Use correct int return type for stream input operations

This makes it safe to compare results to -1 to indicate EOF,
regardless of whether char is considered signed or unsigned;
and so eliminates compiler warnings on platforms such as ARM.

Fixes bug #2058.

11 files changed, 126 insertions, 145 deletions

diff --git a/src/amount.cc b/src/amount.cc
index 4eacdd09..0527c979 100644
--- a/src/amount.cc
+++ b/src/amount.cc
@@ -970,13 +970,13 @@ namespace {
   void parse_quantity(std::istream& in, string& value)
   {
     char buf[256];
-    char c = peek_next_nonws(in);
+    int c = peek_next_nonws(in);
     int max = 255;
     char *p = buf;
     if (c == '-') {
       *p++ = c;
       max--;
-      in.get(c);
+      in.get();
     }
     READ_INTO(in, p, max, c,
               std::isdigit(c) || c == '.' || c == ',');
@@ -1005,10 +1005,10 @@ bool amount_t::parse(std::istream& in, const parse_flags_t& flags)
 
   commodity_t::flags_t comm_flags = COMMODITY_STYLE_DEFAULTS;
 
-  char c = peek_next_nonws(in);
+  int c = peek_next_nonws(in);
   if (c == '-') {
     negative = true;
-    in.get(c);
+    in.get();
     c = peek_next_nonws(in);
   }
 
diff --git a/src/annotate.cc b/src/annotate.cc
index 27261f57..e0fb0d08 100644
--- a/src/annotate.cc
+++ b/src/annotate.cc
@@ -86,33 +86,33 @@ void annotation_t::parse(std::istream& in)
       return;
 
     char buf[256];
-    char c = peek_next_nonws(in);
+    int c = peek_next_nonws(in);
     if (c == '{') {
       if (price)
         throw_(amount_error, _("Commodity specifies more than one price"));
 
-      in.get(c);
-      c = static_cast<char>(in.peek());
+      in.get();
+      c = in.peek();
       if (c == '{') {
-        in.get(c);
+        in.get();
         add_flags(ANNOTATION_PRICE_NOT_PER_UNIT);
       }
 
       c = peek_next_nonws(in);
       if (c == '=') {
-        in.get(c);
+        in.get();
         add_flags(ANNOTATION_PRICE_FIXATED);
       }
 
       READ_INTO(in, buf, 255, c, c != '}');
       if (c == '}') {
-        in.get(c);
+        in.get();
         if (has_flags(ANNOTATION_PRICE_NOT_PER_UNIT)) {
-          c = static_cast<char>(in.peek());
+          c = in.peek();
           if (c != '}')
             throw_(amount_error, _("Commodity lot price lacks double closing brace"));
           else
-            in.get(c);
+            in.get();
         }
       } else {
         throw_(amount_error, _("Commodity lot price lacks closing brace"));
@@ -128,18 +128,18 @@ void annotation_t::parse(std::istream& in)
       if (date)
         throw_(amount_error, _("Commodity specifies more than one date"));
 
-      in.get(c);
+      in.get();
       READ_INTO(in, buf, 255, c, c != ']');
       if (c == ']')
-        in.get(c);
+        in.get();
       else
         throw_(amount_error, _("Commodity date lacks closing bracket"));
 
       date = parse_date(buf);
     }
     else if (c == '(') {
-      in.get(c);
-      c = static_cast<char>(in.peek());
+      in.get();
+      c = in.peek();
       if (c == '@') {
         in.clear();
         in.seekg(pos, std::ios::beg);
@@ -150,13 +150,13 @@ void annotation_t::parse(std::istream& in)
           throw_(amount_error,
                  _("Commodity specifies more than one valuation expresion"));
 
