* dwarf2read.c (dwarf2_compute_name): Check that the first
argument is a pointer.

* dwarf2read.c (read_func_scope, read_structure_type)
(read_common_block): Check for a NULL return from new_symbol.

2010-09-08  Rafael Espindola  <espindola@google.com>

* Makefile.am (memory_test.stdout): Run readelf with -W
* Makefile.in: Regenerate.
* testsuite/memory_test.sh: Make the regexps accept both 32 and 64 bit outputs.

2010-09-08  Rafael Espindola  <espindola@google.com>

* script-sections.cc (Script_sections::add_memory_region): Convert
field precision to int.
* script.cc (script_set_section_region, script_set_section_region):
Convert field precision to int.

gdb/
* dwarf2read.c (dwarf2_read_index): Return on no SEC_HAS_CONTENTS.

gdb/
* dwarf2read.c (read_structure_type): Move processing of
fields and member functions from here...
(process_structure_scope): ... to here.

gdb/testsuite/
* gdb.cp/templates.cc (Empty, FunctionArg): New classes.
(FunctionArg::method): New function.
(empty, arg): New variables.
(main): Call arg.method.
* gdb.cp/templates.exp (test_template_args): New function.
(do_tests): Call it.

* gnu-v3-abi.c (gnuv3_print_method_ptr): Do not use
the domain type.
(gnuv3_make_method_ptr): Likewise.

2010-09-08  Rafael Espindola  <espindola@google.com>

* arm.cc (do_finalize_sections): Create the __exidx_start and
__exdix_end symbols even when the section is missing.

* breakpoint.c (create_breakpoint): Add missing _().

* dwarf2-frame.h (dwarf2_frame_build_info): Delete, unused.

* gdb.threads/threxit-hop-specific.exp: Use "continue" instead
of "next" to proceed over pthread_exit call.

* config/tc-arm.c (create_neon_reg_alias): Deal with case
sensitivity.

* gdb.base/maint.exp: Avoid wildcards against potentially very long
"maint print statistics" output to prevent expect timeouts.

Revert:
2010-09-01  Tom Tromey  <tromey@redhat.com>
* dwarf2read.c, gdbtypes.c, psymtab.c, symfile.h, symtab.c,
symtab.h: Revert earlier change.

    * README: Remove claim that MEMORY is not supported.
        * expression.cc (script_exp_function_origin)
        (script_exp_function_length): Move from here to ...
        * script.cc: ... here.
        (script_set_section_region, script_add_memory)
        (script_parse_memory_attr, script_include_directive): New
        functions.
        * script-sections.cc
        (class Memory_region): New class.
        (class Output_section_definition): Add set_memory_region,
        set_section_vma, set_section_lma and get_section_name methods.
        (class Script_Sections): Add add_memory_region,
        find_memory_region, find_memory_region_origin,
        find_memory_region_length and set_memory_region methods.
        Have set_section_addresses method walk the list of set memory
        regions.
        Extend the print methos to display memory regions.
        * script-sections.h: Add prototypes for new methods.
        Add enum for MEMORY region attributes.
        * yyscript.y: Add support for parsing MEMORY regions.
        * script-c.h: Add prototypes for new functions.
        * testsuite/Makefile.am: Add test of MEMORY region functionality.
        * testsuite/Makefile.in: Regenerate.
        * testsuite/memory_test.sh: New script.
        * testsuite/memory_test.s: New assembler source file.
        * testsuite/memory_test.t: New linker script.

    PR gas/11973
        * config/tc-mn10300.c (md_convert_frag): Zero out top two bytes of
        long call instruction's displacement.

*** empty log message ***

daily update

merge from gcc

Add missing files.

    * ldwrite.c (clone_section): Call bfd_copy_private_section_data on
        newly cloned section.
        * ld-x86-64/split-by-file.rd: New test.
        * ld-x86-64/split-by-file1.s: New test source code.
        * ld-x86-64/split-by-file2.s: New test source code.
        * ld-x86-64/x86-64.exp: Run the new test.
        * readelf.c (process_section_headers): Mention meaning of 'l'
        section flag for x86-64 targets.

Fix custom bcache hash regression.

2010-09-09  Sami Wagiaalla  <swagiaal@redhat.com>

* psymtab.c (add_psymbol_to_bcache): Initialize
obj_section.
memset psymbol.ginfo.value to 0.

