Jim Kingdon [Thu, 30 Dec 1993 15:27:52 +0000 (15:27 +0000)]
* infrun.c (wait_for_inferior): Enable code which assumes that if
we jump into the prologue from another function, then it was a
subroutine call. #if 0 AT_FUNCTION_START; the above code should
take care of this case.
Jim Kingdon [Wed, 29 Dec 1993 20:52:06 +0000 (20:52 +0000)]
* c-exp.y: Revert Kung's change. "..." is not a type, and the
change caused "p (...)0" to dump core.
* gdbtypes.c (check_stub_method): Don't pass "..." to
parse_and_eval_type. This should fix the bug which Kung was
trying to fix.
Jim Kingdon [Wed, 29 Dec 1993 18:42:52 +0000 (18:42 +0000)]
* stabsread.c (define_symbol): If we choose not to combine
two symbols, don't just ignore the second (LOC_REGISTER) one.
* printcmd.c (print_frame_args): If we have a LOC_ARG and a
LOC_REGISTER, use the LOC_ARG not the LOC_REGISTER.
Mark Eichin [Mon, 27 Dec 1993 19:39:55 +0000 (19:39 +0000)]
* config/i386-linux.mt: delete ".sh" from genscripts argument
(change made to genscripts on July 21...)
This is all it took for linux-native ld to build. It isn't necessarily useful,
particularly as gcc is still configured to pass -m486 and -dll-verbose, which
this ld doesn't parse...
Jim Kingdon [Mon, 27 Dec 1993 16:31:56 +0000 (16:31 +0000)]
* dbxread.c: Move default definition of GCC_COMPILED_FLAG_SYMBOL
from here . . .
* symtab.h: . . . to here.
* dbxread.c (record_minimal_symbol): Move check for gcc{,2}_compiled.
and __gnu_compiled* from here . . .
* minsyms.c (prim_record_minimal_symbol_and_info): . . . to here.
* minsyms.c (prim_record_minimal_symbol): Call
prim_record_minimal_symbol_and_info rather than duplicating code.
Jeff Law [Mon, 27 Dec 1993 04:46:05 +0000 (04:46 +0000)]
* dbxread.c (process_one_symbol): Handle stabs-in-som just like
stabs-in-elf.
(pastab_build_psymtabs): Likewise.
* hppa-tdep.c: Change all comments to reference %r3 or frame pointer
rather than %r4.
(frame_chain, skip_prologue, dig_rp_from_stack): Handle %r3 as the
frame pointer.
* config/pa/tm-hppa.h (FP_REGNUM): Define as %r3.
(FIND_FRAME_SAVED_REGS): Handle %r3 as frame pointer.
(CALL_DUMMY): Likewise.
* infrun.c (wait_for_inferior): Remove confusing and inaccurate
stuff about subroutine calls, return, etc., from comment which
says "We've wandered out of the step range.".
Jim Kingdon [Sun, 26 Dec 1993 22:37:47 +0000 (22:37 +0000)]
* gdbtypes.c (force_to_range_type): Use switch statement.
complain() not warning() if the TYPE_CODE isn't one we know how to
deal with gracefully. Use builtin_type_int not
lookup_fundamental_type (the objfile we passed to
lookup_fundamental_type was sometimes NULL).
Jeff Law [Sun, 26 Dec 1993 04:06:32 +0000 (04:06 +0000)]
* hppa-tdep.c (init_extra_frame_info): Correctly adjust the base
of the current frame when "fromleaf" is true. Do not adjust the
frame base of the innermost frame if it is a leaf function.
Jim Kingdon [Sat, 25 Dec 1993 18:44:22 +0000 (18:44 +0000)]
* defs.h: Move definitions of TARGET_*_BIT after include of tm.h.
The old way (using #undef in tm.h) was ugly and asking for
trouble, because it makes it possible for some file to use the
wrong definition. Move definition of HOST_CHAR_BIT after definition
of TARGET_CHAR_BIT.
* config/alpha/tm-alpha.h, config/h8300/tm-h8300.h,
config/h8500/tm-h8500.h, config/z8k/tm-z8k.h: Don't undef TARGET_*_BIT
before defining them.
Jim Kingdon [Fri, 24 Dec 1993 20:43:29 +0000 (20:43 +0000)]
* stabsread.c (read_type): Don't fall through 'S' case (the case it
was falling though happened to do the thing thing ("break;") but that
is hardly a good thing to assume).
Ian Lance Taylor [Thu, 23 Dec 1993 17:29:06 +0000 (17:29 +0000)]
gcc -Wall lint:
* ar.c (main): Put parens around assignment used as truth value.
* objdump.c (display_bfd): Declare return value as void.
(stab_print): Use "" instead of 0 to initialize array.
(dump_stabs_1): Print vma using printf_vma.
(display_info): Put parens around assignment used as truth value.
* strings.c (print_strings): Cast printf arguments.