-        in.get(c);
+        in.get();
         READ_INTO(in, buf, 255, c, c != ')');
         if (c == ')') {
-          in.get(c);
-          c = static_cast<char>(in.peek());
+          in.get();
+          c = in.peek();
           if (c == ')')
-            in.get(c);
+            in.get();
           else
             throw_(amount_error,
                    _("Commodity valuation expression lacks closing parentheses"));
@@ -172,7 +172,7 @@ void annotation_t::parse(std::istream& in)
 
         READ_INTO(in, buf, 255, c, c != ')');
         if (c == ')')
-          in.get(c);
+          in.get();
         else
           throw_(amount_error, _("Commodity tag lacks closing parenthesis"));
 
diff --git a/src/commodity.cc b/src/commodity.cc
index ebe388c8..29c04e59 100644
--- a/src/commodity.cc
+++ b/src/commodity.cc
@@ -297,12 +297,12 @@ void commodity_t::parse_symbol(std::istream& in, string& symbol)
   std::istream::pos_type pos = in.tellg();
 
   char buf[256];
-  char c = peek_next_nonws(in);
+  int c = peek_next_nonws(in);
   if (c == '"') {
-    in.get(c);
+    in.get();
     READ_INTO(in, buf, 255, c, c != '"');
     if (c == '"')
-      in.get(c);
+      in.get();
     else
       throw_(amount_error, _("Quoted commodity symbol lacks closing quote"));
   } else {
@@ -310,46 +310,45 @@ void commodity_t::parse_symbol(std::istream& in, string& symbol)
     while (_p - buf < 255 && in.good() && ! in.eof() && c != '\n') {
       std::size_t    bytes = 0;
       std::ptrdiff_t size  = _p - buf;
-      unsigned char  d     = static_cast<unsigned char>(c);
 
       // Check for the start of a UTF-8 multi-byte encoded string
-      if (d >= 192 && d <= 223 && size < 254)
+      if (c >= 192 && c <= 223 && size < 254)
         bytes = 2;
-      else if (d >= 224 && d <= 239 && size < 253)
+      else if (c >= 224 && c <= 239 && size < 253)
         bytes = 3;
-      else if (d >= 240 && d <= 247 && size < 252)
+      else if (c >= 240 && c <= 247 && size < 252)
         bytes = 4;
-      else if (d >= 248 && d <= 251 && size < 251)
+      else if (c >= 248 && c <= 251 && size < 251)
         bytes = 5;
-      else if (d >= 252 && d <= 253 && size < 250)
+      else if (c >= 252 && c <= 253 && size < 250)
         bytes = 6;
-      else if (d >= 254) // UTF-8 encoding error
+      else if (c >= 254) // UTF-8 encoding error
         break;
 
       if (bytes > 0) {          // we're looking at a UTF-8 encoding
         for (std::size_t i = 0; i < bytes; i++) {
-          in.get(c);
+          c = in.get();
           if (in.bad() || in.eof())
             throw_(amount_error, _("Invalid UTF-8 encoding for commodity name"));
           *_p++ = c;
         }
       }
-      else if (invalid_chars[static_cast<unsigned char>(c)]) {
+      else if (invalid_chars[c]) {
         break;
       }
       else {
-        in.get(c);
+        c = in.get();
         if (in.eof())
           break;
         if (c == '\\') {
-          in.get(c);
+          c = in.get();
           if (in.eof())
             throw_(amount_error, _("Backslash at end of commodity name"));
         }
         *_p++ = c;
       }
 
-      c = static_cast<char>(in.peek());
+      c = in.peek();
     }
     *_p = '\0';
 
diff --git a/src/query.cc b/src/query.cc
index 883bea40..0eb7dd85 100644
--- a/src/query.cc
+++ b/src/query.cc
@@ -220,37 +220,9 @@ test_ident:
   return token_t(token_t::UNKNOWN);
 }
 