*** empty log message ***

daily update

* peXXigen.c: Whitespace.
(pe_print_idata): Correct section limit calculations.  Tidy array
indexing.
(_bfd_XX_print_ce_compressed_pdata): Don't leak memory.

gdb/testsuite/
* gdb.base/printcmds.exp (test_float_accepted): Import gdb_prompt.
Move here, negate and extend by XFAIL the "p 0x1.1" test from ...
(test_float_rejected): ... here.
* gdb.java/jv-print.exp (test_float_accepted): Import gdb_prompt.
Move here, negate and extend by XFAIL the "p 0x1.1" test from ...
(test_float_rejected): ... here.
* gdb.objc/print.exp (test_float_accepted): Import gdb_prompt.
Move here, negate and extend by XFAIL the "p 0x1.1" test from ...
(test_float_rejected): ... here.
* gdb.pascal/print.exp (test_float_accepted): Import gdb_prompt.
Move here, negate and extend by XFAIL the "p 0x1.1" test from ...
(test_float_rejected): ... here.

Sync bootstrap-lto.mk and dfp.m4 with gcc.

2010-09-06  Yao Qi  <yao@codesourcery.com>

* linux-low.c (linux_kill): Remove unused variable.
(linux_stabilize_threads): Likewise.
* server.c (start_inferior): Likewise.
(queue_stop_reply_callback): Likewise.
* tracepoint.c (do_action_at_tracepoint): Likewise.

2010-09-06  Yao Qi  <yao@codesourcery.com>

* linux-low.c (maybe_move_out_of_jump_pad): Restore current_inferior
on return.

* infrun.c (resume): Extend comment on ignoring single-step
requests on vfork parents waiting for a vfork-done.

gdb/
* corelow.c (core_open): Use target_signal_from_host if CORE_GDBARCH
is NULL.
* fork-child.c (startup_inferior) <resume_signal>: Use enum
target_signal type.
* linux-nat.c (linux_nat_resume): Use target_signal_to_host before
calling strsignal.  Use enum target_signal type for saved_signo.
(linux_handle_extended_wait) <signo>: Use enum target_signal type.
(linux_nat_wait_1): Use enum target_signal type for signo.  Use
target_signal_to_host before calling strsignal.
* remote-m32r-sdi.c (m32r_wait, m32r_detach): Replace 0 by
TARGET_SIGNAL_0.

gdb/gdbserver/
* target.c (mywait) <TARGET_WAITKIND_EXITED>: Fix to use INTEGER.

gdb/
* common/signals.c (ANY): Remove.
(SET): No longer use ANY.

include/gdb/
* signals.def: Replace all ANY uses by SET with specific numbers.
* signals.h (ANY): Remove.

2010-09-06  Yao Qi  <yao@codesourcery.com>

* infrunc(resume): When inferior is waiting_for_vfork_done,
clear step and don't use displaced stepping.

* Makefile.in (install-only): Replace $IPA_DEPFILES with
"$(IPA_DEPFILES)".

daily update

*** empty log message ***

*** empty log message ***

daily update

* breakpoint.c (can_use_hardware_watchpoint): Handle the first
value specially.

testsuite/
* gdb.base/watchpoint.exp: Call test_inaccessible_watchpoint
before disabling hardware watchpoints.
(test_inaccessible_watchpoint): Check that hardware watchpoints
are used.  Test for watchpoints on a constant address.

gdb/
Code cleanup - split print_stop_reason.
* infrun.c (enum inferior_stop_reason): Remove.
(print_stop_reason): Remove by splitting into ...
(print_exited_reason, print_signal_exited_reason)
(print_no_history_reason, print_signal_received_reason)
(print_end_stepping_range_reason): ... these new functions.  Update
the preceding comment.
(handle_inferior_event): Change the calls to print_exited_reason,
print_signal_exited_reason, print_no_history_reason,
print_signal_received_reason, print_end_stepping_range_reason.
(handle_step_into_function, handle_step_into_function_backward):
Change the calls to print_end_stepping_range_reason.

