https://gcc.gnu.org/bugzilla/show_bug.cgi?id=123816
Bug ID: 123816
Summary: [M68K][Fixed-Point] Internal compilation error when
using typedef
Product: gcc
Version: 14.3.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: target
Assignee: unassigned at gcc dot gnu.org
Reporter: djipi.mari at gmail dot com
Target Milestone: ---
I have adapted the gcc 14.3.0 config file, and compiled with the
--enable-fixed-point so gcc can compile with the fixed-point types.
I'm using gcc 14.3.0 but I think the issue can also apply to more recent gcc
versions.
About the configure options:
--prefix="/c/Projects/GNU/build" \
--target=m68k-elf \
--with-cpu=m68000 \
--with-newlib --with-gnu-as --with-gnu-ld \
--without-headers \
--enable-fixed-point \
--enable-languages=c,c++ \
--enable-lto \
--enable-silent-rules \
--enable-static \
--disable-bootstrap \
--disable-checking \
--disable-docs \
--disable-hosted-libstdcxx \
--disable-libgomp \
--disable-libssp \
--disable-libstdcxx-pch \
--disable-libstdcxx-verbose \
--disable-libquadmath \
--disable-multilib \
--disable-nls \
--disable-plugins \
--disable-shared \
--disable-threads \
--disable-tls
And I have added these lines in the gcc configure with the mips, arm &
loongarch parts:
m68k*)
enable_fixed_point=yes
;;
The gcc compiler is generated without problems.
For a project, I wanted to use the typedef instruction.
typedef _Accum vec_t;
vec_t a;
_Accum b;
...
if (a < b)
...
The test line will generate this compilation error:
sorry, unimplemented: GCC cannot support operators with integer types and
fixed-point types that have too many integral and fractional bits together
40 | if (a < b)
| ^~
testaccum.c:40: confused by earlier errors, bailing out
If I do the following line, no compilation error is generated
...
if ((_Accum)a < b)
It seems typedef may be confused with fixed point type.