-void query_t::lexer_t::token_t::unexpected()
+void query_t::lexer_t::token_t::expected(char wanted)
 {
-  kind_t prev_kind = kind;
-
-  kind = UNKNOWN;
-
-  switch (prev_kind) {
-  case END_REACHED:
-    throw_(parse_error, _("Unexpected end of expression"));
-  case TERM:
-    throw_(parse_error, _f("Unexpected string '%1%'") % *value);
-  default:
-    throw_(parse_error, _f("Unexpected token '%1%'") % symbol());
-  }
-}
-
-void query_t::lexer_t::token_t::expected(char wanted, char c)
-{
-  kind = UNKNOWN;
-
-  if (c == '\0' || c == -1) {
-    if (wanted == '\0' || wanted == -1)
-      throw_(parse_error, _("Unexpected end"));
-    else
-      throw_(parse_error, _f("Missing '%1%'") % wanted);
-  } else {
-    if (wanted == '\0' || wanted == -1)
-      throw_(parse_error, _f("Invalid char '%1%'") % c);
-    else
-      throw_(parse_error, _f("Invalid char '%1%' (wanted '%2%')") % c % wanted);
-  }
+  throw_(parse_error, _f("Missing '%1%'") % wanted);
 }
 
 expr_t::ptr_op_t
diff --git a/src/query.h b/src/query.h
index f7d48d32..3297cbd6 100644
--- a/src/query.h
+++ b/src/query.h
@@ -191,8 +191,7 @@ public:
 #endif
       }
 
-      void unexpected();
-      void expected(char wanted, char c = '\0');
+      void expected(char wanted);
     };
 
     token_t token_cache;
diff --git a/src/times.cc b/src/times.cc
index bab4e0de..127cdd69 100644
--- a/src/times.cc
+++ b/src/times.cc
@@ -1621,17 +1621,10 @@ void date_parser_t::lexer_t::token_t::unexpected()
 
 void date_parser_t::lexer_t::token_t::expected(char wanted, char c)
 {
-  if (c == '\0' || c == -1) {
-    if (wanted == '\0' || wanted == -1)
-      throw_(date_error, _("Unexpected end"));
-    else
-      throw_(date_error, _f("Missing '%1%'") % wanted);
-  } else {
-    if (wanted == '\0' || wanted == -1)
-      throw_(date_error, _f("Invalid char '%1%'") % c);
-    else
-      throw_(date_error, _f("Invalid char '%1%' (wanted '%2%')") % c % wanted);
-  }
+  if (wanted == '\0')
+    throw_(date_error, _f("Invalid char '%1%'") % c);
+  else
+    throw_(date_error, _f("Invalid char '%1%' (wanted '%2%')") % c % wanted);
 }
 
 namespace {
diff --git a/src/token.cc b/src/token.cc
index 76bf5106..63865829 100644
--- a/src/token.cc
+++ b/src/token.cc
@@ -38,7 +38,7 @@ namespace ledger {
 
 int expr_t::token_t::parse_reserved_word(std::istream& in)
 {
-  char c = static_cast<char>(in.peek());
+  int c = in.peek();
 
   if (c == 'a' || c == 'd' || c == 'e' || c == 'f' ||
       c == 'i' || c == 'o' || c == 'n' || c == 't') {
@@ -131,7 +131,8 @@ void expr_t::token_t::parse_ident(std::istream& in)
   kind   = IDENT;
   length = 0;
 
-  char c, buf[256];
+  int c;
+  char buf[256];
   READ_INTO_(in, buf, 255, c, length, std::isalnum(c) || c == '_');
 
   value.set_string(buf);
@@ -146,9 +147,9 @@ void expr_t::token_t::next(std::istream& in, const parse_flags_t& pflags)
   if (! in.good())
     throw_(parse_error, _("Input stream no longer valid"));
 
-  char c = peek_next_nonws(in);
+  int c = peek_next_nonws(in);
 
