* prj-proc.adb, make.adb, mlib-prj.adb, prj.adb, prj.ads, clean.adb,
prj-nmsc.adb, prj-env.adb (Project_List_Table, Project_Element):
removed. Lists of projects are now implemented via standard malloc
rather than through the table.
2009-04-24 Thomas Quinot <quinot@adacore.com>
* sem_ch12.adb: Minor reformatting
* g-trasym.adb: Minor reformatting
* exp_ch6.adb: Minor reformatting
2009-04-24 Robert Dewar <dewar@adacore.com>
* layout.adb (Layout_Type): For packed array type, copy unset
size/alignment fields from the referenced Packed_Array_Type.
2009-04-24 Bob Duff <duff@adacore.com>
* lib-load.adb (Make_Instance_Unit): Revert previous change, no
longer needed after sem_ch12 changes.
* sem.adb (Walk_Library_Items): Include with's in some debugging
printouts.
* prj.ads, prj-nmsc.adb (Unit_Project): removed, since in fact we were
only ever using the Project field.
2009-04-24 Ed Schonberg <schonberg@adacore.com>
* sem_ch12.adb (Build_Instance_Compilation_Unit_Nodes): Do not set
Body_Required on the generated compilation node. The new node is linked
to its body, but both share the same file, so we do not set this flag
on the new unit so as not to create a spurious dependency on a
non-existent body in the ali file for the instance.
* prj.adb, prj.ads, prj-nmsc.adb (Check_File, Record_Ada_Source,
Add_Source): merge some code between those. In particular change where
file normalization is done to avoid a few extra calls to
Canonicalize_File_Name. This also removes the need for passing
Current_Dir in a number of subprograms.
2009-04-24 Bob Duff <duff@adacore.com>
* lib-load.adb (Make_Instance_Unit): In the case where In_Main is
False, assign the correct unit to the Cunit field of the new table
entry. We want the spec unit, not the body unit.
* rtsfind.adb (Make_Unit_Name, Maybe_Add_With): Simplify calling
interface for these.
(Maybe_Add_With): Check whether we're trying to a with on the current
unit, and avoid creating such directly self-referential with clauses.
(Text_IO_Kludge): Add implicit with's for the generic pseudo-children of
[[Wide_]Wide_]Text_IO. These are needed for Walk_Library_Items,
and matches existing comments in the spec.
* sem.adb (Walk_Library_Items): Add various special cases to make the
assertions pass.
* sem_ch12.adb (Build_Instance_Compilation_Unit_Nodes): Use Body_Cunit
instead of Parent (N), for uniformity.
* sem_res.adb: additional optimization to inhibit creation of
redundant transient scopes.
2009-04-24 Bob Duff <duff@adacore.com>
* rtsfind.ads: Minor comment fix
2009-04-24 Emmanuel Briot <briot@adacore.com>
* prj-proc.adb, prj-nmsc.adb (Find_Ada_Sources,
Get_Path_Name_And_Record_Ada_Sources): merged, since these were
basically doing the same work (for explicit or implicit sources).
(Find_Explicit_Sources): renamed to Find_Sources to better reflect its
role. Rewritten to share some code (testing that all explicit sources
have been found) between ada_only and multi_language modes.
2009-04-24 Jerome Lambourg <lambourg@adacore.com>
* sem_prag.adb (Check_Form_Of_Interface_Name): Allow space in Ext_Name
for CLI imported types.
(Analyze_Pragma): Allow CIL or Java imported functions returning
access-to-subprogram types.
* make.adb, prj.adb, prj.ads, makeutl.adb, makeutl.ads:
(Project_Data.Dir_Path): field removed, since it can be computed
directly from .Directory, and is needed only once when processing the
project is buildgpr.adb or make.adb
* loop-iv.c (replace_single_def_regs): Look for REG_EQUAL notes;
follow chains of regs with a single definition, and allow expressions
that are function_invariant_p.
* prj-proc.adb, make.adb, mlib-prj.adb, prj.adb, prj.ads, makeutl.adb,
clean.adb, prj-nmsc.adb, prj-env.adb, prj-env.ads (Project_Data.Seen):
field removed. This is not a property of the
project, just a boolean used to traverse the project tree, and storing
it in the structure prevents doing multiple traversal in parallel.
