Ian Lance Taylor [Wed, 17 Nov 1993 23:06:12 +0000 (23:06 +0000)]
* write.c (is_dnrange): Do not define if md_relax_frag is defined.
(relax_segment): If md_relax_frag is defined, use it to handle a
frag of type rs_machine_dependent rather than looking through
md_relax_table.
* write.c (relax_and_size_seg): Don't set SEC_HAS_CONTENTS for a
bss section. Set SEC_RELOC if there are any relocations, even for
a zero size section.
* write.c (write_relocs): In RELOC_EXPANSION_POSSIBLE case, based
data offset on reloc[0]->address rather than reloc[j]->address, so
that multiple relocs can affect different memory locations.
* write.c (chain_frchains_together, relax_and_size_seg,
adjust_reloc_syms, write_relocs): Make third argument PTR, not
char *, to match definition of bfd_map_over_sections.
Ian Lance Taylor [Wed, 17 Nov 1993 23:03:42 +0000 (23:03 +0000)]
* read.c (read_a_source_file): If we find a bad pseudo-op,
do a continue to go on to the next line rather than a break.
Removed duplicate bad pseudo-op code which was never executed.
* read.c (s_lcomm): Do not require a comma after the name.
* read.c (s_lcomm): Set bss flag for .sbss section if used.
Kung Hsu [Wed, 17 Nov 1993 18:07:58 +0000 (18:07 +0000)]
Modified Files:
ChangeLog eval.c valops.c
* valops.c (typecmp): improve prototype matching when calling
a method. Make 'p (*(ostream *) &cout) << "lll" ' to work.
* eval.c(evalute_subexp): fix operator search problem when call
like p x.'operator+'(i).
Stu Grossman [Wed, 17 Nov 1993 07:56:28 +0000 (07:56 +0000)]
* i386ly-nat.c, i386lynx-nat.c, m68kly-nat.c: Remove. Move
common code into lynx-nat.c.
* lynx-nat.c: New module. Contains portable code for Lynx native
stuff (mostly ptrace related).
* config/i386/i386lynx.mh (NATDEPFILES): i386ly-nat.o -> lynx-nat.o
* config/m68k/m68klynx.mh (NATDEPFILES): i386ly-nat.o -> lynx-nat.o
* config/nm-lynx.h, config/tm-lynx.h: New files to contain
non-architecture specific native and target defs.
* config/i386/nm-i386lynx.h, config/i386/tm-i386lynx.h,
config/m68k/nm-m68klynx.h, config/m68k/tm-m68klynx.h: Move all
(arch) portable stuff into ../{tm nm}-lynx.h.
Stu Grossman [Wed, 17 Nov 1993 07:17:45 +0000 (07:17 +0000)]
* i386ly-nat.c, i386lynx-nat.c, m68kly-nat.c: Remove. Move
common code into lynx-nat.c.
* lynx-nat.c: New module. Contains portable code for Lynx native
stuff (mostly ptrace related).
* config/i386/i386lynx.mh (NATDEPFILES): i386ly-nat.o -> lynx-nat.o
* config/m68k/m68klynx.mh (NATDEPFILES): i386ly-nat.o -> lynx-nat.o
* config/nm-lynx.h, config/tm-lynx.h: New files to contain
non-architecture specific native and target defs.
* config/i386/nm-i386lynx.h, config/i386/tm-i386lynx.h,
config/m68k/nm-m68klynx.h, config/m68k/tm-m68klynx.h: Move all
(arch) portable stuff into ../{tm nm}-lynx.h.
Jim Kingdon [Wed, 17 Nov 1993 04:24:34 +0000 (04:24 +0000)]
* symtab.c (gdb_mangle_name): Only assume that the physname is
the entire mangled name if it looks like the mangled name of a
constructor. Needed for testsuite to work with GCC 2.4.5.
Stu Grossman [Wed, 17 Nov 1993 01:38:11 +0000 (01:38 +0000)]
* i386ly-nat.c, i386lynx-nat.c, m68kly-nat.c: Remove. Move
common code into lynx-nat.c.