-  if (in.eof() || c == -1) {
+  if (in.eof()) {
     kind = TOK_EOF;
     return;
   }
@@ -162,10 +163,10 @@ void expr_t::token_t::next(std::istream& in, const parse_flags_t& pflags)
 
   switch (c) {
   case '&':
-    in.get(c);
-    c = static_cast<char>(in.peek());
+    in.get();
+    c = in.peek();
     if (c == '&') {
-      in.get(c);
+      in.get();
       kind = KW_AND;
       length = 2;
       break;
@@ -173,10 +174,10 @@ void expr_t::token_t::next(std::istream& in, const parse_flags_t& pflags)
     kind = KW_AND;
     break;
   case '|':
-    in.get(c);
-    c = static_cast<char>(in.peek());
+    in.get();
+    c = in.peek();
     if (c == '|') {
-      in.get(c);
+      in.get();
       kind = KW_OR;
       length = 2;
       break;
@@ -185,23 +186,23 @@ void expr_t::token_t::next(std::istream& in, const parse_flags_t& pflags)
     break;
 
   case '(':
-    in.get(c);
+    in.get();
     kind = LPAREN;
     break;
   case ')':
-    in.get(c);
+    in.get();
     kind = RPAREN;
     break;
 
   case '[': {
-    in.get(c);
+    in.get();
 
     char buf[256];
     READ_INTO_(in, buf, 255, c, length, c != ']');
     if (c != ']')
       expected(']', c);
 
-    in.get(c);
+    in.get();
     length++;
 
     date_interval_t timespan(buf);
@@ -222,7 +223,7 @@ void expr_t::token_t::next(std::istream& in, const parse_flags_t& pflags)
     READ_INTO_(in, buf, 4095, c, length, c != delim);
     if (c != delim)
       expected(delim, c);
-    in.get(c);
+    in.get();
     length++;
     kind = VALUE;
     value.set_string(buf);
@@ -230,10 +231,10 @@ void expr_t::token_t::next(std::istream& in, const parse_flags_t& pflags)
   }
 
   case '{': {
-    in.get(c);
+    in.get();
     amount_t temp;
     temp.parse(in, PARSE_NO_MIGRATE);
-    in.get(c);
+    c = in.get();
     if (c != '}')
       expected('}', c);
     length++;
@@ -243,10 +244,10 @@ void expr_t::token_t::next(std::istream& in, const parse_flags_t& pflags)
   }
 
   case '!':
-    in.get(c);
-    c = static_cast<char>(in.peek());
+    in.get();
+    c = in.peek();
     if (c == '=') {
-      in.get(c);
+      in.get();
       symbol[1] = c;
       symbol[2] = '\0';
       kind = NEQUAL;
@@ -254,7 +255,7 @@ void expr_t::token_t::next(std::istream& in, const parse_flags_t& pflags)
       break;
     }
     else if (c == '~') {
-      in.get(c);
+      in.get();
       symbol[1] = c;
       symbol[2] = '\0';
       kind = NMATCH;
@@ -265,10 +266,10 @@ void expr_t::token_t::next(std::istream& in, const parse_flags_t& pflags)
     break;
 
   case '-':
-    in.get(c);
-    c = static_cast<char>(in.peek());
+    in.get();
+    c = in.peek();
     if (c == '>') {
-      in.get(c);
+      in.get();
       symbol[1] = c;
       symbol[2] = '\0';
       kind = ARROW;
@@ -278,27 +279,26 @@ void expr_t::token_t::next(std::istream& in, const parse_flags_t& pflags)
     kind = MINUS;
     break;
   case '+':
-    in.get(c);
+    in.get();
     kind = PLUS;
     break;
 
   case '*':
-    in.get(c);
+    in.get();
     kind = STAR;
     break;
 
   case '?':
-    in.get(c);
+    in.get();
     kind = QUERY;
     break;
   case ':':
-    in.get(c);
-    c = static_cast<char>(in.peek());
+    in.get();
     kind = COLON;
     break;
 
   case '/': {
-    in.get(c);
+    in.get();
     if (pflags.has_flags(PARSE_OP_CONTEXT)) { // operator context
       kind = SLASH;
     } else {                    // terminal context
@@ -307,7 +307,7 @@ void expr_t::token_t::next(std::istream& in, const parse_flags_t& pflags)
       READ_INTO_(in, buf, 4095, c, length, c != '/');
       if (c != '/')
         expected('/', c);
-      in.get(c);
+      in.get();
       length++;
 
       kind = VALUE;
@@ -317,10 +317,10 @@ void expr_t::token_t::next(std::istream& in, const parse_flags_t& pflags)
   }
 
   case '=':
-    in.get(c);
-    c = static_cast<char>(in.peek());
+    in.get();
+    c = in.peek();
     if (c == '~') {
-      in.get(c);
+      in.get();
       symbol[1] = c;
       symbol[2] = '\0';
       kind = MATCH;
@@ -328,7 +328,7 @@ void expr_t::token_t::next(std::istream& in, const parse_flags_t& pflags)
       break;
     }
     else if (c == '=') {
-      in.get(c);
+      in.get();
       symbol[1] = c;
       symbol[2] = '\0';
       kind = EQUAL;
@@ -339,10 +339,9 @@ void expr_t::token_t::next(std::istream& in, const parse_flags_t& pflags)
     break;
 
   case '<':
-    in.get(c);
-    if (static_cast<char>(in.peek()) == '=') {
-      in.get(c);
-      symbol[1] = c;
+    in.get();
+    if (in.peek() == '=') {
+      symbol[1] = in.get();
       symbol[2] = '\0';
       kind = LESSEQ;
       length = 2;
@@ -352,10 +351,9 @@ void expr_t::token_t::next(std::istream& in, const parse_flags_t& pflags)
     break;
 
   case '>':
-    in.get(c);
-    if (static_cast<char>(in.peek()) == '=') {
-      in.get(c);
-      symbol[1] = c;
+    in.get();
+    if (in.peek() == '=') {
+      symbol[1] = in.get();
       symbol[2] = '\0';
       kind = GREATEREQ;
       length = 2;
@@ -365,17 +363,17 @@ void expr_t::token_t::next(std::istream& in, const parse_flags_t& pflags)
     break;
 
   case '.':
-    in.get(c);
+    in.get();
     kind = DOT;
     break;
 
   case ',':
-    in.get(c);
+    in.get();
     kind = COMMA;
     break;
 
   case ';':
-    in.get(c);
+    in.get();
     kind = SEMI;
     break;
 
@@ -420,7 +418,7 @@ void expr_t::token_t::next(std::istream& in, const parse_flags_t& pflags)
         if (in.fail() || ! in.good())
           throw_(parse_error, _("Failed to reset input stream"));
 
-        c = static_cast<char>(in.peek());
+        c = in.peek();
         if (c != -1) {
           if (! std::isalpha(c) && c != '_')
             expected('\0', c);
@@ -503,19 +501,20 @@ void expr_t::token_t::unexpected(const char wanted)
   }
 }
 
-void expr_t::token_t::expected(const char wanted, char c)
+void expr_t::token_t::expected(const char wanted, const int c)
 {
-  if (c == '\0' || c == -1) {
-    if (wanted == '\0' || wanted == -1)
+  if (c == -1) {
+    if (wanted == '\0')
       throw_(parse_error, _("Unexpected end"));
     else
       throw_(parse_error, _f("Missing '%1%'") % wanted);
   } else {
-    if (wanted == '\0' || wanted == -1)
-      throw_(parse_error, _f("Invalid char '%1%'") % c);
+    char ch = c;
+    if (wanted == '\0')
+      throw_(parse_error, _f("Invalid char '%1%'") % ch);
     else
       throw_(parse_error,
-             _f("Invalid char '%1%' (wanted '%2%')") % c % wanted);
+             _f("Invalid char '%1%' (wanted '%2%')") % ch % wanted);
   }
 }
 
