Tom Tromey [Mon, 15 Apr 1996 21:14:30 +0000 (21:14 +0000)]
* Makefile.in (ALL_MODULES): Include all-apache.
(CROSS_CHECK_MODULES): Include check-apache.
(INSTALL_MODULES): Include install-apache.
(all-apache): New target.
* configure.in: Added apache everywhere perl is seen.
Ian Lance Taylor [Mon, 15 Apr 1996 16:28:19 +0000 (16:28 +0000)]
* config/tc-mips.c (load_address): Cast X_add_number to valueT
before comparing against MAX_GPREL_OFFSET, so that negative
numbers are handled correctly.
(macro): Likewise.
Stu Grossman [Sat, 13 Apr 1996 21:33:59 +0000 (21:33 +0000)]
* remote-nindy.c (nindy_open): Acquire more target state so that
user can attach to a previously running program.
* (nindy_fetch_registers nindy_store_registers): Get rid of fp
conversion code. That's all handled in {extract store}_floating
now.
* utils.c (floatformat_to_double): Don't bias exponent when
handling zero's, denorms or NaNs.
* config/i960/tm-i960.h (REGISTER_CONVERT_TO_VIRTUAL
REGISTER_CONVERT_TO_RAW): Change to using DOUBLST and
FLOATFORMAT_TO/FROM_DOUBLEST macros.
* config/i960/tm-nindy960.h: Undefine
REGISTER_CONVERT_TO_VIRTUAL, REGISTER_CONVERT_TO_RAW, and
REGISTER_CONVERTIBLE. These are no longer necessary now that all
the magic happens in extract/store_floating.
Peter Schauer [Sat, 13 Apr 1996 10:01:49 +0000 (10:01 +0000)]
* eval.c (evaluate_subexp_standard): Report error when attempting to
evaluate subscripts for types which cannot be subscripted.
* valarith.c (value_x_binop, value_x_unop): Add noside parameter.
Return a zero value with the return type of the member function
if noside is EVAL_AVOID_SIDE_EFFECTS instead of calling the member
function.
* values.h (value_x_binop, value_x_unop): Update prototypes
accordingly.
* eval.c (evaluate_subexp_standard): Update all callers of
value_x_binop, value_x_unop accordingly.
* valarith.c (value_neg, value_complement): Perform ANSI C/C++
integral promotion on operands.
Fred Fish [Sat, 13 Apr 1996 04:38:55 +0000 (04:38 +0000)]
* README: Update for 4.16 release.
* configure.in (AC_CHECK_FUNCS): Also check for sbrk.
* configure: Regenerate with autoconf.
* config.in: Regenerate with autoheader.
* main.c (main): Only use sbrk() when HAVE_SBRK is defined.
* top.c (command_loop): Ditto.
Stu Grossman [Fri, 12 Apr 1996 16:53:08 +0000 (16:53 +0000)]
* defs.h: Define TARGET_{FLOAT DOUBLE LONG_DOUBLE}_FORMAT
defaults for bi-endian targets. Replace function pointers for
floatformat routines with macros. No need for these to be runtime
selectable.
* findvar.c: Get rid of floatformat function pointers. Use
macros in extract_floating and store_floating.
* remote-nindy.c (nindy_fetch_registers nindy_store_registers):
Use floatformat macros.
Fred Fish [Fri, 12 Apr 1996 06:14:46 +0000 (06:14 +0000)]
From: Miles Bader <miles@gnu.ai.mit.edu>
* configure.in (AC_CHECK_HEADERS): check for endian.h.
Use AC_CHECK_TOOL to find AR & RANLIB. Add AC_PROG_AWK.
Add host & target cases for i[345]86-*-gnu*.
* config.in: Regenerate with autoheader.
* configure: Regenerate with autoconf.
* Makefile.in (AR, AWK): Set from corresponding autoconf substs.
(init.c): Don't scan mig-generated files.
* defs.h (endian.h): Include if HAVE_ENDIAN_H defined.
* config/nm-m3.h (ATTACH_NO_WAIT): Define.
* infcmd.c (attach_command): Use "#ifndef ATTACH_NO_WAIT"
rather than "#ifndef MACH".
Ian Lance Taylor [Thu, 11 Apr 1996 21:35:01 +0000 (21:35 +0000)]
* config/tc-sparc.c (last_insn): New static variable.
(md_assemble): Warn about putting floating point branches in a
delay slot. If architecture is less than v9, insert NOP
instructions between floating point instructions and floating
point branches. (The SunOS assembler does both these operations.)
