* ldlang.c (lang_process): don't pass null pointers when
abs_output_section is what is required.
* ldwrite.c (ldwrite): use malloc to allocate the largest space
used, and pass that down.
* relax.c,relax.h (write_relaxnorel): use the passed malloc area rather
than alloca.
Fred Fish [Wed, 26 Aug 1992 23:05:55 +0000 (23:05 +0000)]
* cplus-dem.c: Reorder args to most local functions to make them
follow the consistent pattern of struct work_stuff pointer first,
followed by pointer into string being demangled, followed by
pointer to the string where the result is being stored, followed
by other args. Also generally replace most variables of name
"type" with "mangled" for consistency. They all pointed to the
same string but half were one name and half the other.
* cplus-dem.c (LONGERNAMES): Remove #ifndef's.
* cplus-dem.c (demangle_qualified): Rewrite to clean up and allow
use whenever a qualified name is being demangled.
* cplus-dem.c (do_type): Qualified names are treated similarly
to fundamental types, not as modifiers of fundamental types.
Ken Raeburn [Wed, 26 Aug 1992 04:08:45 +0000 (04:08 +0000)]
* config/tc-m68k.c (init_table): Now const. Always include 68851
data, so that "bc" is available to 68040 cache instructions. Added
"tt0", "tt1", and 68ec030 variants.
(md_assemble): Complain if 68000 (only) and 68881 are specified.
(enum _register): Added TT0, TT1.
(m68k_ip, cases '3' and 't'): Handle new operand type codes. Pass
line number correctly in "internal error" messages. Don't print
architecture-mismatch message for operand errors.
From Colin Smith (colin@wrs.com):
* config/tc-m68k.c (m68k_ip, case '_'): Use addword twice rather than
install_operand.
Ken Raeburn [Wed, 26 Aug 1992 04:02:42 +0000 (04:02 +0000)]
* config/tc-m68k.c (init_table): Now const. Always include 68851
data, so that "bc" is available to 68040 cache instructions. Added
"tt0", "tt1", and 68ec030 variants.
(md_assemble): Complain if 68000 (only) and 68881 are specified.
(enum _register): Added TT0, TT1.
(m68k_ip, cases '3' and 't'): Handle new operand type codes. Pass
line number correctly in "internal error" messages. Don't print
architecture-mismatch message for operand errors.
From Colin Smith (colin@wrs.com):
* config/tc-m68k.c (m68k_ip, case '_'): Use addword twice rather than
install_operand.
Ken Raeburn [Wed, 26 Aug 1992 03:46:22 +0000 (03:46 +0000)]
m68k.h (m68ec030): Defined as alias for 68030.
(m68k_opcodes): New type characters "3" for 68030 MMU regs and "t" for
immediate 0-7 added. Set up some opcodes (ptest, bkpt) to use them.
Tightened description of "fmovex" to distinguish it from some "pmove"
encodings. Added "pmove" for 68030 MMU regs, cleaned up descriptions that
claimed versions were available for chips not supporting them. Added
"pmovefd".
Roland Pesch [Tue, 25 Aug 1992 21:20:06 +0000 (21:20 +0000)]
Markup changes to run through TeX without errors.
WARNING: (1) this has not yet been proofread as a printout---just eliminated
problems that TeX complained about.
(2) extraneous line breaks introduced in some examples, sometimes
in ridiculously long strings; this may falsify examples. Places
where I did this are marked by FIXME comments.
Roland Pesch [Tue, 25 Aug 1992 20:56:41 +0000 (20:56 +0000)]
Some improvements in Texinfo markup:
1) use makeinfo node defaulting
2) correct bad xref
3) use @display, @exdent, etc. as first approximation to Julia's
clever typewriter markup; **THIS WILL NEED STUDY AND REVISION**
4) introduce many nodes rather than ---- in "detailed reference" appendix
* coffcode.h (coff_add_missing_symbols): if symbols don't come
from a coff file (csym is null), dont deref them
* ieee.c (parse_expression): get the answer right when adding an
abs+(sec+off), (ieee_generic_stat_arch_elt): call ieee_object_p
on elts so that filename is filled in.
Fred Fish [Mon, 24 Aug 1992 19:00:24 +0000 (19:00 +0000)]
* infrun.c (handle_command): Rewrite to allow multiple signal
numbers, signal number ranges, and to recognize "all" to mean all
signals except those used by the debugger.
* infrun.c (SET_SIGS, UNSET_SIGS): Macros used in handle_command
to set or reset actions for specific signals.
John Gilmore [Mon, 24 Aug 1992 08:20:14 +0000 (08:20 +0000)]
FIXME. Half-done documentation for the macros used to configure GDB's
sources. What's there now is a list of all macros, and at least one
source file in which they are referenced.
Fred Fish [Mon, 24 Aug 1992 00:13:11 +0000 (00:13 +0000)]
* coffread.c (decode_type): Call alloc_type to alloc new
types.