(Project_Data.Checked): also removed, since it was playing the same role
as Seen when we had two nested loops, and this is no longer necessary
(For_All_Imported_Projects): removed, since in fact there was already
the equivalent in For_Every_Project_Imported. The latter was rewritten
to use a local hash table instead of Project_Data.Seen
Various loops were rewritten to use For_Every_Project_Imported, thus
removing the need for Project_Data.Seen. This avoids a lot of code
duplication
* sem_res.adb (Resolve_Actuals): Do not create blocks around code
statements, even though the actual of the call is a concatenation,
because the argument is static, and we want to preserve warning
messages about sequences of code statements that are not marked
volatile.
* sem_warn.adb: remove obsolete comment about warning being obsolete
* s-tasren.adb (Task_Do_Or_Queue): If a timed entry call is being
requeued and the delay has expired while within the accept statement
that executes the requeue, do not perform the requeue and indicate that
the timed call has been aborted.
2009-04-24 Emmanuel Briot <briot@adacore.com>
* mlib-prj.adb, prj.adb, prj.ads, prj-nmsc.adb, prj-env.adb
(Has_Ada_Sources, Has_Foreign_Sources): new subprograms
(Project_Data.Ada_Sources_Present, Foreign_Sources_Present): removed,
since they can be computed from the above.
* gnatcmd.adb: Call Prj.Env.Initialize with the Project_Tree
* prj-env.adb: Move all global variables to the private part of the
project tree data.
Access these new components instead of the global variables no longer
in existence.
(Add_To_Path): New Project_Tree_Ref parameter, to access the new
components that were previously global variables.
* prj-env.ads (Initialize): New Project_Tree_Ref parameter
(Set_Mapping_File_Initial_State_To_Empty): New Project_Tree_Ref
parameter.
* prj-nmsc.adb (Compute_Unit_Name): New Project_Tree_Ref parameter to
be able to call Set_Mapping_File_Initial_State_To_Empty with it.
* prj.adb (Initialize): Do not call Prj.Env.Initialize
(Reset): Do not call Prj.Env.Initialize. Instead, initialize the new
components in the private part of the project tree data.
* prj.ads (Private_Project_Tree_Data): new components moved from
Prj.Env: Current_Source_Path_File, Current_Object_Path_File,
Ada_Path_Buffer, Ada_Path_Length, Ada_Prj_Include_File_Set,
Ada_Prj_Objects_File_Set, Fill_Mapping_File.
PR middle-end/39867
* fold-const.c (fold_cond_expr_with_comparison): When folding
> and >= to MAX, make sure the MAX uses the same type as the
comparison operands.
testsuite:
2009-04-24 Paolo Bonzini <bonzini@gnu.org>
* config/frv/frv.c (frv_frame_access): Do not use reg+reg
addressing for DImode accesses.
(frv_print_operand_address): Handle PLUS case.
* config/frv/frv.h (FIXED_REGISTERS): Mark link register as
fixed.
* s-osinte-darwin.adb, s-osinte-darwin.ads: lwp_self now returns the
mach thread id.
2009-04-24 Emmanuel Briot <briot@adacore.com>
* prj-env.adb, prj-env.ads (Body_Path_Name_Of, Spec_Path_Name_Of,
Path_Name_Of_Library_Unit_Body): rEmove unused subprograms.
(For_All_Imported_Projects): new procedure
(For_All_Source_Dirs, For_All_Object_Dirs): Rewritten based on the
above rather than duplicating code.
* prj-proc.adb, prj.adb, prj.ads, prj-nmsc.adb, prj-env.adb
(Source_Id, Source_Data): use a real list to store sources rather than
using an external table to store the elements. This makes code more
efficient and more readable.
* prj-proc.adb, prj.adb, prj.ads, prj-nmsc.adb, prj-env.adb
(Source_Iterator): new type.
This removes the need for having the sources on three different
lists at the project tree, project and language level. They are now
on a single list.
