(pv_is_array_ref): Remove.
(s390_on_stack): Remove.
(S390_NUM_SPILL_SLOTS): Remove.
(struct s390_prologue_data): Remove members 'spill' and 'back_chain',
add members 'gpr_slot', 'fpr_slot', and 'back_chain_saved_p'.
(s390_store): Track all stores of orginal incoming registers to the
stack constant offsets relative to the CFA, instead of only tracking
stores into specific spill slots.
(s390_load): Likewise.
(s390_analyze_prologue): Adapt to struct s390_prologue_data changes.
(s390_prologue_frame_unwind_cache): Likewise. Only track registers
defined as call-saved by the ABI.
(s390_push_dummy_call): Use bottom of dummy call argument save area
as return value, not the top. Do not store to the called function's
register save area.
(s390_unwind_dummy_id): Adapt accordingly.
* ada-exp.y, ada-lex.l: Use xmalloc and xrealloc through out.
* inf-ptrace.c: Include "gdb_wait.h" instead of <wait.h>.
* Makefile.in (.y.c, .l.c): Don't replace xmalloc with xxmalloc.
Update dependencies.
(ada_evaluate_subexp): Change error message for
OP_FUNCALL case so as not to indicate an internal error.
From Joel Brobecker; committed by Paul Hilfinger.
* ada-lang.c (ada_template_to_fixed_record_type_1): Detect bogus
type sizes in XVE/XVU types, and emit a warning when that happens,
instead of rounding up the size of the resulting type.
Some minor reformatting.
(ada_is_aligner_type): Do not consider a type as an aligner type
if there is a corresponding XVS type as well.
(ada_evaluate_subexp): For TERNOP_SLICE, handle cases where we
have several levels of pointer indirection to the array.
From Jerome Guitton; committed by Paul Hilfinger.
* ada-lang.c (check_size): New function.
(coerce_unspec_val_to_type): call check_size to make sure
that the object size is reasonable.
(ada_evaluate_subexp): in UNOP_IND: Ditto. In TERNOP_SLICE:
Implements the case when we get a reference when we evaluate
the expression of one of the bounds of the slice.
(ada_value_struct_elt): when a packed array field is
a reference, use ada_coerce_ref to dereference it; use
ada_value_ind only if it is a pointer.
a bit to avoid the necessity of an advance declaration. Remove
declaration.
(set_reg_offset): Move implemenation up.
(mips16_get_imm): Likewise.
(mips16_scan_prologue): Likewise.
(reset_saved_regs): Likewise.
(mips32_scan_prologue): Likewise.
Remove sp as a parameter, compute it internally instead. Move
prologue scanning limit from heuristic_proc_desc to here.
(mips16_scan_prologue): Likewise.
(mips_insn16_frame_cache): Replace call to heuristic_proc_desc
by call to mips16_scan_prologue.
(mips_insn32_frame_cache): Replace call to heuristic_proc_desc
by call to mips32_scan_prologue.
(heuristic_proc_desc): Remove code that's no longer necessary.
Update calls to mips16_scan_prologue and mips32_scan_prologue.
(mips_skip_prologue): Update calls to mips16_scan_prologue
and mips32_scan_prologue.
mips16_skip_prologue. Now return the address of the first
instruction past the function prologue.
(mips16_skip_prologue): Remove, no longer necessary.
(mips_skip_prologue): Replace call to mips16_skip_prologue
by call to mips16_scan_prologue.
mips32_skip_prologue. Now return the address of the first
instruction past the function prologue.
(mips32_skip_prologue): Remove. No longer necessary.
(mips16_skip_prologue): Add parameter end_pc instead of
computing it.
(mips_skip_prologue): Compute the upper limit for the
prologue scanning. Update call to mips16_skip_prologue.
Replace call to mips32_skip_prologue by call to
mips32_scan_prologue.
(fetch_fpregs): Use i387_supply_fxsave to fill the floating-point
registers when the floating-point state isn't initialized.
* Makefile.in (i386gnu-nat.o): Update dependencies.
(mips_insn16_frame_cache): Pass frame cache in call to
heuristic_proc_desc. Move some code to mips16_heuristic_proc_desc.
Remove code that became redundant as a consequence.
(mips32_heuristic_proc_desc): No longer compute a fake
procedure descriptor. Compute the full frame cache instead.
Some minor comment reformatting.
(sparc_linux_sigtramp_start): Change first argument from CORE_ADDR
to struct frame_info *. Use safe_frame_unwind_memory instead of
deprecated_read_memory_nobpt.
(sparc32_linux_sigtramp_start, sparc32_linux_rt_sigtramp_start):
Change argumen from CORE_ADDR to struct frame_info *. Adjust for
changed signature of sparc_linux_sigtramp_start.
(sparc32_linux_pc_in_sigtramp): Remove function.
(sparc32_linux_sigtramp_p): New function.
(sparc32_linux_sigtramp_frame_sniffer): Use
sparc32_linux_sigtramp_p instead of sparc32_linux_pc_in_sigtramp.
* target.h (to_stopped_data_address): Change prototype to
take a CORE_ADDR pointer and return an int.
* target.c (update_current_target): Change to_stopped_data_address
to match new prototype.
(debug_to_stopped_data_address): Change appropriately.
* breakpoint.c (bpstat_stop_status): Change call to
target_stopped_data_address to use new prototype.
* frv-tdep.c (frv_have_stopped_data_address): New function.
(frv_stopped_data_address): Change to new prototype and
functionality.
* ia64-linux-nat.c (ia64_stopped_data_address): Change to new
prototype and functionality.
(ia64_stopped_by_watchpoint): New function.
* i386-nat.c (i386_stopped_data_address): Change to new
prototype and functionality.
(i386_stopped_by_watchpoint): New function.
* remote.c (remote_stopped_data_address): Change to new prototype
and functionality.
* remote-m32r-sdi.c (m32r_stopped_data_address): Ditto.
* config/frv/tm-frv.h (frv_stopped_data_address): Change prototype.
(STOPPED_BY_WATCHPOINT): Change to use frv_have_stopped_data_address.
* config/i386/nm-i386.h (STOPPED_BY_WATCHPOINT): Change to use
new i386_stopped_by_watchpoint function.
(i386_stopped_by_watchpoint): New prototype.
(i386_stoppped_data_address): Change to new prototype.
* config/ia64/nm-linux.h (STOPPED_BY_WATCHPOINT): Change to use
new ia64_stopped_by_watchpoint function.
(ia64_stopped_by_watchpoint): New prototype.
(ia64_stopped_data_address): Ditto.
modulus of modular type.
* ada-lang.h (ada_modulus): Change return type to unsigned.
* ada-lang.c (ada_language_arch_info): Correct type of
string_char_type.
(ada_modulus): Return unsigned result.
(ada_lookup_symbol): Remove user disambiguation and modify comment
accordingly.
2004-10-08 Eric Botcazou <ebotcazou@act-europe.fr>
Committed by Paul Hilfinger.
* ada-lang.c (decode_packed_array): On big-endian targets,
left-justify the value if it comes with a modular type.
2004-10-08 Joel Brobecker <brobecker@gnat.com>
Committed by Paul Hilfinger.
* ada-lang.c (template_to_static_fixed_type): Fix a small typo.
(is_name_suffix): Add support for "___JM" suffixes.
(ada_check_typedef): Renames ada_completed_type.
Update all users of CHECK_TYPEDEF, check_typedef, and
ada_complete_type to call ada_check_typedef.
* ada-typeprint.c: Likewise.
* ada-valprint.c: Likewise.
* m32r-tdep.c (decode_prologue): Support functions written
in assembler language. Recognize trap instructions as the
end of prologue.
(m32r_frame_unwind_cache): Ditto.
(m32r_skip_prologue): Extend search limit. Quit analyzing
prologue if pc's location is not readable.