* stabsread.c (read_array_type, read_range_type, define_symbol):
Call alloc_type to alloc new types.
* stabsread.c (define_symbol): Move dbl_valu symbol field data
from type_obstack to symbol_obstack.
* stabsread.c (define_symbol): Move typedef_sym from type_obstack
to symbol_obstack.
* gdbtypes.h (TYPE_ALLOC): New macro to allocate space for data
associated with a type, using the same mechanism as was used to
allocate space for the type structure itself.
* coffread.c (patch_type, coff_read_struct_type,
coff_read_enum_type): Use TYPE_ALLOC.
* dwarfread.c (struct_type): Use TYPE_ALLOC.
* gdbtypes.c (create_array_type, check_stub_method,
allocate_cplus_struct_type): Use TYPE_ALLOC.
* mipsread.c (parse_symbol, parse_type): Use TYPE_ALLOC.
* stabsread.c (read_struct_type, read_array_type, read_enum_type,
read_range_type): Use TYPE_ALLOC.
Fred Fish [Sun, 23 Aug 1992 18:09:19 +0000 (18:09 +0000)]
* breakpoint.c (breakpoint_re_set): Select a default source
symtab if one is not currently selected.
* utils.c (query): Call wrap_here before building the output
query string, to turn off wrapping and flush any buffered output.
Otherwise our query may end up in the wrap buffer and never be
seen by the user.
* eval.c (evaluate_subexp): Report error when attempting to
evaluate subscript for types without a target type, rather
than dumping core by using the NULL pointer.
* symfile.c (symbol_file_command): Forget current_source_symtab
and current_source_line when discarding symbol table.
Stu Grossman [Fri, 21 Aug 1992 22:35:35 +0000 (22:35 +0000)]
* remote.c (remote_open): Fix baud rate setting to make -b flag
work. (remote_wait): Change 'T' message parser to deal with new
improved format which allows stub to send an arbitrary bunch of
registers.
* sparc-stub.c: General cleanups. (trap_low, handle_exception):
make all this re-entrant by storing all state on the stack. Clean
up memory error trapping. (computeSignal, set_debug_traps):
make it all table driven. Make a start at a baud rate setting command.
Ken Raeburn [Fri, 21 Aug 1992 19:21:52 +0000 (19:21 +0000)]
From p3:
* messages.c (as_warn): Use fputs, not fprintf, with a buffer that
has already been formatted (but may still contain %-characters).
(as_bad): Likewise.
Ian Lance Taylor [Wed, 19 Aug 1992 18:27:48 +0000 (18:27 +0000)]
This is for PR 628.
Wed Aug 19 11:20:59 1992 Ian Lance Taylor (ian@cygnus.com)
* tc-m68k.c, tc-m68kmote.c: the cas2 instruction is supposed to be
written with indirection on the last two operands, which can be
either data or address registers. Added a new operand type 'r'
which accepts either register type. Added '(' to notend stuff in
tc-m68kmote.c to accept (a0):(a2) in cas2 instruction.
Ian Lance Taylor [Wed, 19 Aug 1992 17:26:47 +0000 (17:26 +0000)]
Wed Aug 19 10:23:27 1992 Ian Lance Taylor (ian@cygnus.com)
* remote-vx.c: redefine malloc to avoid buggy declaration on
RS/6000 <rpc/types.h>.
xm-rs6000.h: include <sys/select.h> to define fd_set for
<rpc/svc.h> on RS/6000.
Roland Pesch [Tue, 18 Aug 1992 23:02:35 +0000 (23:02 +0000)]
Tue Aug 18 15:59:13 1992 Roland H. Pesch (pesch@fowanton.cygnus.com)
* gdbinv-s.m4.in: refrain from using @cartouche for just a few
examples (not consistent w others).
gdb.texinfo: issue disclaimer paragraph on cmdline options only
for generic vn of doc
Roland Pesch [Tue, 18 Aug 1992 01:33:09 +0000 (01:33 +0000)]
Mon Aug 17 18:07:59 1992 Roland H. Pesch (pesch@fowanton.cygnus.com)
* gdb.texinfo, none.m4, h8.m4, gdbinv-s.m4.in: improve H8/300
conditionals; introduce a few generic switches that may be
useful for other cross-dev or dos-hosted configs.
* input-scrub.c (input_scrub_pop, input_scrub_push): memcpy was
being used with args swapped, causing occasional lossage when
refilling buffers after an include file.
Roland Pesch [Mon, 17 Aug 1992 18:28:26 +0000 (18:28 +0000)]
Thanks to Zoo watchfulness:
(1) reflect current name of linker (now "ld", not "gld");
(2) fill in missing portion of a reference to full manual;
(3) fix old random typo spotted while proofing above.