* prj.adb, prj.ads, prj-nmsc.adb, prj-env.adb (Language_Index): renamed
to Language_Ptr to better reflect its new implementation.
(Project_Data.First_Languages_Processing): renamed to Languages now
that the field with that name is no longer used
(Project_Data.Languages): removed, no longer used, and duplicates
information already available through First_Language_Processing.
(Prj.Language_Index): now an actual pointer, instead of an index into
a table. This makes the list somewhat more obvious, but more importantly
removes the need to pass a pointer to the project_tree_data in a few
places, and makes accessing the attributes of a languages more
efficient.
* fe.h (Set_Identifier_Casing): Add const to second parameter.
* gcc-interface/misc.c (internal_error_function): Make copy of retur
from pp_formatted_text before assigning BUFFER to it.
(gnat_init): Likewise for main_input_filename and gnat_argv.
(gnat_printable_name): Remove cast from call to Set_Identifier_Casing.
* ttypes.ads (Target_Double_Float_Alignment): New variable.
(Target_Double_Scalar_Alignment): Likewise.
* get_targ.ads (Get_Strict_Alignment): Adjust external name.
(Get_Double_Float_Alignment): New imported function.
(Get_Double_Scalar_Alignment): Likewise.
* layout.adb (Set_Elem_Alignment): Take into account specific caps for
the alignment of "double" floating-point types and "double" or larger
scalar types, as parameterized by Target_Double_Float_Alignment and
Target_Double_Scalar_Alignment respectively.
* gcc-interface/gigi.h (double_float_alignment): Declare.
(double_scalar_alignment): Likewise.
(is_double_float_or_array): Likewise.
(is_double_scalar_or_array): Likewise.
(get_target_double_float_alignment): Likewise.
(get_target_double_scalar_alignment): Likewise.
* gcc-interface/targtyps.c (get_strict_alignment): Rename into...
(get_target_strict_alignment): ...this.
(get_target_double_float_alignment): New function.
(get_target_double_scalar_alignment): Likewise.
* gcc-interface/decl.c (gnat_to_gnu_entity) <E_Signed_Integer_Subtype>:
Test the presence of an alignment clause for under-aligned integer
types. Take into account specific caps for the alignment of "double"
floating-point types and "double" or larger scalar types, as
parameterized by Target_Double_Float_Alignment and
Target_Double_Scalar_Alignment respectively.
(validate_alignment): Likewise.
* gcc-interface/trans.c (Attribute_to_gnu) <Attr_Alignment>: Likewise.
(gigi): Initialize double_float_alignment and double_scalar_alignment.
* gcc-interface/utils.c (double_float_alignment): New global variable.
(double_scalar_alignment): Likewise.
(is_double_float_or_array): New predicate.
(is_double_scalar_or_array): Likewise.
* gcc-interface/decl.c (gnat_to_gnu_entity) <E_Array_Type>: If an
alignment is specified, do not promote that of the component type
beyond it.
<E_Array_Subtype>: Likewise.
* config/spu/spu.h (enum spu_builtin_type): Move here from
spu-builtins.h.
(struct spu_builtin_description): Likewise. Add GTY marker.
Do not use enum spu_function_code or enum insn_code.
(spu_builtins): Add extern declaration.
* config/spu/spu.c: Do not include "spu-builtins.h".
(enum spu_function_code, enum spu_builtin_type_index,
V16QI_type_node, V8HI_type_node, V4SI_type_node, V2DI_type_node,
V4SF_type_node, V2DF_type_node, unsigned_V16QI_type_node,
unsigned_V8HI_type_node, unsigned_V4SI_type_node,
unsigned_V2DI_type_node): Move here from spu-builtins.h.
(spu_builtin_types): Make static. Add GTY marker.
(spu_builtins): Add extern declaration with GTY marker.
Include "gt-spu.h".
* config/spu/spu-c.c: Do not include "spu-builtins.h".
(spu_resolve_overloaded_builtin): Do not use spu_function_code.
Check programmatically whether all parameters are scalar.
* gimplify.c (gimplify_modify_expr_rhs) <VAR_DECL>: Do not do a direct
assignment from the constructor either if the target is volatile.
ada/
* einfo.ads (Is_True_Constant): Lift restriction on atomic objects.
