gdb/
[binutils-gdb.git] / gdb / ChangeLog
index dc8e6c0bbdf8c0beb8e9d0133604a65f8e066760..d274e2f02bd9eb2bbcd5ae88dd329f86ae6a5a7f 100644 (file)
@@ -1,3 +1,299 @@
+2009-06-27  Daniel Jacobowitz  <dan@codesourcery.com>
+           Jan Kratochvil  <jan.kratochvil@redhat.com>
+
+       * NEWS: Document inlined function support.
+       * Makefile.in (SFILES): Add inline-frame.c.
+       (COMMON_OBS): Add inline-frame.o.
+       * block.c (contained_in): Rewrite to use lexical nesting.
+       (block_linkage_function): Skip inlined function blocks.
+       (block_inlined_p): New.
+       * block.h (struct block): Update comment.
+       (block_inlined_p): New prototype.
+       * blockframe.c (get_frame_block): Handle inlined functions.
+       (get_frame_function): Do not use block_linkage_function.
+       (block_innermost_frame): Use get_frame_block and contained_in.
+       * breakpoint.c (watchpoint_check): Remove extra reinit_frame_cache.
+       Skip over inlined functions.  Simplify epilogue check.
+       (bpstat_check_breakpoint_conditions): Use get_stack_frame_id.
+       Update comments.
+       (set_momentary_breakpoint): Only accept non-inlined frames.
+       (watch_command_1): Use frame_unwind_caller_pc and
+       frame_unwind_caller_id instead of get_prev_frame.
+       (until_break_command): Likewise.  Use get_stack_frame_id.
+       * buildsym.c (end_symtab): Set SYMBOL_SYMTAB for block functions.
+       * dwarf2loc.c (dwarf_expr_frame_base): Use block_linkage_function.
+       * dwarf2read.c (process_die): Handle DW_TAG_inlined_subroutine.
+       (read_func_scope, new_symbol): Likewise.  Handle arguments specially
+       for inlined functions without call site information.
+       (inherit_abstract_dies): Allow tag mismatch for inlined subroutines.
+       (die_specification): Treat DW_AT_abstract_origin as a specification.
+       (read_type_die): Handle DW_TAG_inlined_subroutine.
+       * frame-unwind.c (frame_unwind_init): Add inline_frame_unwind.
+       * frame.c (fprint_frame_id): Print inline depth.
+       (fprint_frame_type): Handle INLINE_FRAME and SENTINEL_FRAME.
+       (skip_inlined_frames, get_stack_frame_id): New.
+       (frame_unwind_caller_id): Use skip_inlined_frames.
+       (frame_id_inlined_p): New.
+       (frame_id_eq): Make the logic match the comments.  Add inline_depth
+       check.
+       (frame_id_inner): Handle inlined functions.
+       (frame_unwind_pc): New function, copied from frame_unwind_caller_pc.
+       (frame_unwind_caller_pc): Use skip_inlined_frames and frame_unwind_pc.
+       (get_prev_frame_1): Check for inline frames.  Split out frame
+       allocation to get_prev_frame_raw.
+       (get_prev_frame_raw): New function.
+       (get_prev_frame): Handle inline frames.
+       (get_frame_pc): Use frame_unwind_pc.
+       (get_frame_address_in_block): Skip inlined frames on both sides.
+       (pc_notcurrent): Delete.
+       (find_frame_sal): Rewrite to handle inline call sites.  Use
+       get_frame_address_in_block.
+       (deprecated_update_frame_pc_hack): Make static.
+       * frame.h: Update comments.
+       (struct frame_id): Add inline_depth.
+       (enum frame_type): Add INLINE_FRAME.
+       (frame_id_inlined_p, get_stack_frame_id): New prototypes.
+       * gdbthread.h (struct thread_info): Add step_stack_frame_id field.
+       * infcmd.c (set_step_frame): New function.
+       (step_once): Use set_step_frame.  Handle inlined functions.
+       (until_next_command): Use set_step_frame.
+       (finish_backward), finish_forward): Use get_stack_frame_id.
+       (finish_command): Support inlined functions.
+       * inferior.h (set_step_info): New prototype.
+       * infrun.c (RESUME_ALL): Use minus_one_ptid.
+       (clear_proceed_status): Clear step_stack_frame_id.
+       (init_wait_for_inferior): Call clear_inline_frame_state.