*** empty log message ***

daily update

Check flag_code instead of use_rela_relocations for 64bit.

gas/

2010-09-03  H.J. Lu  <hongjiu.lu@intel.com>

PR gas/11974
* config/tc-i386.c (i386_finalize_immediate): Check flag_code
instead of use_rela_relocations for 64bit.

gas/testsuite/

2010-09-03  H.J. Lu  <hongjiu.lu@intel.com>

PR gas/11974
* gas/i386/immed64.s: Add more movabs tests.
* gas/i386/immed64.d: Updated.

binutils/
* dwarf.c (regname): New declaration.
(decode_location_expression): Print for registers also regname output.

binutils/testsuite/
* binutils-all/objdump.W: Update DW_OP_reg5 expected output.

gas/testsuite/
* gas/elf/dwarf2-1.d: Update DW_OP_reg5 expected output.
* gas/elf/dwarf2-2.d: Likewise.
* gas/i386/dw2-compress-1.d: Likewise.

gdb/
* breakpoint.c (save_breakpoints): Use RETURN_MASK_ALL.
* cli-out.c: Include vec.h.
(cli_field_fmt, cli_spaces, cli_text, cli_message, cli_flush): New
variable stream, initialize it, use it.
(cli_redirect): New function comment.  Replace the stream and
original_stream fields by the new streams field.  Remove the
original_stream != NULL conditional, assert error on NULL instead.
(out_field_fmt, field_separator): New variable stream, initialize it, use it.
(cli_out_data_ctor): Assert non-NULL stream.  Replace the stream and
original_stream fields by the new streams field.
(cli_out_set_stream): Replace the stream field by the new streams
field.
* cli-out.h: Include vec.h.
(ui_filep): New typedef, call DEF_VEC_P for it.
(struct cli_ui_out_data): Replace the stream and original_stream
fields by the new streams field.
* cli/cli-logging.c (set_logging_redirect): Call ui_out_redirect with
NULL first.  Extend the comment.
(handle_redirections): Call ui_out_redirect with output.
* python/py-breakpoint.c (bppy_get_commands): Move ui_out_redirect
calls outside of the TRY_CATCH block.

gdb/testsuite/
* gdb.base/ui-redirect.exp: New file.

    * gas/cfi/cfi-i386.d: Use objdump -Wf instead of readelf.
        * gas/cfi/cfi-i386.s: Remove .type directives.
        * gas/cfi/reloc-pe-i386.d: Adjust test for i386.

libdecnumber/

2010-09-03  Andreas Krebbel  <Andreas.Krebbel@de.ibm.com>

    * configure.ac: Use the GCC_AC_ENABLE_DECIMAL_FLOAT macro.
    * Makefile.in: Add aclocal.m4 dependency to dfp.m4.
    * configure: Regenerate.
    * aclocal.m4: Regenerate.
config/

2010-09-03  Andreas Krebbel  <Andreas.Krebbel@de.ibm.com>

    * dfp.m4: New file.

GDB 7.2 released.

daily update

*** empty log message ***

Select 32-bit pei for 64-bit pei configurations.

Back out the change introducing the "record pic" feature.

gdb/ChangeLog:

Back out the following change:
| 2010-06-29  Hui Zhu  <teawater@gmail.com>
| * record.c (set_record_pic_cmdlist,
| show_record_pic_cmdlist): New variables.
| (set_record_pic_command,
| show_record_pic_command): New functions.
| (record_pic_function, record_pic_line, record_pic_enum,
| set_record_pic_type, record_pic_hide_nofunction,
| record_pic_hide_nosource, record_pic_hide_same): New variables.
| (record_pic_fputs): New function.
| (function_list, node_list, edge_list): New struct.
| (function_list, node_list, edge_list): New variables.
| (record_pic_cleanups, record_pic_node,
| record_pic_edge, cmd_record_pic): New functions.
| (_initialize_record): Add new commands for record pic.

* gas/cfi/reloc-pe-i386.d, gas/cfi/reloc-pe-i386.s: New test.

* dw2gencfi.c (TC_DWARF2_EMIT_OFFSET): Provide default.
(output_fde): Use it.  Make sure to fully init exp before using it.
testsuite/
* gas/cfi/cfi-common-1.d: Use objdump instead of readelf to dump.
* gas/cfi/cfi-common-2.d, gas/cfi/cfi-common-3.d,
gas/cfi/cfi-common-4.d, gas/cfi/cfi-common-6.d,
gas/cfi/cfi-common-7.d, gas/cfi/cfi-x86_64.d: Likewise.
* gas/cfi/cfi-x86_64.s: Remove .type directives.
* gas/cfi/cfi.exp: Run for pecoff objects too.
* gas/cfi/reloc-pe-i386.d, gas/cfi/reloc-pe-i386.s: New test.