* sinfo.ads (Object Declaration): Likewise.
(Assignment Statement): Likewise.
* freeze.adb (Expand_Atomic_Aggregate): Remove useless test.
Do not force Is_True_Constant to false on the temporary.
(Freeze_Entity): Do not force Is_True_Constant to false on names on
the RHS of object declarations.
* gcc-interface/trans.c (lvalue_required_p) <N_Object_Declaration>:
New case. Return 1 if the object is atomic.
<N_Assignment_Statement>: Likewise.
* gcc-interface/decl.c (gnat_to_gnu_entity) <E_Modular_Integer_Subtype>
For packed array types, make the original array type a parallel type
for the modular type and its JM wrapper if the type is bit-packed.
<E_Array_Subtype>: Likewise. Do not generate the special XA parallel
record type for packed array types. Remove kludge.
gcc/cp/ChangeLog:
PR c++/38228
* pt.c (unify): Do not allow the result of a template argument
deduction to be a METHOD_TYPE.
* cvt.c (cp_convert): Report a meaningful error for non-valid use
of pointer to member functions during conversions.
* call.c (build_new_op): Report a meaningful error for non-valid
use of pointer to member functions in binary expressions.
* typeck.c (invalid_nonstatic_memfn_p): Do not crash when EXPR is
NULL;
gcc/testsuite/ChangeLog:
PR c++/38228
* g++.dg/expr/bound-mem-fun.C: New test.
* gcc-interface/gigi.h (create_index_type): Adjust head comment.
* gcc-interface/decl.c (gnat_to_gnu_entity) <E_Signed_Integer_Subtype>:
Use front-end predicates to compute signedness and precision.
<E_String_Literal_Subtype>: Fold range type.
Make sure to set longest_float_type_node to a scalar type.
(elaborate_entity): Use consistent Constraint_Error spelling.
(substitute_in_type) <INTEGER_TYPE>: Always copy the type.
* gcc-interface/misc.c (gnat_print_type) <INTEGER_TYPE>: Use brief
output for the modulus, if any.
<ENUMERAL_TYPE>: Likewise for the RM size.
* gcc-interface/trans.c (gnat_to_gnu): Use consistent Constraint_Error
spelling.
* gcc-interface/utils.c (finish_record_type): Really test the alignment
of BLKmode bit-fields to compute their addressability.
(create_index_type): Adjust comments.
(create_param_decl): Create the biased subtype manually.
* gcc-interface/utils2.c (build_component_ref): Use consistent
Constraint_Error spelling.
* gcc-interface/ada-tree.h (union lang_tree_node): Use standard idiom.
(SET_TYPE_LANG_SPECIFIC): Likewise. Fix formatting.
(SET_DECL_LANG_SPECIFIC): Likewise.
Reorder macros.
* gcc-interface/decl.c (gnat_to_gnu_entity) <E_Signed_Integer_Subtype>:
Update comment about use of build_range_type.
<E_Array_Type, E_Array_Subtype>: Use consistent naming convention.
<E_Array_Subtype>: Rework comments about TYPE_ACTUAL_BOUNDS and add
check for other cases of overloading.
* gcc-interface/trans.c (gigi): Use size_int in lieu of build_int_cst.
* gcc-interface/utils2.c (build_call_raise): Fix off-by-one error.
Use size_int in lieu of build_int_cst.
(build_call_alloc_dealloc): Use build_index_2_type in lieu of
build_range_type.
* config/v850/v850.md (epilogue): Remove suppressed code.
(return): Rename to return_simple and remove test of frame size.
* config/v850/v850.c (expand_epilogue): Rename call to gen_return
to gen_return_simple.
2009-04-22 Andrew Pinski <andrew_pinski@playstation.sony.com>
PR C/31499
* c-typeck.c (process_init_element): Treat VECTOR_TYPE like ARRAY_TYPE
and RECORD_TYPE/UNION_TYPE. When outputing the actual element and the
value is a VECTOR_CST, the element type is the element type of the
vector.