Save the last instruction opcode.
(sparc_ip): Add pinsn parameter. Change caller.
PR 355.
Stu Grossman [Thu, 11 Apr 1996 21:17:45 +0000 (21:17 +0000)]
* dcache.c: Add prototypes. Make many functions static.
* (dcache_peek dcache_fetch dcache_poke): Make dcache_fetch and
dcache_poke call dcache_xfer_memory directly in order to fix
problems with turning off dcache. dcache_peek is now unnecessary,
so it goes away.
* defs.h: Define new macros HOST_{FLOAT DOUBLE LONG_DOUBLE}_FORMAT
and TARGET_{FLOAT DOUBLE LONG_DOUBLE}_FORMAT to specify a pointer
to a struct floatformat. This allows for better handling of
targets whose floating point formats differ from the host by more
than just byte order.
* (floatformat_to_long_double floatformat_from_long_double):
Prototypes for new functions in utils.c.
* (floatformat_to_doublest floatformat_from_doublest): Prototypes
for pointers to floating point conversion functions. The actual
function uses either double or long double if the host supports it.
* findvar.c (floatformat_to_doublest floatformat_from_doublest):
Initialize to point at correct function depending on HAVE_LONG_DOUBLE.
* (extract_floating store_floating): Rewrite. Now, if host fp
format is the same as the target, we just do a copy. Otherwise,
we call floatformat_{to from}_doublest.
* remote-nindy.c (nindy_xfer_inferior_memory): Change param
`write' to `should_write'.
* utils.c (floatformat_to_long_double
floatformat_from_long_double): New routines that implement long
double versions of functions in libiberty/floatformat.c.
* config/i960/tm-i960.h (TARGET_LONG_DOUBLE_FORMAT): Define this for
i960 extended real (80 bit) numbers.
* nindy-share/nindy.c (ninMemGet ninMemPut): Return number of bytes
actually read or written.
Jeff Law [Wed, 10 Apr 1996 20:40:24 +0000 (20:40 +0000)]
* scripttempl/h8300.sc: Add the "8-bit area" in the upper 256
bytes of the address space. Put data from the ".eight" sections
into the 8-bit area.
* scripttempl/h8300h.sc: Likewise.
8-bit area changes for hmse.
Ian Lance Taylor [Wed, 10 Apr 1996 18:21:08 +0000 (18:21 +0000)]
* versados.c (versados_scan): Check bfd_alloc return value.
(versados_object_p): Check all bfd_read return values. Require
that lang field be less than 10, to avoid confusion with Intel Hex
files.
Ian Lance Taylor [Wed, 10 Apr 1996 18:11:00 +0000 (18:11 +0000)]
* libaout.h (WRITE_HEADERS): Write out the relocs even if there
aren't any symbols.
* aoutx.h (NAME(aout,swap_std_reloc_out)): Set r_index to N_ABS,
not 0, for an absolute symbol.
(NAME(aout,swap_ext_reloc_out)): Likewise.
PR 7387.
Ian Lance Taylor [Wed, 10 Apr 1996 17:01:35 +0000 (17:01 +0000)]
* ihex.c (ihex_scan): Accept a length of 4 for record type 5.
(ihex_write_object_contents): For a large start address, output
the full 32 bit address in record type 5.
PR 9422.
* ch-valprint.c (chill_val_print): Remove call to calculate_array_length.
(calculate_array_length): Move function from here ...
* ch-exp.c (calculate_array_length): ... to here.
(parse_primval): If we have a symbol with an array type
and the length is 0, call calculate_array_length.
* aoutx.h (aout_link_add_symbols): Always call add_dynamic_symbols
entry point, not just for DYNAMIC objects.
* sunos.c (sunos_add_dynamic_symbols): Always call
sunos_create_dynamic_sections. If called with a non DYNAMIC
object, don't do anything else.
(sunos_add_one_symbol): Don't call sunos_create_dynamic_sections.
Fred Fish [Tue, 9 Apr 1996 06:09:33 +0000 (06:09 +0000)]
* Makefile.in (scm-exp.o, scm-lang.o, scm-valprint.o): Add targets and
dependencies.
* scm-lang.c (gdb_string.h): Include.
* objfiles.c (add_to_objfile_sections): Cast second arg of obstack_grow
call to correct type (char *).
* cp-valprint.c (cp_print_static_field): Ditto.