* lynx-nat.c: New module. Contains portable code for Lynx native
stuff (mostly ptrace related).
* config/i386/i386lynx.mh (NATDEPFILES): i386ly-nat.o -> lynx-nat.o
* config/m68k/m68klynx.mh (NATDEPFILES): i386ly-nat.o -> lynx-nat.o
* config/nm-lynx.h, config/tm-lynx.h: New files to contain
non-architecture specific native and target defs.
* config/i386/nm-i386lynx.h, config/i386/tm-i386lynx.h,
config/m68k/nm-m68klynx.h, config/m68k/tm-m68klynx.h: Move all
(arch) portable stuff into ../{tm nm}-lynx.h.
Jim Kingdon [Tue, 16 Nov 1993 20:29:03 +0000 (20:29 +0000)]
* a68v-nat.c: Replace with new version from Troy Rollo. The
version I am replacing appears to be an old copy of sun3-nat.c.
* dstread.c (dst_symfile_read): Replace sort_all_symtab_syms call
with loop.
Jim Kingdon [Tue, 16 Nov 1993 19:51:59 +0000 (19:51 +0000)]
* Makefile.in (TAGS): Depend on TAGFILES_{NO,WITH}_SRCDIR.
* Makefile.in: (HFILES,TAGFILES): Split into _WITH_SRCDIR and
_NO_SRCDIR versions.
(TAGS): Only add srcdir to TAGFILES_NO_SRCDIR.
(This is part of a long saga involving me putting srcdir on
everything (perhaps for now-obsolete reasons, I forget), Rich
removing the srcdir from everything, Stu putting it back some
places for Sun make, and me just now getting around to fixing
`make TAGS' again).
* remote-e7000.c: New file to cope with the Hitachi E7000 ICE.
* remote-utils.c, remote-utils.h (gr_load_image): New function to
download to target.
* h8300-tdep.c, h8500-tdep.c, remote-z8k.c, sh-tdep.c z8k-tdep.c
(sim_load): delete.
* remote-sim.c (sim_load): Use gr_load_image.
* config/sh/sh.mt: Add remote-e7000
Jim Kingdon [Mon, 15 Nov 1993 20:13:30 +0000 (20:13 +0000)]
* gdbtypes.h: Add TYPE_FLAG_TARGET_STUB.
* gdbtypes.c (check_stub_type): On TYPE_FLAG_TARGET_STUB, do
what cleanup_undefined_types does for arrays, except we clear
TYPE_FLAG_TARGET_STUB if we fix up the type.
* stabsread.c (cleanup_undefined_types): Add comments about how
doing arrays here is no longer the clean way to do it.
(read_array_type): Set TYPE_FLAG_TARGET_STUB as well as calling
add_undefined_type.
* c-typeprint.c, ch-typeprint.c: Move call to check_stub_type
outside switch so it happens for all type codes.
* cp-valprint.c (cp_print_value_fields): Recurse to val_print,
instead of c_val_print, so that check_stub_type gets called.
* gdbtypes.h, gdbtypes.c, m2-lang.c, ch-lang.c, c-lang.c: Remove
TYPE_FLAG_SIGNED. It was inconsistently set, never checked
(except in recursive_dump_type), and is pointless.
Jeff Law [Mon, 15 Nov 1993 08:43:36 +0000 (08:43 +0000)]
* som.c (som_object_setup): Do not create dummy ".text", ".data",
and ".bss" sections.
(setup_sections): Do not set SEC_HAS_CONTENTS if a section's size
is zero. Recognize BSS type sections and turn off SEC_LOAD and
SEC_DATA (so binutils/size works). Set the correct value for
a section's _raw_size.
(som_slurp_symbol_table): Program entry points, and millicode are
also functions. Mark them as such. Also mark L$* symbols as
debugging symbols.
Jim Kingdon [Mon, 15 Nov 1993 05:36:36 +0000 (05:36 +0000)]
* stabsread.c: Remove #if 0'd dbx_class_complaint. We now handle
this (more or less) gracefully, and complain() was never a good
way of dealing with this.
