacer)%-(partial_account)\n");
+    "%20(print_balance(strip(display_total), 20))"
+    "  %(depth_spacer)%-(partial_account)\n");
 
   HANDLER(equity_format_).on("\n%D %Y%C%P\n%/    %-34W  %12t\n");
 
@@ -191,7 +192,7 @@ value_t report_t::fn_print_balance(call_scope_t& args)
 
   std::ostringstream out;
   args[0].strip_annotations(what_to_keep())
-    .print(out, *first_width, *latter_width,
+    .print(out, *first_width, latter_width ? *latter_width : -1,
 	   HANDLED(date_format_) ?
 	   HANDLER(date_format_).str() : optional<string>());
 
diff --git a/src/system.hh b/src/system.hh
index bffd041c..96b574e5 100644
--- a/src/system.hh
+++ b/src/system.hh
@@ -138,10 +138,7 @@ typedef std::ostream::pos_type ostream_pos_type;
 #include <gmp.h>
 #include <mpfr.h>
 #include "sha1.h"
-#define SUPPORT_UNICODE 1
-#if defined(SUPPORT_UNICODE)
 #include "utf8.h"
-#endif
 
 #ifdef HAVE_LIBEDIT
 #include <editline/readline.h>
diff --git a/src/utils.h b/src/utils.h
index d4011442..93cb9a4a 100644
--- a/src/utils.h
+++ b/src/utils.h
@@ -546,50 +546,52 @@ inline char * next_element(char * buf, bool variable = false) {
 
 inline char peek_next_nonws(std::istream& in) {
   char c = in.peek();
-  while (! in.eof() && std::isspace(c)) {
+  while (in.good() && ! in.eof() && std::isspace(c)) {
     in.get(c);
     c = in.peek();
   }
   return c;
 }
 
-#define READ_INTO(str, targ, size, var, cond) {				\
-  char * _p = targ;							\
-  var = str.peek();							\
-  while (! str.eof() && var != '\n' && (cond) && _p - targ < size) {	\
-    str.get(var);							\
-    if (str.eof())							\
-      break;								\
-    if (var == '\\') {							\
-      str.get(var);							\
-      if (in.eof())							\
-	break;								\
-    }									\
-    *_p++ = var;							\
-    var = str.peek();							\
-  }									\
-  *_p = '\0';								\
-}
+#define READ_INTO(str, targ, size, var, cond) {		\
+    char * _p = targ;					\
+    var = str.peek();					\
+    while (str.good() && ! str.eof() && var != '\n' &&	\
+	   (cond) && _p - targ < size) {		\
+      str.get(var);					\
+      if (str.eof())					\
+	break;						\
+      if (var == '\\') {				\
+	str.get(var);					\
+	if (in.eof())					\
+	  break;					\
+      }							\
+      *_p++ = var;					\
+      var = str.peek();					\
+    }							\
+    *_p = '\0';						\
+  }
 
-#define READ_INTO_(str, targ, size, var, idx, cond) {			\
-  char * _p = targ;							\
-  var = str.peek();							\
-  while (! str.eof() && var != '\n' && (cond) && _p - targ < size) {	\
-    str.get(var);							\
-    if (str.eof())							\
-      break;								\
-    idx++;								\
-    if (var == '\\') {							\
-      str.get(var);							\
-      if (in.eof())							\
-	break;								\
-      idx++;								\
-    }									\
-    *_p++ = var;							\
-    var = str.peek();							\
-  }									\
-  *_p = '\0';								\
-}
+#define READ_INTO_(str, targ, size, var, idx, cond) {	\
+    char * _p = targ;					\
+    var = str.peek();					\
+    while (str.good() && ! str.eof() && var != '\n' &&	\
+	   (cond) && _p - targ < size) {		\
+      str.get(var);					\
+      if (str.eof())					\
+	break;						\
+      idx++;						\
+      if (var == '\\') {				\
+	str.get(var);					\
+	if (in.eof())					\
+	  break;					\
+	idx++;						\
+      }							\
+      *_p++ = var;					\
+      var = str.peek();					\
+    }							\
+    *_p = '\0';						\
+  }
 
 } // namespace ledger
 
diff --git a/src/value.cc b/src/value.cc
index 482dca21..a39eada2 100644
--- a/src/value.cc
+++ b/src/value.cc
@@ -1271,7 +1271,7 @@ void value_t::print(std::ostream&           out,
     break;
 
   case AMOUNT:
-    out << as_amount();
+    as_amount().right_justify(out, first_width);
     break;
 
   case STRING: