Improve error handling in Python parser

diff --git a/bootstrap/scan-gram.l b/bootstrap/scan-gram.l
index 9419da7..84aee8a 100644 (file)
--- a/bootstrap/scan-gram.l
+++ b/bootstrap/scan-gram.l
@@ -248,7 +248,7 @@ eqopt    ([[:space:]]*=)?
     unput('%token-table')
   }
 
-  "%"{id}                      state.complain(y_tab.yylloc, state.complaint, 'invalid directive: {0:s}'.format(quote(yytext)))
+  "%"{id}                      state.complain(y_tab.yylloc, state.complaint, 'invalid directive: {0:s}'.format(state.quote(yytext)))
 
   "="                          return y_tab.EQUAL
   "|"                          return y_tab.PIPE
@@ -278,7 +278,7 @@ eqopt    ([[:space:]]*=)?
 
   /* Identifiers may not start with a digit.  Yet, don't silently
      accept "1FOO" as "1 FOO".  */
-  {int}{id}                    state.complain(y_tab.yylloc, state.complaint, 'invalid identifier: %s', quote(yytext))
+  {int}{id}                    state.complain(y_tab.yylloc, state.complaint, 'invalid identifier: {0:s}'.format(state.quote(yytext)))
 
   /* Characters.  */
   "'" {
@@ -394,7 +394,7 @@ eqopt    ([[:space:]]*=)?
     BEGIN(SC_BRACKETED_ID)
   }
 
-  [^\[%A-Za-z0-9_<>{}\"\'*;|=/, \f\n\t\v]+|. state.complain(y_tab.yylloc, state.complaint, '%s: %s', ngettext('invalid character', 'invalid characters', gram_leng), quote_mem(yytext, gram_leng))
+  [^\[%A-Za-z0-9_<>{}\"\'*;|=/, \f\n\t\v]+|. state.complain(y_tab.yylloc, state.complaint, '{0:s}: {1:s}'.format('invalid character' if len(yytext) == 1 else 'invalid characters', state.quote(yytext)))
 
   <<EOF>> {
     y_tab.yylloc.start = scanner_cursor.copy()
@@ -463,7 +463,7 @@ eqopt    ([[:space:]]*=)?
   {id} {
     global bracketed_id_str, bracketed_id_loc
     if bracketed_id_str is not None:
-      state.complain(y_tab.yylloc, state.complaint, 'unexpected identifier in bracketed name: %s', quote(yytext))
+      state.complain(y_tab.yylloc, state.complaint, 'unexpected identifier in bracketed name: {0:s}'.format(state.quote(yytext)))
     else:
       bracketed_id_str = yytext
       bracketed_id_loc = y_tab.yylloc
@@ -481,7 +481,7 @@ eqopt    ([[:space:]]*=)?
       state.complain(y_tab.yylloc, state.complaint, 'an identifier expected')
   }
 
-  [^\].A-Za-z0-9_/ \f\n\t\v]+|.        state.complain(y_tab.yylloc, state.complaint, '{0:s}: {1:s}'.format('invalid character in bracketed name' if gram_leng == 1 else 'invalid characters in bracketed name', quote_mem(yytext, gram_leng)))
+  [^\].A-Za-z0-9_/ \f\n\t\v]+|.        state.complain(y_tab.yylloc, state.complaint, '{0:s}: {1:s}'.format('invalid character in bracketed name' if len(yytext) == 1 else 'invalid characters in bracketed name', state.quote(yytext)))
 
   <<EOF>> {
     BEGIN(bracketed_id_context_state)
@@ -636,7 +636,7 @@ eqopt    ([[:space:]]*=)?
   "<"+ {
     global nesting
     obstack_for_string.append(yytext)
-    nesting += gram_leng
+    nesting += len(yytext)
   }
 
   <<EOF>>                      unexpected_eof(token_start, '>')
@@ -651,7 +651,7 @@ eqopt    ([[:space:]]*=)?
   \\[0-7]{1,3} {
     c = strtoul(yytext + 1, None, 8)
     if not c or 0x7f * 2 + 1 < c:
-      state.complain(y_tab.yylloc, state.complaint, 'invalid number after \\-escape: %s', yytext + 1)
+      state.complain(y_tab.yylloc, state.complaint, 'invalid number after \\-escape: {0:s}'.format(yytext[1:]))
     else:
       obstack_for_string.append(chr(c))
       rpl_sprintf(gram_piece_temp, '<AST_Text_Escape character="%d">', int(c))
@@ -663,7 +663,7 @@ eqopt    ([[:space:]]*=)?
   \\x[0-9abcdefABCDEF]+ {
     c = strtoul(yytext + 2, None, 16)
     if not c or 0x7f * 2 + 1 < c:
-      state.complain(y_tab.yylloc, state.complaint, 'invalid number after \\-escape: %s', yytext + 1)
+      state.complain(y_tab.yylloc, state.complaint, 'invalid number after \\-escape: {0:s}'.format(yytext[1:]))
     else:
       obstack_for_string.append(chr(c))
       rpl_sprintf(gram_piece_temp, '<AST_Text_Escape character="%d">', int(c))
@@ -727,7 +727,7 @@ eqopt    ([[:space:]]*=)?
   \\(u|U[0-9abcdefABCDEF]{4})[0-9abcdefABCDEF]{4} {
     c = convert_ucn_to_byte(yytext)
     if c <= 0:
-      state.complain(y_tab.yylloc, state.complaint, 'invalid number after \\-escape: %s', yytext + 1)
+      state.complain(y_tab.yylloc, state.complaint, 'invalid number after \\-escape: {0:s}'.format(yytext[1:]))
     else:
       obstack_for_string.append(chr(c))
       rpl_sprintf(gram_piece_temp, '<AST_Text_Escape character="%d">', c)
@@ -736,12 +736,12 @@ eqopt    ([[:space:]]*=)?
       gram_piece_append('</AST_Text_Escape>')
   }
   \\(.|\n) {
-    p = yytext + 1
-    if c_isspace(int(*p)) and c_isprint(int(*p)):
-      p = quote(p)
-    else:
-      p = quotearg_style_mem(escape_quoting_style, p, 1)
-    state.complain(y_tab.yylloc, state.complaint, 'invalid character after \\-escape: %s', p)
+    p = yytext[1:]
+    if True: #c_isspace(int(*p)) and c_isprint(int(*p)):
+      p = state.quote(p)
+    #else:
+    #  p = quotearg_style_mem(escape_quoting_style, p, 1)
+    state.complain(y_tab.yylloc, state.complaint, 'invalid character after \\-escape: {0:s}'.format(p))
   }
 }
 
@@ -985,7 +985,7 @@ eqopt    ([[:space:]]*=)?
 def scan_integer(number, base, loc):
   num = int(number, base)
   if 0x7fffffff < num:
-    state.complain(y_tab.yylloc, state.complaint, 'integer out of range: %s', quote(number))
+    state.complain(y_tab.yylloc, state.complaint, 'integer out of range: {0:s}'.format(state.quote(number)))
     num = 0x7fffffff
   return num
 
@@ -1009,18 +1009,18 @@ def scan_integer(number, base, loc):
 
 def unexpected_end(start, msg, token_end):
   loc = state.location(start.copy(), scanner_cursor.copy())
-  scanner_cursor.column -= i
+  scanner_cursor.column -= len(token_end)
   unput(token_end)
-  token_end = quote(token_end)
+  token_end = state.quote(token_end)
   if token_end == '\'\\\'\'':
     token_end = '"\'"'
-  state.complain(y_tab.yylloc, state.complaint, msg, token_end)
+  state.complain(y_tab.yylloc, state.complaint, msg.format(token_end))
 
 def unexpected_eof(start, token_end):
-  unexpected_end(start, 'missing {0:s} at end of file'.format(token_end))
+  unexpected_end(start, 'missing {0:s} at end of file', token_end)
 
 def unexpected_newline(start, token_end):
-  unexpected_end(start, 'missing {0:s} at end of line'.format(token_end))
+  unexpected_end(start, 'missing {0:s} at end of line', token_end)
 
 #def gram_scanner_initialize():
 #  global obstack_for_string

diff --git a/bootstrap/state.py b/bootstrap/state.py
index f4eb74c..a87299a 100644 (file)
--- a/bootstrap/state.py
+++ b/bootstrap/state.py
@@ -72,3 +72,6 @@ precedence_assoc = 4
 
 destructor = 0
 printer = 1
+
+def quote(str):
+  return '"{0:s}"'.format(str.replace('\\', '\\\\').replace('"', '\\"'))