+       (init_execution_control_state): Make static.
+       (set_step_info): New function.
+       (init_thread_stepping_state): Do not set the symtab or line here.
+       (stepped_in_from): New function.
+       (handle_inferior_event): Handle inlined functions.  Use set_step_info.
+       (insert_step_resume_breakpoint_at_frame): Use get_stack_frame_id.
+       (struct inferior_status): Add step_stack_frame_id.
+       (save_inferior_status, restore_inferior_status): Save and restore
+       step_stack_frame_id.
+       * inline-frame.c, inline-frame.h: New files.
+       * minsyms.c (prim_record_minimal_symbol_and_info): Use XCALLOC.
+       * regcache.c (regcache_write_pc): Call reinit_frame_cache.
+       * s390-tdep.c (s390_prologue_frame_unwind_cache): Handle INLINE_FRAME.
+       * stack.c (frame_show_address): New.
+       (print_frame_info, print_frame): Use it.
+       (find_frame_funname): Use get_frame_function.  Handle inlined blocks.
+       (frame_info): Mark inlined functions.
+       (backtrace_command_1): Use get_current_user_frame.
+       (print_frame_local_vars, print_frame_label_vars): Update comments.
+       (return_command): Refuse inlined functions.
+       * symtab.c (lookup_symbol_aux_local): Stop at inlined function
+       boundaries.
+       (find_function_start_sal): Avoid inlined functions.
+       (completion_list_add_fields): New function.
+       (default_make_symbol_completion_list): Use it.  Use block_static_block
+       and block_global_block.  Check for inlined functions.
+       (skip_prologue_using_sal): Avoid line number comparison across
+       inlining.
+       * symtab.h (struct symbol): Add is_inlined.
+       (SYMBOL_INLINED): New.
+       * target.c (target_resume): Call clear_inline_frame_state.
+       * valops.c (value_of_variable): Check block_inlined_p.
+
+2009-06-27  Daniel Jacobowitz  <dan@codesourcery.com>
+
+       * frame.c (frame_unwind_id): Renamed to ...
+       (frame_unwind_caller_id): ... this.  All callers updated.
+       (frame_pc_unwind): Renamed to ...
+       (frame_unwind_caller_pc): ... this.  All callers updated.
+       * frame.h: Document frame_unwind_caller_WHAT functions.
+       (frame_unwind_id): Renamed to ...
+       (frame_unwind_caller_id): ... this.
+       (frame_pc_unwind): Renamed to ...
+       (frame_unwind_caller_pc): ... this.
+       * hppa-tdep.c (hppa_find_unwind_entry_in_block): Correct comment.
+       * stack.c (parse_frame_specification_1): Do not rely on
+       frame_unwind_id.
+
+2009-06-27  Michael Snyder  <msnyder@vmware.com>
+
+       * infrun.c (handle_inferior_event): Improve handling of
+       reverse-stepping into and over shared library functions.
+
+2009-06-25  Karen Osmond  <karen.osmond@gmail.com>
+
+       * linespec.c (find_method): Add new not_found_ptr parameter to
+       bring in line with symtab_from_filename, decode_variable.
+       (decode_compound): Likewise.  Also propagate not_found_ptr to 
+       find_method.
+       (decode_line_1): Propagate not_found_ptr to decode_compound.
+       (cplusplus_error): Now throws NOT_FOUND_ERROR.
+
+2009-06-25  Karen Osmond  <karen.osmond@gmail.com>
+
+       * MAINTAINERS (Write After Approval): Add self.
+
+2009-06-23  Ulrich Weigand  <uweigand@de.ibm.com>
+
+       * valops.c (value_one): Reimplement broken decimal-float case.
+
+2009-06-23  Ulrich Weigand  <uweigand@de.ibm.com>
+
+       * eval.c (evaluate_subexp_standard): Use expression architecture
+       instead of current_gdbarch.
+
+       * infrun.c (handle_inferior_event): Avoid unncessary reference
+       to current_gdbarch and get_current_frame ().
+
+2009-06-23  Ulrich Weigand  <uweigand@de.ibm.com>
+
+       * breakpoint.h (set_longjmp_breakpoint): Add THREAD argument.
+       (enum bptype): Add bp_longjmp_master.
+
+       * breakpoint.c (create_longjmp_master_breakpoint): New function.