diff --git a/src/token.h b/src/token.h
index 5d9f1fac..db592475 100644
--- a/src/token.h
+++ b/src/token.h
@@ -126,7 +126,7 @@ struct expr_t::token_t : public noncopyable
   void next(std::istream& in, const parse_flags_t& flags);
   void rewind(std::istream& in);
   void unexpected(const char wanted = '\0');
-  void expected(const char wanted, const char c = '\0');
+  void expected(const char wanted, const int c);
   void expected(const kind_t wanted);
 };
 
diff --git a/src/utils.h b/src/utils.h
index c9146dd7..9af1251a 100644
--- a/src/utils.h
+++ b/src/utils.h
@@ -540,25 +540,25 @@ inline char * next_element(char * buf, bool variable = false) {
   return NULL;
 }
 
-inline char peek_next_nonws(std::istream& in) {
-  char c = static_cast<char>(in.peek());
+inline int peek_next_nonws(std::istream& in) {
+  int c = in.peek();
   while (in.good() && ! in.eof() && std::isspace(c)) {
-    in.get(c);
-    c = static_cast<char>(in.peek());
+    in.get();
+    c = in.peek();
   }
   return c;
 }
 
 #define READ_INTO(str, targ, size, var, cond) {         \
     char * _p = targ;                                   \
-    var = static_cast<char>(str.peek());                \
+    var = str.peek();                                   \
     while (str.good() && ! str.eof() && var != '\n' &&  \
            (cond) && _p - targ < size) {                \
-      str.get(var);                                     \
+      var = str.get();                                  \
       if (str.eof())                                    \
         break;                                          \
       if (var == '\\') {                                \
-        str.get(var);                                   \
+        var = str.get();                                \
         if (in.eof())                                   \
           break;                                        \
         switch (var) {                                  \
@@ -572,22 +572,22 @@ inline char peek_next_nonws(std::istream& in) {
         }                                               \
       }                                                 \
       *_p++ = var;                                      \
-      var = static_cast<char>(str.peek());              \
+      var = str.peek();                                 \
     }                                                   \
     *_p = '\0';                                         \
   }
 
 #define READ_INTO_(str, targ, size, var, idx, cond) {   \
     char * _p = targ;                                   \
-    var = static_cast<char>(str.peek());                \
+    var = str.peek();                                   \
     while (str.good() && ! str.eof() && var != '\n' &&  \
            (cond) && _p - targ < size) {                \
-      str.get(var);                                     \
+      var = str.get();                                  \
       if (str.eof())                                    \
         break;                                          \
       idx++;                                            \
       if (var == '\\') {                                \
-        str.get(var);                                   \
+        var = str.get();                                \
         if (in.eof())                                   \
           break;                                        \
         switch (var) {                                  \
@@ -602,7 +602,7 @@ inline char peek_next_nonws(std::istream& in) {
         idx++;                                          \
       }                                                 \
       *_p++ = var;                                      \
-      var = static_cast<char>(str.peek());              \
+      var = str.peek();                                 \
     }                                                   \
     *_p = '\0';                                         \
   }
diff --git a/test/regress/2058_1.test b/test/regress/2058_1.test
new file mode 100644
index 00000000..06043007
--- /dev/null
+++ b/test/regress/2058_1.test
@@ -0,0 +1,8 @@
+2021/1/2 Test
+    A  $1.00
+    B
+
+test -p 'last %^@' bal -> 1
+__ERROR__
+Error: Invalid char '%'
+end test
diff --git a/test/regress/2058_2.test b/test/regress/2058_2.test
new file mode 100644
index 00000000..3c6c2a7e
--- /dev/null
+++ b/test/regress/2058_2.test
@@ -0,0 +1,11 @@
+2021/1/2 Test
+    A  $1.00
+    B
+
+test --limit 'date>=[2020/12/31' register A -> 1
+__ERROR__
+While parsing value expression:
+  (date>=[2020/12/31)&((account =~ /A/))
+
+Error: Missing ']'
+end test