Jim Kingdon [Sun, 14 Nov 1993 06:40:47 +0000 (06:40 +0000)]
* stabsread.c (read_type): Skip the colon when reading a
cross-reference. Only complain, not error_type, on unrecognized
cross-reference types. error_type, not dump core, if the colon is
missing.
Jeff Law [Fri, 12 Nov 1993 23:43:24 +0000 (23:43 +0000)]
* som.c (struct reloc_queue): New structure to keep track of
the last four multibyte relocations emitted.
(enum pa_symbol_type): Type to fully describe the symbol types
associated with .import/.export assembler directives.
Jeff Law [Fri, 12 Nov 1993 23:08:13 +0000 (23:08 +0000)]
* som.h (struct som_symbol): Add new fields to hold additional
information needed to build/write symbol tables and fixup streams.
(struct som_section_data_struct): Add new fields to hold additional
information needed to build/write space and subspace headers.
(som_symbol_data): New accessor macro for SOM symbol information.
(R_HPPA_*): Basic relocation types to be used by the assembler.
Jeff Law [Fri, 12 Nov 1993 22:54:41 +0000 (22:54 +0000)]
* config/tc-hppa.c (hppa_fix_struct): Use a real type for the
field selector, rather than an int. All uses of field selectors
fixed.
(tc_gen_reloc): For SOM PLABELs, always set addend to zero for now.
(md_apply_fix_1): Do not call hppa_field_adjust for any PLABEL
field.
Kung Hsu [Fri, 12 Nov 1993 19:34:24 +0000 (19:34 +0000)]
Modified Files:
ChangeLog stabsread.c
* stabsread.c (patch_block_stabs, define_symbol, read_type): in
g++ template instantiation, nested class can be part of the
params, and '::' can gets into symbol or type names. This is
to fix the problem. (fix pr3837)
Stu Grossman [Fri, 12 Nov 1993 18:59:43 +0000 (18:59 +0000)]
* coff-solib.c (coff_solib_add): Cast result of alloca().
* m68k-tdep.c (m68k_saved_pc_after_call): Get rid of
GDB_TARGET_IS_SUN3. Use more general SYSCALL_TRAP macro.
* config/m68k/m68klynx.mh (NATDEPFILES): Remove exec.o (it's
already in TDEPFILES).
* config/m68k/tm-m68k.h (SAVED_PC_AFTER_CALL): Use
m68k_saved_pc_after_call.
* Remove all Sun3 specific stuff.
* (FIX_CALL_DUMMY): Cast arg to bfd_putb32 to unsigned char *.
* config/m68k/tm-m68klynx.h: Define SYSCALL_TRAP as trap #10.
Disable REMOTE_BREAKPOINT mechanism.
* config/m68k/tm-sun3.h: Get rid of GDB_TARGET_IS_SUN3.
* Protect from multiple includion.
* Move Sun3 specific stuff from tm-m68k.h to here.
* Define SYSCALL_TRAP as trap #0.
* Remove def of SAVED_PC_AFTER_CALL (now in tm-m68k.h).
* gdbserver/low-lynx.c: Redo all register store/fetch stuff to
make it portable for 386 and 68k.
Jim Kingdon [Fri, 12 Nov 1993 17:10:56 +0000 (17:10 +0000)]
* trad-core.c (trad_unix_core_file_p): If new hook
TRAD_CORE_ALLOW_ANY_EXTRA_SIZE defined, then skip the check for the
corefile being too big.
* hosts/i386sco.h: Define it.
Jim Kingdon [Fri, 12 Nov 1993 16:35:59 +0000 (16:35 +0000)]
* mips-tdep.c (init_extra_frame_info): Check to see whether the
registers mentioned in the proc_desc have been saved. This
generalizes mips_in_lenient_prologue in the sense that we keep
searching until we've found saves for all the registers, not just
look for a "lenient prologue" pattern.
* mips-tdep.c: #if 0 lenient prologue code.
* mips-tdep.c (heuristic_proc_desc): Don't assume a host short
is 16 bits.