This is the mail archive of the binutils@sourceware.org mailing list for the binutils project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

gas not accepting clang exceptions on powerpc



comparing gcc verses clang-llvm assembler output for PowerPC (apple-powerbook, OSX 10.4, using installed gcc/gas, gcc -version --> 4.0.1, as -version --> 1.38)


gas does not complain about differencing two symbols from different sections
in a ".long" directive, but does complain when first defining that difference as
a symbol name.


Is this really the way gas is supposed to work, or is this version
of gas too old, or what ---?

example: ------------------------------------------------------------------------ -

        .section        __TEXT,__text,regular,pure_instructions
        .globl  __Z3barv
__Z3barv:
Leh_func_begin0:
        mflr r0

.section __TEXT,__eh_frame,coalesced,no_toc +strip_static_syms+live_support
.globl __Z3barv.eh
__Z3barv.eh:
.long Leh_func_begin0-. ;;; <--- this is the syntax gcc uses, no complaints from gas


Ltmp20:
Ltmp21 = Leh_func_begin0-Ltmp20 ;;; <--- this is the syntax clang uses, gas complains
.long Ltmp21


Ltmp31 = Leh_func_begin0-. ;;; <--- experiment, gas also complains
.long Ltmp31





sincerely, Peter Lawrence.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]