+       (update_breakpoints_after_exec): Handle bp_longjmp_master
+       breakpoints.  Call create_longjmp_master_breakpoint.
+       (print_it_typical, bpstat_stop_status, bpstat_what,
+       print_one_breakpoint_location, allocate_bp_location, mention,
+       delete_command, breakpoint_re_set_one): Handle bp_longjmp_master.
+       (breakpoint_re_set): Call create_longjmp_master_breakpoint.
+       (create_longjmp_breakpoint): Delete.
+       (set_longjmp_breakpoint): Add THREAD argument.  Reimplement
+       to install momentary clones of bp_longjmp_master breakpoints.
+
+       * infcmd.c (step_1): Pass thread to set_longjmp_breakpoint.
+
+2009-06-23  Cary Coutant  <ccoutant@google.com>
+
+       * dwarf2read.c (dwarf_decode_lines): Ignore rows where is_stmt is 0.
+       Set basic_block to 0 after a special opcode.
+
+2009-06-23  Jan Kratochvil  <jan.kratochvil@redhat.com>
+
+       Fix a regression by the mmap patch from 2009-06-16.
+       * symfile.c (reread_symbols): Move sym_finish and clear_objfile_data
+       calls before deleting the OBJFILE obstack.  Extend the comment.
+
+2009-06-23  Ulrich Weigand  <uweigand@de.ibm.com>
+
+       * jv-lang.h (JAVA_OBJECT_SIZE): Remove.
+       (get_java_object_header_size): Add GDBARCH parameter.
+       * jv-lang.c (get_java_object_header_size): Add GDBARCH parameter.
+       Use it instead of current_gdbarch.
+       (evaluate_subexp_java): Replace JAVA_OBJECT_SIZE with call to
+       get_java_object_header_size.
+       * jv-valprint.c (java_value_print): Likewise.
+
+2009-06-23  Sami Wagiaalla  <swagiaal@redhat.com>
+
+       * dwarf2read.c (process_die): Handle import statements
+       (DW_TAG_imported_declaration, case DW_TAG_imported_module)
+       (read_import_statement): New.
+       (read_func_scope): Update using_directives to point to current context
+       (read_lexical_block_scope): Ditto.
+       * cp-support.h: Added prototype for cp_add_using.
+       * cp-namespace.c: Removed local context_stack.
+       (cp_initialize_namespace): Deleted.
+       (cp_finalize_namespace): Deleted.
+       (cp_add_using_directive): Use using_directives instead of using_list.
+       (cp_add_using): No longer static.
+       * buildsym.h: Created global using_direct variable.
+       Created using_direct variable in context_stack.
+       * buildsym.c (finish_block): Set using directives for the block under
+       construction.
+       (start_symtab): Removed call to cp_initialize_namespace().
+       (end_symtab): Removed call to cp_finalize_namespace().
+       (push_context): Save and reset using_directives.
+       * block.c (block_using): Return using directives for given
+       block instead of static block.
+
+2009-06-23  Ulrich Weigand  <uweigand@de.ibm.com>
+
+       * target-descriptions.h (struct type): Do not declare.
+       (struct tdesc_type): Declare.
+       (tdesc_named_type): Change return type to struct tdesc_type *.
+       (tdesc_record_type): Remove.
+       (tdesc_create_vector): Add prototype.
+       (tdesc_create_union): Likewise.
+       (tdesc_add_field): Likewise.
+
+       * target-descriptions.c (struct tdesc_reg): Replace gdb_type
+       member by tdesc_type member.
+       (struct tdesc_type_field, tdesc_type_field): Add type and vector def.
+       (struct tdesc_type, tdesc_type_p): Likewise.
+       (struct tdesc_arch_reg, tdesc_arch_reg): Likewise.
+       (struct tdesc_feature): Change type of "types" to VEC(tdesc_type_p).
+       (struct tdesc_arch_data): Replace registers member by arch_regs.
+
+       (tdesc_predefined_types): Change to array of pre-defined
+       struct tdesc_type structures.  Add "code_ptr" and "data_ptr".
+       (tdesc_named_type): Return struct tdesc_type * instead of GDB type.
+       (tdesc_gdb_type): New function.
+
+       (target_find_description): Update for data structure changes.
+       (tdesc_data_cleanup): Likewise.
+       (tdesc_numbered_register): Likewise.
+       (tdesc_use_registers): Likewise.