Ken Raeburn [Tue, 21 Dec 1993 17:24:10 +0000 (17:24 +0000)]
(aix386_core_file_p): Use cd_regs[0] for computing the offsetof because AIX
/bin/cc does not like to take the address of an array. (From Minh Tran-Le.)
Ken Raeburn [Mon, 20 Dec 1993 21:53:01 +0000 (21:53 +0000)]
* write.c (adjust_reloc_syms) [RELOC_REQUIRES_SYMBOL]: If no symbol is present
(i.e., relocation against absolute), create a phony local symbol, and use it in
the reloc.
(fixup_segment): When adjusting a reloc with an absolute symbol, have
TC_FORCE_RELOCATION control clearing add_symbolP too.
Ken Raeburn [Mon, 20 Dec 1993 16:48:00 +0000 (16:48 +0000)]
* as.c (main): Only invoke md_end if it's defined as a macro.
* tc.h (md_end): Don't declare it.
* config/tc-*.[ch] (md_end): Deleted, in cases where it doesn't do anything.
* config/tc-vax.c (vip_end): Deleted null function.
* config/tc-mips.c (md_mips_end): Renamed from md_end.
* config/tc-mips.h (md_mips_end): Declare.
(md_end): New macro, calls md_mips_end.
* write.c (write_object_file): Don't close output file.
* as.c (main): Close output file (if needed) after calling listing_print, which
should be after calling write_object_file, which sets the frag addresses.
Kung Hsu [Fri, 17 Dec 1993 18:51:17 +0000 (18:51 +0000)]
Modified Files:
ChangeLog symtab.c c-valprint.c
* symtab (decode_line_1): fix a bug when position char is not
set correctly.
* c-valprint (c_val_print): handle vtbl printing when vtbl is not
set up yet.
Jeff Law [Fri, 17 Dec 1993 05:16:01 +0000 (05:16 +0000)]
* config/tc-hppa.c (md_begin): Initialize "dummy_label".
(tc_gen_reloc, SOM version): For relocations which have no symbol,
set sym_ptr_ptr to dummy_label. Avoids lossage in generic BFD code.
Ken Raeburn [Thu, 16 Dec 1993 21:31:22 +0000 (21:31 +0000)]
* config/tc-vax.c (tc_aout_fix_to_chars): Local variable NBYTES_R_LENGTH now
const.
* config/obj-*.c, config/tc-*.c: Omit superfluous "return" statements at ends
of functions. Don't check for null return from hash_new, since it won't return
at all if there's no memory available. Also, check for null return from
hash_insert, rather than zero-length string, as success indicator.
Ken Raeburn [Wed, 15 Dec 1993 16:28:46 +0000 (16:28 +0000)]
Some changes from trying to run test suite in California last week:
* Makefile.in (site.exp): Don't set ASFLAGS. Quote value of OBJDUMPFLAGS in
case it's empty. Use temporary names until the end; make creating site.exp the
final step.
(check): Pass in ASFLAGS.
* gas/all/gas.exp: Use all_ones proc. Change regexp for matching
C comments to avoid bugs in latest expect code.
* config/h8300/tm-h8300.h (BREAKPOINT): Insn changed to sleep.
(DECP_PC_AFTER_BREAK): Now is 0.
* config/h8500/tm-h8500.h (REGISTER_BYTES, REGISTER_BYTE,
REGISTER_NAMES): update to new view. (INIT_EXTRA_FRAME_INFO): No
extra frame info now.
* config/sh/sh.h (NOP): Define NOP insn.
* config/z8k/tm-z8k.h (BIG): is now sim_z8001_mode.
* config/z8k/z8ksim.mt (TDEPFILES): Add remote-sim.o to list.
* ser-go32.c: Lint. (strncasecmp): Removed, now in libiberty.
(go32_readchar): Special handling for faster polling. (async
structure): Volatile.
* h8300-tdep.c (print_register_hook): Allocate and use the right
number bytes for the raw register.
* h8500-tdep.c (regoff, frame_find_saved_reg, examine_prologue):
deleted. (h8500_register_size, h8500_register_virtual_type, ):
Use new way of counting registers.
* remote-e7000.c (echo_index): deleted. (expect): Better handling
of user interrupts. (expect_prompt): Remove never used log file
support. (want, want_nopc): Add support for H8/300H.
(fetch_regs_from_dump): Treat \r and \n as whitespace.
(e7000_drain): Send an "end" command before waiting for output to
stop. (e7000_wait): Cope with H8/300H, better handling of user
interrupts. (why_stop, expect_n, sub2_from_pc): New function.
* remote-utils.c (gr_load_image): call fflush and QUIT more regularly.
* utils.c (notice_quit): New function for polling for user interrupts.
Ian Lance Taylor [Sat, 11 Dec 1993 22:55:55 +0000 (22:55 +0000)]
Made many changes to eliminate gcc warnings. Made various
cosmetic changes, declared various things in header files, removed
various extern declarations from .c files. No substantive
changes.