This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug fortran/55210] cannot use macro #define FOO 'a' in a conditional
- From: "burnus at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: Mon, 05 Nov 2012 15:32:24 +0000
- Subject: [Bug fortran/55210] cannot use macro #define FOO 'a' in a conditional
- Auto-submitted: auto-generated
- References: <bug-55210-4@http.gcc.gnu.org/bugzilla/>
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55210
Tobias Burnus <burnus at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |NEW
Last reconfirmed| |2012-11-05
CC| |burnus at gcc dot gnu.org
Ever Confirmed|0 |1
--- Comment #1 from Tobias Burnus <burnus at gcc dot gnu.org> 2012-11-05 15:32:24 UTC ---
If fails in cpp_interpret_string at
1426 if (!APPLY_CONVERSION (cvt, base, p - base, &tbuf))
Because "func" is a NULL pointer:
(gdb) p cvt
$13 = {func = 0x0, cd = 0x0, width = 0}
Backtrace:
#0 cpp_interpret_string (pfile=pfile@entry=0x15d9a90,
from=from@entry=0x15da320, count=count@entry=1, to=to@entry=0x7fffffffd1c0,
type=<optimized out>) at
/home/tob/projects/gcc-git/gcc/libcpp/charset.c:1426
#1 0x0000000000e71b7d in cpp_interpret_charconst (pfile=0x15d9a90,
token=0x15da318, pchars_seen=0x7fffffffd30c, unsignedp=0x7fffffffd3f0)
at /home/tob/projects/gcc-git/gcc/libcpp/charset.c:1615
#2 0x0000000000e5fe85 in eval_token (virtual_location=4868, token=0x15da318,
pfile=0x15d9a90)
at /home/tob/projects/gcc-git/gcc/libcpp/expr.c:964
#3 _cpp_parse_expr (pfile=<optimized out>, is_if=<optimized out>) at
/home/tob/projects/gcc-git/gcc/libcpp/expr.c:1150
#4 0x0000000000e599b4 in do_if (pfile=0x15d9a90) at
/home/tob/projects/gcc-git/gcc/libcpp/directives.c:1953
#5 0x0000000000e5b1b1 in _cpp_handle_directive (pfile=0x15d9a90,
indented=<optimized out>)
at /home/tob/projects/gcc-git/gcc/libcpp/directives.c:492
#6 0x0000000000e6ea6f in _cpp_scan_out_logical_line
(pfile=pfile@entry=0x15d9a90, macro=macro@entry=0x0)
at /home/tob/projects/gcc-git/gcc/libcpp/traditional.c:635
#7 0x0000000000e6ef6a in _cpp_read_logical_line_trad (pfile=0x15d9a90) at
/home/tob/projects/gcc-git/gcc/libcpp/traditional.c:306
#8 0x000000000056fa20 in scan_translation_unit_trad (pfile=<optimized out>) at
/home/tob/projects/gcc-git/gcc/gcc/fortran/cpp.c:773
The cvt variable is set via a call to "converter_for_type" with the argument
type=CPP_CHAR. The function contains:
1346 switch (type)
1347 {
1348 default:
1349 return pfile->narrow_cset_desc;
(and also handles CPP_UTF8STRING, CPP_CHAR16, CPP_CHAR32, CPP_WCHAR and
CPP_WSTRING).
The CPP_CHAR is probably generated in lex.c's lex_string
The function narrow_cset_desc should get set via:
cpp_init_iconv (cpp_reader *pfile)
which has:
pfile->narrow_cset_desc = init_iconv_desc (pfile, ncset, SOURCE_CHARSET);
where "nset" gets set via:
const char *ncset = CPP_OPTION (pfile, narrow_charset);
If one calls "cpp" for the file (which matches calling "cc1 -E"), the type is
also CPP_CHAR, but pfile->narrow_cset_desc alias cvt.func is
convert_no_conversion.
* * *
I wonder whether one should call cpp_init_iconv - as C/C++ does in
../gcc/c-family/c-opts.c's c_common_init:
/* Set up preprocessor arithmetic. Must be done after call to
c_common_nodes_and_builtins for type nodes to be good. */
cpp_opts->precision = TYPE_PRECISION (intmax_type_node);
cpp_opts->char_precision = TYPE_PRECISION (char_type_node);
cpp_opts->int_precision = TYPE_PRECISION (integer_type_node);
cpp_opts->wchar_precision = TYPE_PRECISION (wchar_type_node);
cpp_opts->unsigned_wchar = TYPE_UNSIGNED (wchar_type_node);
cpp_opts->bytes_big_endian = BYTES_BIG_ENDIAN;
/* This can't happen until after wchar_precision and bytes_big_endian
are known. */
cpp_init_iconv (parse_in);