* dwarf.c (init_dwarf_regnames_i386, init_dwarf_regnames_x86_64): New.
(init_dwarf_regnames): Use them.
* dwarf.h: Declare them.
* objdump.c (dump_dwarf): Use bfd_get_arch + bfd_get_mach to set up
the regnames, rather than using elf_machine_code.

    * objdump.c (dump_dwarf): Use bfd_arch_bits_per_address.

Fix ChangeLog entries cli-logging.c -> cli/cli-logging.c .

* config.in, configure: Regenerated.
* configure.ac: Check for waitpid.
* ser-pipe.c (pipe_close): Wait for the program to exit.

gdb/
* cli-logging.c: Include gdb_assert.h.
(set_logging_overwrite): New function.
(logging_redirect): New comment.
(logging_no_redirect_file, set_logging_redirect)
(pop_output_files) <logging_no_redirect_file>: New.
(handle_redirections) <!logging_redirect>: New variable
no_redirect_file.  Remove file autoclose for tee_file_new.  No longer
discard cleanup for the close of former OUTPUT.  Set
LOGGING_NO_REDIRECT_FILE.
(handle_redirections) <logging_redirect>: gdb_assert
LOGGING_NO_REDIRECT_FILE.
(show_logging_command) <logging_redirect handling>: Adjust messages
for SAVED_FILENAME not NULL.
(_initialize_cli_logging): Install set_logging_overwrite and
set_logging_redirect.

Add test for -file-list-exec-source-files

gdb/testsuite/ChangeLog:

        * gdb.dwarf/dw2-filename.S, gdb.dwarf/dw2-filename.exp: New files.

gdb/
* iq2000-tdep.c (iq2000_scan_prologue): Initialize SAL.END.

gdb/
* Makefile.in (GDB_WARN_CFLAGS_NO_FORMAT): Convert it to the no- form.
(monitor.o): Replace $(INTERNAL_WARN_CFLAGS) by $(INTERNAL_CFLAGS) and
add $(GDB_WARN_CFLAGS_NO_FORMAT).
(printcmd.o): Replace $(INTERNAL_CFLAGS_BASE) by $(INTERNAL_CFLAGS).

2010-09-02  Yao Qi  <yao@codesourcery.com>

        * linux-nat.c (status_to_str): Use WTERMSIG to extract the signal
        number from a WIFSIGNALED status.

*** empty log message ***

daily update

* symtab.h (lookup_type_symbol): Declare.
* symtab.c (lookup_symbol_in_language_full): Rename from
lookup_symbol_in_language.  Add 'for_type' argument.
(lookup_symbol_in_language): New function.
(lookup_type_symbol): Likewise.
(lookup_symbol_aux): Add 'for_type' argument.
(match_symbol_aux): New function.
(lookup_symbol_aux_symtabs): Use expand_one_symtab_matching.
(match_transparent_type): New function.
(basic_lookup_transparent_type): Use expand_one_symtab_matching.
* symfile.h (struct quick_symbol_functions)
<pre_expand_symtabs_matching>: Remove.
<expand_one_symtab_matching>: New field.
* psymtab.c (expand_one_symtab_matching_psymtabs): New function.
(pre_expand_symtabs_matching_psymtabs): Remove.
(psym_functions): Update.
* gdbtypes.c (lookup_typename): Use lookup_type_symbol.
* dwarf2read.c (dw2_lookup_symbol): Update comment.
(dw2_pre_expand_symtabs_matching): Remove.
(dw2_expand_one_symtab_matching): New function.
(dwarf2_gdb_index_functions): Update.

[NEWS] powerpc-lynxos gdbserver support

This adds an entry in the NEWS file announcing the GDBserver support
on powerpc-lynxos.

gdb/ChangeLog:

        * NEWS: Add entry announcing GDBserver support on powerpc-lynxos.

   * mi/mi-main.c (mi_cmd_list_target_features): Add `reverse'
       as a feature reported by -list-target-features.

       * gdb.texinfo (GDB/MI Miscellaneous Commands): Document new
       feature `reverse' output by -list-target-features.

gdbserver support for powerpc-lynxos (4.x)

This patch adds support for powerpc-lynxos.

gdbserver/ChangeLog:

        * gdbserver/lynx-low.c, gdbserver/lynx-low.h,
        gdbserver/lynx-ppc-low.c: New files.
        * Makefile.in (lynx_low_h): New variable.
        (lynx-low.o, lynx-ppc-low.o): New rules.
        * configure.ac: On LynxOS, link with -lnetinet.
        * configure.srv: Add handling of powerpc-*-lynxos* targets.
        * configure: regenerate.