* somsolib.c (som_solib_create_inferior_hook): Add a declaration
for external find_unwind_entry function (from hppa-tdep.c).
* remote-pa.c (remote_write_bytes, remote_read_bytes): Change
type of second arg to "char *" to be type compatible with
dcache.
(remote_wait): Cast second arg to strtol to correct type.
* hppa-tdep.c (compare_unwind_entries): Change argument types to
"const void *" to be type compatible with qsort, and then
assign to local args prior to use.
Jeff Law [Mon, 8 Apr 1996 21:34:12 +0000 (21:34 +0000)]
* infptrace.c (kill_inferior): Remove call to "kill"; update
comments.
The "kill" call was declared bogus years ago and is causing problems
on hpux10 (PR 9410).
Modified remote-e7000.c so that :23 isn't appended to the target port
if __WIN32__ is defined. This is necessary to make e7000 support work
for WinGDB.
From: Robert Lipe <robertl@dgii.com>
Add support for SCO OpenServer 5 (a.k.a. 3.2v5*) This
target is an SVR3.2 with COFF, ELF, and shared libes, but
no /proc.
* config/i386/i386sco5.mh: New file.
* config/i386/nm-i386sco5.h: New file.
Fred Fish [Sat, 6 Apr 1996 10:19:19 +0000 (10:19 +0000)]
* symfile.c (INLINE_ADD_PSYMBOL): Remove ifdef.
(add_psymbol_to_list): Add an arg for passing CORE_ADDR values and
use it, rather than calling add_psymbol_addr_to_list.
(add_psymbol_addr_to_list): Delete.
(add_psymbol_to_list): Make psymbol static to avoid random data in
gaps due to alignment of structure members.
* symfile.h (INLINE_ADD_PSYMBOL, ADD_PSYMBOL_TO_LIST,
ADD_PSYMBOL_ADDR_TO_LIST): Remove. Real world tests show no
performance improvements by inlining via complicated macros and
they just make gdb larger and harder to maintain.
* dwarfread.c (add_enum_psymbol): Replace ADD_PSYMBOL_TO_LIST
and/or ADD_PSYMBOL_ADDR_TO_LIST macro(s) with call to
add_psymbol_to_list with appropriate long or CORE_ADDR args.
(add_partial_symbol): Ditto.
* partial-stab.h: Ditto.
* os9kread.c (read_os9k_psymtab): Ditto
* mdebugread.c (parse_partial_symbols): Ditto.
(handle_psymbol_enumerators): Ditto.
(demangle.h): Include.
* hpread.c (hpread_build_psymtabs): Ditto.
(hpread_build_psymtabs): Ditto.
(demangle.h): Include
Jeff Law [Sat, 6 Apr 1996 00:21:35 +0000 (00:21 +0000)]
* compile.c (decode): Use "bit" to hold L_3 immediates instead
of holding them in "abs". Handle ABS8MEM memory references aka
8-bit area. Replace ABSMOV references with ABS8MEM.
So we've got a chance of simulating something like btst #0,@40:8 correctly.
hmse.
Jeff Law [Sat, 6 Apr 1996 00:14:04 +0000 (00:14 +0000)]
* h8300-dis.c (bfd_h8_disassemble): Use "bit" for L_3 immediates,
not "abs", which may be needed for the absolute in something
like btst #0,@10:8. Print L_3 immediates separately from other
immediates. Change ABSMOV reference to ABS8MEM.
One day we'll actually disassemble btst #0,@10:8 correctly... But not
yet. hmse.
Jeff Law [Fri, 5 Apr 1996 23:56:39 +0000 (23:56 +0000)]
* h8300.h (ABS8MEM): Renamed from ABSMOV. Remove ABSMOV
from all insns.
(ABS8SRC,ABS8DST): Add ABS8MEM.
(add.l): Fix reg+reg variant.
(eepmov.w): Renamed from eepmovw.
(ldc,stc): Fix many cases.
We'll use ABS8MEM for access to the 8-bit area.
Fixes from running the h8300 testsuite.
hmse.
Jeff Law [Fri, 5 Apr 1996 23:50:28 +0000 (23:50 +0000)]
* gas/h8300/{addsubh.s,bitops1h.s,bitops2h.s}: New h8300h tests.
* gas/h8300/{bitops3h.s,bitops4h.s,branchesh.s}: New h8300h tests.
* gas/h8300/{cbranchh.s,compareh.s,decimalh.s}: New h8300h tests.
* gas/h8300/{divmulh.s,incdech.s,logicalh.s}: New h8300h tests.