Ken Raeburn [Mon, 17 Aug 1992 16:04:19 +0000 (16:04 +0000)]
* archures.c, cpu-m68k.c, cpu-sparc.c: Undo 16 June changes.
* libaout.h (struct aout_backend_data): New callback "set_sizes".
* aout-target.h (callback): Don't set page/seg sizes.
(mkobject): Ditto.
(set_sizes): New routine.
(backend_data): Point to it.
* aoutf1.h (sunos4_set_sizes): Heed architecture when setting sizes.
(sunos4_aout_backend): Point to it.
* aoutx.h (set_arch_mach): Call set_sizes callback.
Stu Grossman [Sat, 15 Aug 1992 09:34:25 +0000 (09:34 +0000)]
* sparc-stub.c: New file. Mix it with your SPARClite
application, and it will speak GDB remote protocol!
* remote.c (remote_wait): Change 'T' (expedited reply) message to
deal with arbitrary registers. Needed for sparc-stub,
Ken Raeburn [Sat, 15 Aug 1992 02:57:12 +0000 (02:57 +0000)]
* config/tc-m68k.c (m68k_ip): If instruction is invalid for the
selected architecture, print a message saying so and listing what
processors support it, rather than saying "operands mismatch".
Ken Raeburn [Sat, 15 Aug 1992 02:50:47 +0000 (02:50 +0000)]
* config/tc-m68k.c (m68k_ip): If instruction is invalid for the
selected architecture, print a message saying so and listing what
processors support it, rather than saying "operands mismatch".
Per Bothner [Sat, 15 Aug 1992 02:27:58 +0000 (02:27 +0000)]
* aout-target.h: Make _bfd_slurp_extended_name_table be
the default. Given that we *write* the suckers (for long
archive member names), we really ought to be able to read them!
* trad-core.c: Don't include <machine/reg.h>. It doesn't
seem to be needed, and many machines don't have it.
Per Bothner [Fri, 14 Aug 1992 21:23:56 +0000 (21:23 +0000)]
* m68k.h: Merged m68kmri.h and m68k.h (using the former as a
base). Also cleaned up duplicates, re-ordered instructions for
the sake of dis-assembling (so aliases come after standard names).
* m68kmri.h: Now just defines some macros, and #includes m68k.h.
Fred Fish [Fri, 14 Aug 1992 19:14:15 +0000 (19:14 +0000)]
* cplus-dem.c (cplus_markers): Add table for gnu style and
use throughout, in place of compile time constant CPLUS_MARKER.
* cplus-dem.c (ARM_VTABLE_STRING, ARM_VTABLE_STRLEN): Add.
* cplus-dem.c (cfront_special): New function, as suggested
by pfieland@stratus.com.
* cplus-dem.c (forget_types): New function.
* cplus-dem.c (cplus_demangle): Call gnu_special, moved from
demangle_prefix().
* cplus-dem.c (mop_up): Call forget_types().
* cplus-dem.c (AUTO_DEMANGLING, GNU_DEMANGLING, LUCID_DEMANGLING):
Use throughout, instead of checking current_demangling_style.
* cplus-dem.c (demangle_signature): When finding an explicit
start of function args, forget all remembered types for
lucid/cfront style demangling.
* cplus-dem.c (demangle_prefix): In a sequence of two or more
underbar characters, use last pair as the delimiter. Hoist
gnu_special() call up to cplus_demangle(). Call cfront_special()
when appropriate.
* cplus-dem.c (cplus_special): Fix virtual table name demangling
for inherited classes.
* cplus-dem.c (demangle_args): Document quirks of numbered
references to previously seen types.
* dbxread.c (read_ofile_symtab, process_one_symbol):
Use AUTO_DEMANGLING rather than explicitly checking
current_demangling_style.
* demangle.h: Add some comments.
* demangle.h (AUTO_DEMANGLING, GNU_DEMANGLING, LUCID_DEMANGLING,
CFRONT_DEMANGLING): New macros.
* dwarfread.c (LCC_PRODUCER): Remove trailing space, which is
not found in the actual producer string produced by lcc.
* dwarfread.c (handle_producer): Use AUTO_DEMANGLING rather
than explicitly checking current_demangling_style.
John Gilmore [Thu, 13 Aug 1992 19:02:32 +0000 (19:02 +0000)]
* breakpoint.c (enable_longjmp_breakpoint,
disable_longjmp_breakpoint, set_longjmp_resume_breakpoint):
Check for duplicate breakpoints at the same address. Bug report
and preliminary fix from Dave Morrison, <drmorris@us.oracle.com>.
Fred Fish [Wed, 12 Aug 1992 18:17:48 +0000 (18:17 +0000)]
* buildsym.c (end_symtab): Document that end_symtab can return
NULL under non-error conditions.
* dwarfread.c (read_file_scope): Guard against dereferencing NULL
returned from end_symtab for empty compilation units.