+       (tdesc_create_reg): Likewise.
+       
+       (tdesc_find_arch_register): New function.
+       (tdesc_find_register): Use it.
+       (tdesc_register_type): Use tdesc_gdb_type to generate GDB type from
+       tdesc_type on demand; cache generated types in tdesc_arch_reg.
+
+       (tdesc_free_type): New function.
+       (tdesc_free_feature): Call it.
+
+       (tdesc_create_vector): New function.
+       (tdesc_create_union): Likewise.
+       (tdesc_add_field): Likewise.
+       (tdesc_record_type): Remove.
+
+       (tdesc_type_id): Remove.
+       (maint_print_c_tdesc_cmd): Update for data structure changes.  Update
+       generated code to create tdesc_type structures instead of GDB types.
+
+       * xml-tdesc.c: Do not include "gdbtypes.h".
+       (struct tdesc_parsing_data): Change type of current_union to
+       struct tdesc_type *.
+       (tdesc_start_reg): Do not special-case "code_ptr" or "data_ptr".
+       (tdesc_start_union): Call tdesc_create_union.
+       (tdesc_start_field): Call tdesc_add_field.
+       (tdesc_start_vector): Call tdesc_create_vector.
+       (tdesc_end_union): Remove.
+       (feature_children): Remove reference to tdesc_end_union.
+
+       * features/arm-with-iwmmxt.c: Regenerate.
+       * features/mips-linux.c: Regenerate.
+       * features/mips64-linux.c: Regenerate.
+       * features/rs6000/powerpc-32.c: Regenerate.
+       * features/rs6000/powerpc-32l.c: Regenerate.
+       * features/rs6000/powerpc-403.c: Regenerate.
+       * features/rs6000/powerpc-403gc.c: Regenerate.
+       * features/rs6000/powerpc-505.c: Regenerate.
+       * features/rs6000/powerpc-601.c: Regenerate.
+       * features/rs6000/powerpc-602.c: Regenerate.
+       * features/rs6000/powerpc-603.c: Regenerate.
+       * features/rs6000/powerpc-604.c: Regenerate.
+       * features/rs6000/powerpc-64.c: Regenerate.
+       * features/rs6000/powerpc-64l.c: Regenerate.
+       * features/rs6000/powerpc-7400.c: Regenerate.
+       * features/rs6000/powerpc-750.c: Regenerate.
+       * features/rs6000/powerpc-860.c: Regenerate.
+       * features/rs6000/powerpc-altivec32.c: Regenerate.
+       * features/rs6000/powerpc-altivec32l.c: Regenerate.
+       * features/rs6000/powerpc-altivec64.c: Regenerate.
+       * features/rs6000/powerpc-altivec64l.c: Regenerate.
+       * features/rs6000/powerpc-e500.c: Regenerate.
+       * features/rs6000/powerpc-e500l.c: Regenerate.
+       * features/rs6000/powerpc-isa205-32l.c: Regenerate.
+       * features/rs6000/powerpc-isa205-64l.c: Regenerate.
+       * features/rs6000/powerpc-isa205-altivec32l.c: Regenerate.
+       * features/rs6000/powerpc-isa205-altivec64l.c: Regenerate.
+       * features/rs6000/powerpc-isa205-vsx32l.c: Regenerate.
+       * features/rs6000/powerpc-isa205-vsx64l.c: Regenerate.
+       * features/rs6000/powerpc-vsx32.c: Regenerate.
+       * features/rs6000/powerpc-vsx32l.c: Regenerate.
+       * features/rs6000/powerpc-vsx64.c: Regenerate.
+       * features/rs6000/powerpc-vsx64l.c: Regenerate.
+       * features/rs6000/rs6000.c: Regenerate.
+
 2009-06-23  Joel Brobecker  <brobecker@adacore.com>
 
        * ada-lang.c (decode_packed_array): If the value is a pointer to
        * c-lang.c (classify_type): Iterate over typedefs.
        * c-valprint.c (textual_element_type): Iterate over typedefs.
 
+2009-06-22  Tristan Gingold  <gingold@adacore.com>
+
+       * darwin-nat.c (darwin_xfer_partial): Use host_address_to_string
+       instead of "%p".
+
 2009-06-22  Jan Kratochvil  <jan.kratochvil@redhat.com>
 
        PR gdb/9988: