[Bug c/52952] Wformat location info is bad (wrong column number)

["manu at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52952

--- Comment #22 from Manuel LÃpez-IbÃÃez <manu at gcc dot gnu.org> 2013-04-01 14:17:45 UTC ---
(In reply to comment #13)
> and didn't need to be translated.  So, printf ("%.*d"); (the common case)
> wouldn't have to be recorded, while printf (R"<<<(%)>>>" "." R"(*)" "d"); would

BTW, your testcase does not work for me and confuses GCC:

/home/manuel/loc_offset.c:7:21: error: unterminated raw string
   __builtin_printf (R"<<<(%)>>>" R"(*d)" );
                     ^
/home/manuel/loc_offset.c:7:3: error: expected expression at end of input
   __builtin_printf (R"<<<(%)>>>" R"(*d)" );
   ^
/home/manuel/loc_offset.c:7:3: error: expected declaration or statement at end
of input

It is also very unclear to me how to re-process such a token stream in order to
find out the real location+offset.

A further issue is that for:

#define FORMAT ".*d"
#include <stdio.h>
void f() {
   printf(
"%"
FORMAT);
}

the token representing "%" is overriden when we call cpp_get_token next time:

@@ -1003,10 +1004,13 @@ lex_string (const cpp_token *tok, tree *
      concatenation.  It is 'true' if we have seen an '@' before the
      current string, and 'false' if not.  We must see exactly one or
      zero '@' before each string.  */
   bool objc_at_sign_was_seen = false;

+  // This token is overwritten! How can that happen?
+  const cpp_token *prev_tok = tok;
+
  retry:
   tok = cpp_get_token (parse_in);
   switch (tok->type)
     {
     case CPP_PADDING:

so prev_tok above may or may not change by the call of cpp_get_token.