Import vsnprintf from libiberty if not available.

vasprintf and vsnprintf are not available on LynxOS, at least not on
version 4.0.0.  This patch updates the configury to import them from
libiberty if not available out of the box.

gdbserver/
        * Makefile.in (vasprintf.o, vsnprintf.o): New rules.
        * configure.ac: Add check for vasprintf and vsnprintf.
        * configure, config.in: Regenerate.
        * server.h (vasprintf, vsnprintf): Add conditional declarations.

Generate regformats/rs6000/powerpc-32.dat

I will need this description for the ppc-lynxos debugger where, AFAICT,
we only have access to the core GP and FP registers.  So this adds
powerpc-32 to the list of targets for which a .dat file needs  to be
generated.

2010-04-30  Joel Brobecker  <brobecker@adacore.com>

        * features/Makefile (WHICH): Add rs6000/powerpc-32.
        (powerpc-32.o, powerpc-32.c): New rules.
        (clean): Also remove powerpc-32.c.
        * regformats/rs6000/powerpc-32.dat: Generate.

[gdbserver] Move malloc.h include to server.h.

This patch moves all includes of malloc.h, which were introduced
purely to get access to alloca's declaration, to server.h, next
to the include of alloca.h.

There is one exception: gdbreplay.c, which does not include server.h.
In this case, the include of alloca.h was simply moved up a bit, next
to the include of malloc.h.

gdb/gdbserver/ChangeLog:

        * gdbreplay.c: Move include of alloca.h up, next to include of
        malloc.h.
        * server.h: Add include of malloc.h.
        * mem-break.c: Remove include of malloc.h.
        * server.c, tracepoint.c, utils.c, win32-low.c: Likewise.

Put ChangeLog entries in the right ChangeLog file.

* netbsd-core.c (netbsd_core_core_file_pid): Renamed to ...
(netbsd_core_file_pid): ... this.

Fix typo

    * ld-scripts/section-match.exp: New file.  Runs ths
        section-match-1 test.
        * ld-scripts/section-match-1.d: New file.
        * ld-scripts/section-match-1.s: New file.
        * ld-scripts/section-match-1.t: New file.

2010-09-01  Tristan Gingold  <gingold@adacore.com>

* coffcode.h (coff_slurp_line_table): Add a cast.
(coff_slurp_reloc_table): Ditto.

Compile memmem.o without -Werror.

This reproduces the same approach as in GDB to allow us to build
specific files without -Werror.

gdb/gdbserver/ChangeLog:

        * Makefile.in (INTERNAL_CFLAGS_BASE): New variable. Extracted
        from INTERNAL_CFLAGS.
        (INTERNAL_WARN_CFLAGS): New variable.
        (INTERNAL_CFLAGS): Adjust, using INTERNAL_WARN_CFLAGS.

use xsnprintf instead of snprintf.

snprintf is not available on LynxOS, so I changed the calls to snprintf
to calls to xsnprintf, which should be strictly equivalent.

gdb/gdbserver/ChangeLog:

        * utils.c (xsnprintf): Make non-static.
        * server.h: Add xsnprintf declaration.
        * linux-low.c, nto-low.c, target.c, thread-db.c, tracepoint.c:
        replace calls to snprintf by calls to xsnprintf throughout.

include alloca.h if available.

On LynxOS, alloca is defined in allocal.h.  This fixes one warning that
occurs when building GDBserver for LynxOS.

gdb/gdbserver/ChangeLog:

        * configure.ac: Add configure check for alloca.
        * configure, config.in: Regenerate.
        * server.h: Include alloca.h if it exists.
        * gdbreplay.c: Include alloca.h if it exists.

*** empty log message ***

daily update

Fix "pushw imm16" for x86-64 disassembler.

gas/testsuite/

2010-08-31  H.J. Lu  <hongjiu.lu@intel.com>

PR binutils/11960
* gas/i386/opcode-intel.d: Updated.
* gas/i386/x86-64-opcode.d: Likewise.

* gas/i386/x86-64-opcode.s: Add a "pushw imm16" test.

opcodes/

2010-08-31  H.J. Lu  <hongjiu.lu@intel.com>

PR binutils/11960
* i386-dis.c (sIv): New.
(dis386): Replace Iq with sIv on "pushT".
(reg_table): Replace T with {T|} on callT, JcallT, jmpT and JjmpT.
(x86_64_table): Replace {T|}/{P|} with P.
(putop): Add 'w' to 'T'/'P' if needed for Intel syntax.
(OP_sI): Update v_mode.  Remove w_mode.

Create and use a specialized bcache type for psymbols

2010-08-31  Sami Wagiaalla  <swagiaal@redhat.com>

* symfile.c (reread_symbols): Use psymbol_bcache_free, and
psymbol_bcache_init.
* psymtab.h (psymbol_bcache_init): New function prototype.
(psymbol_bcache_free): New function prototype.
(psymbol_bcache_get_bcache): New function prototype.
* psymtab.c (psymbol_bcache_init): New function.
(psymbol_bcache_free): New function.
(psymbol_bcache_full): New function.
(psymbol_bcache_get_bcache): New function.
(add_psymbol_to_bcache): use psymbol_bcache_full.
* objfiles.h (psymbol_cache): Change type of psymbol_cache to
psymbol_bcache.
* symmisc.c (print_symbol_bcache_statistics): Updated.
(print_objfile_statistics): Updated.
* objfiles.c (allocate_objfile): Use psymbol_bcache_init to initialize
psymbol_cache.
(free_objfile): Use psymbol_bcache_free.

PR c++/11961:
* dwarf2read.c (new_symbol_full) <DW_TAG_template_type_param>:
Don't set TYPE_NAME on the type.

2010-08-31  Michael Snyder  <msnyder@msnyder-server.eng.vmware.com>

* infrun.c (set_exec_direction_func): Error out if target does not
support reverse execution.

* gdb.dwarf2/dw4-sig-types.exp: Use .x for executable name.

gdb/
Make linux_get_siginfo_type `type *' unique.
* linux-tdep.c (linux_gdbarch_data_handle, struct linux_gdbarch_data)
(init_linux_gdbarch_data, get_linux_gdbarch_data): New.
(linux_get_siginfo_type): New variable linux_gdbarch_data.  Initialize
it.  Use linux_gdbarch_data->siginfo_type as a persistent storage.
(_initialize_linux_tdep): New.

gdb/
Code cleanup.
* defs.h (find_memory_region_ftype): New typedef.
(exec_set_find_memory_regions): Use it.
* exec.c (exec_set_find_memory_regions): Use find_memory_region_ftype.
* fbsd-nat.c (fbsd_find_memory_regions): Likewise.
* gcore.c (objfile_find_memory_regions): Likewise.
* gnu-nat.c (gnu_find_memory_regions): Likewise.
* linux-nat.c (linux_nat_find_memory_regions): Likewise.
* procfs.c (iterate_over_mappings_cb_ftype): Remove.
(iterate_over_mappings): Rename iterate_over_mappings_cb_ftype to
find_memory_region_ftype.
(insert_dbx_link_bpt_in_region): Likewise.
(iterate_over_mappings): Likewise.  Drop the comment part about the
function prototype.
(find_memory_regions_callback): Use find_memory_region_ftype.
(proc_find_memory_regions): Likewise.
(info_mappings_callback): Rename iterate_over_mappings_cb_ftype to
find_memory_region_ftype.
* target.c (dummy_find_memory_regions): Use find_memory_region_ftype.
* target.h (struct target_ops) <to_find_memory_regions>: Likewise.

Enable custom bcache hash function.

2010-08-25  Sami Wagiaalla  <swagiaal@redhat.com>

* psymtab.c (add_psymbol_to_bcache): Remove 'static' from
'static partial_symbol psymbol'.
(psymbol_hash): New function.
(psymbol_compare): New function.
* bcache.c (hash_continue): New.
(hash): Use hash_continue.
* bcache.c: Add hash_function and compare_function
pointers to bcache struct.
(bcache_full): Use bcache->hash_function, and
bcache->compare_function.
(bcache_compare): New function.
(bcache_xmalloc): Take hash_function and
compare_function arguments and initialize the
bcach's pointers.
Updated comment.
* objfiles.c (allocate_objfile): Updated.
* symfile.c (reread_symbols): Updated.
* python/py-type.c (typy_richcompare): Updated.

    * ldlex.l (WILDCHAR): Add the caret and exclamation point
        characters, so that they can be used inside regexps.

Remove an extra "@item".