* gas/h8300/{misch.s,movbh.s,movwh.s}: New h8300h tests.
* gas/h8300/{pushpoph.s,rotshifth.s}: New h8300h tests.
* gas/h8300/h8300.exp: Run them.
More tests.
Stan Shebs [Fri, 5 Apr 1996 21:52:38 +0000 (21:52 +0000)]
* gdbtk.c (running_now): New global variable.
(gdb_cmd): Test it before executing any command.
(gdbtk_call_command): Set it when inferior is running.
* gdbtk.tcl (gdbtk_tcl_busy, gdbtk_tcl_idle): Enable and
disable interaction with command window's text appropriately.
* emultempl/elf32.em (gld${EMULATION_NAME}_check_ld_so_conf): New
static function, if ${host} = ${target}.
(gld${EMULATION_NAME}_after_open): Call check_ld_so_conf to find a
needed shared library if ${host} = $[target}.
Fri Apr 5 15:13:10 1996 Andreas Schwab <schwab@issan.informatik.uni-dortmund.de>
* config/atof-ieee.c: Fix handling of denormalized extended
precision numbers and overflow/underflow detection.
(MAX_PRECISION, X_PRECISION, P_PRECISION): Changed from 6 to 5, to
not include the 16 bit gap in the m68k extended precision format.
Fri Apr 5 18:11:25 1996 James G. Smith <jsmith@cygnus.co.uk>
* emulparams/elf32{b,l}4300.sh (MACHINE): Add explicit
architecture number.
* scripttempl/elf.sc: Use $MACHINE definition if present.
* configure.tgt (targ_extra_emuls): Force 4100 build to use same
template as 4300.
Jeff Law [Fri, 5 Apr 1996 17:13:28 +0000 (17:13 +0000)]
* gas/h8300/{addsub.s,bitops1.s,bitops2.s}: New h8300 tests.
* gas/h8300/{bitops3.s,bitops4.s,branches.s}: New h8300 tests.
* gas/h8300/{cbranch.s,compare.s,decimal.s}: New h8300 tests.
* gas/h8300/{divmul.s,incdec.s,logical.s}: New h8300 tests.
* gas/h8300/{misc.s,movb.s,movw.s}: New h8300 tests.
* gas/h8300/{pushpop.s,rotshift.s}: New h8300 tests.
* gas/h8300/h8300.exp: Run them.
To keep me from screwing up the assembler...
Jason Molenda [Fri, 5 Apr 1996 11:18:57 +0000 (11:18 +0000)]
* Makefile.in (BASE_FLAGS_TO_PASS): pass down $(MAKE).
w/o this, if you use GNU make on a Solaris platform (eg), but you don't
have GNU make at the front of your path, when the make/ subdir cds down
into make/glob and does a make there, you will end up picking up the
first make that is in your path no matter what you do.
No PR, I just noticed this while I was working on something else.
Fred Fish [Fri, 5 Apr 1996 04:22:08 +0000 (04:22 +0000)]
* configure.in: Check for setpgid function.
* config.in: Regenerate with autoheader.
* configure: Regenerate with autoconf.
* inflow.c (_initialize_inflow): Only try to use _SC_JOB_CONTROL
if it is actually defined.
(gdb_setpgid): Use HAVE_SETPGID.
* ch-exp.c: Change include of <string.h> to "gdb_string.h".
* c-exp.y: Ditto.
* f-exp.y: Ditto.
* m2-exp.y: Ditto.
* c-exp.y: Include <ctype.h>.
* serial.c: Ditto.
* config/m68k/nm-news.h: Add typedef for pid_t which is
apparently missing from <sys/types.h>. Enclose entire
file in NM_NEWS_H ifndef and define when included.
* config/mips/nm-news-mips.h: Ditto.
* config/m68k/tm-m68k.h (REGISTER_CONVERT_TO_VIRTUAL,
REGISTER_CONVERT_TO_RAW): Change name of temporary variable.
Fred Fish [Fri, 5 Apr 1996 04:14:36 +0000 (04:14 +0000)]
* configure.in (mips-sony-bsd*): Fix typo in AC_DEFINE.
(AC_CHECK_HEADERS): Check for sys/time.h
(AC_HEADER_TIME): Add macro.
* configure: Rebuild.
* sysdep.h: Use TIME_WITH_SYS_TIME and HAVE_SYS_TIME_H
to control what combination of <time> and <sys/time.h>
get included.