2010-08-31  H.J. Lu  <hongjiu.lu@intel.com>

* gdb.texinfo (i386 Features): Remove an extra "@item".

2010-08-31  Kai Tietz  <kai.tietz@onevision.com>

        * config/obj-coff-seh.c (obj_coff_seh_save): Correct comparison.
        (obj_coff_seh_stackalloc): Likewise.

    * ld.texinfo (MEMORY): Clarify the behaviour of the ! character in
        a region's attributes.

* config/obj-elf.c (obj_elf_init_stab_section): Fix assertion.

* scripttempl/mmo.sc: Move assignment to "Main" inside .text
output section statement.

*** empty log message ***

daily update

* config/obj-coff-seh.c: Rewrite the entire file.
(symtab, symptr, reltab, relcount, relsize): Remove.
(seh_ctx_root, seh_ctx): Remove.
(xdata_seg, xdata_subseg, pdata_seg): New.
(switch_xdata, switch_pdata): New.
(verify_context, verify_context_and_target, skip_whitespace_and_comma):
New parsing functions.  Rewrite all parsing functions to use them.
(obj_coff_seh_32): Fix != arm thinko.
(obj_coff_seh_handler): For x64, don't accept handler pointer here,
only flags.
(obj_coff_seh_handlerdata): New.
(do_seh_endproc): Split out of ...
(obj_coff_seh_endproc): ... here.
(obj_coff_seh_proc): Use it, if needed.
(seh_x64_make_prologue_element): Use XRESIZEVEC, symbol_temp_new_now.
(seh_x64_read_reg): Remove mm_regs alternative.  Tidy integer reg
alternatives.  Don't slurp commas.
(seh_read_offset): Remove.
(obj_coff_seh_pushframe): Split out from obj_coff_seh_push.
(obj_coff_seh_scope): Remove.
(obj_coff_seh_save): Decide UWOP_SAVE_* vs _FAR immediately.
(obj_coff_seh_stackalloc): Decide _SMALL vs _LARGE immediately.
(out_one, out_two, out_four): New.
(seh_x64_write_prologue_data, seh_x64_size_prologue_data,
seh_x64_write_function_xdata, write_function_xdata): Rewrite
from seh_x64_write_xdata, seh_needed_unwind_info, seh_store_elm_data,
seh_getelm_data_size, seh_getsize_of_unwind_entry,
seh_make_unwind_entry, seh_getsize_unwind_data, and
seh_create_unwind_data.
(seh_arm_write_function_pdata): Rewrite from seh_arm_create_pdata.
(write_function_pdata): Rewrite from make_function_entry_pdata.
(seh_write_text_eh_data, make_function_entry_pdata,
seh_arm_create_pdata, seh_arm_write_pdata, seh_reloc, save_relocs,
seh_symbol_init, seh_symbol, quick_section, seh_emit_rva,
seh_emit_long, seh_make_globl, seh_make_section2, seh_make_section,
seh_make_xlbl_name, make_seh_text_label, seh_fill_pcsyms,
seh_needed_unwind_info, seh_store_elm_data, seh_getelm_data_size,
seh_getsize_of_unwind_entry, seh_make_unwind_entry,
seh_getsize_unwind_data, seh_create_unwind_data,
seh_make_function_entry_xdata, seh_x64_makescope_elem): Remove.
* config/obj-coff-seh.h (SEH_CMDS): Remove seh_savemm, seh_scope.
Add seh_handlerdata.  Adjust function/what arguments for
seh_savereg, seh_pushframe, seh_stackalloc.
(struct seh_prologue_element): Adjust members to closer match
the elements of the UNWIND_CODE structure.
(struct seh_scope_elem): Remove.
(struct seh_context): Replace char* members with symbolS or
expressionS as appropriate.  Sort members by ARM/x64 applicability.
Remove obsolete stuff wrt direct symbol and reloc manipulation.

gdb
2010-08-30  Andre Poenitz  <andre.poenitz@nokia.com>
    Tom Tromey  <tromey@redhat.com>

PR python/11792:
* python/py-value.c (valpy_get_dynamic_type): New function.
(value_object_getset): Add "dynamic_type".
(valpy_get_type): Fail on error.
gdb/doc
PR python/11792:
* gdb.texinfo (Values From Inferior): Document dynamic_type.
gdb/testsuite
PR python/11792:
* gdb.python/py-value.exp (test_subscript_regression): Add
dynamic_type test.