2009-04-22 Andrew Pinski <andrew_pinski@playstation.sony.com>
PR C/31499
* gcc.dg/vector-init-1.c: New testcase.
* gcc.dg/vector-init-2.c: New testcase.
* gcc-interface/utils2.c (build_binary_op) <PLUS_EXPR>: If operation's
type is an enumeral or a boolean type, change it to an integer type
with the same mode and signedness.
* config/rs6000/linux-unwind.h (get_regs): Remove type
puns. Change the type of `pc' to an array of unsigned ints and
update all users. Constify frame24.
gcc/cp/ChangeLog:
PR c++/39639
* parser.c (cp_parser_template_argument_list): Display an error
when an ellipsis is not preceded by a parameter pack. Also, warn
about variadic templates usage without -std=c++0x.
gcc/testsuite/ChangeLog:
PR c++/39639
* g++.dg/cpp0x/pr39639.C: New test.
manu [Wed, 22 Apr 2009 15:32:18 +0000 (15:32 +0000)]
2009-04-22 Manuel Lopez-Ibanez <manu@gcc.gnu.org>
PR c++/14875
* c-common.c (c_parse_error): Take a token_flags parameter.
Use token_type for the token type instead.
Pass token_flags to cpp_type2name.
* c-common.h (c_parse_error): Update declaration.
* c-parser.c (c_parser_error): Pass 0 as token flags.
libcpp/
* lex.c (cpp_type2name): Take a flags parameter. Call
cpp_named_operator2name for named operators and cpp_digraph2name
for digraphs.
(cpp_digraph2name): New.
(cpp_spell_token): Use it.
(cpp_output_token): Likewise.
* include/cpplib.h (cpp_type2name): Update declaration.
* init.c (cpp_named_operator2name): New.
* internal.h (cpp_named_operator2name): Declare.
cp/
* parser.c (cp_parser_error): Pass token->flags to c_parse_error.
testsuite/
* g++.dg/parse/parser-pr14875.C: New.
* g++.dg/parse/parser-pr14875-2.C: New.
* g++.dg/parse/error6.C: Update match string.
* make.adb, prj-env.adb, prj-env.ads, prj-nmsc.adb, prj.adb,
prj.ads (Create_Mapping_File): merge the two versions for Ada_Only and
Multi_Language modes, to avoid code duplication.
(Project_Data.Include_Language): Removed.
* tempdir.adb (Create_Temp_File): Add a diagnostic in verbose mode when
the temporary file cannot be created, indicating the directory when the
creation was attempted.
2009-04-22 Emmanuel Briot <briot@adacore.com>
* prj-env.adb, prj-env.ads (Create_Mapping_File): we need to compare
the language names, not their indices.
* function.c (expand_function_end): Do not emit a jump to the "naked"
return label for fall-through returns.
* except.c (sjlj_emit_function_exit): Always place the call to the
unregister function at the location installed by expand_function_end.
* prj-proc.adb, prj-nmsc.adb (Check_Naming_Schemes): split into several
smaller subprograms.
Renamed to Check_File_Naming_Schemes to avoid confusion with the
other Check_Naming_Schemes functions that plays a totally different
role.
(Check_Unit_Based_Lang, Check_File_Based_Lang): new subprograms,
extracted from the above. These were partially rewritten to avoid
unnecessary code and temporary variables.
(Compute_Unit_Name): new subprogram, merge of Check_Unit_Based_Lang
and Get_Unit (which for now still exist since they contain mode-specific
code)
* prj.ads, prj.adb, prj-nmsc.adb, prj-proc.adb (Recursive_Process):
Remove duplicated code.
(Canonical_Case_File_Name): new subprogram
(Check_And_Normalize_Unit_Names): new subprogram
(Write_Attr): new subprogram
Better sharing of code
(Check_Naming_Ada_Only, Check_Naming_Multi_Lang): new subprogram, to
split Check_Naming and help find duplicated code
(Check_Common): new subprogram, sharing code between ada_only and
multi_language mode.
(Naming_Data.Dot_Repl_Loc): field removed
* prj-proc.adb, prj-nmsc.adb (Load_Naming_Exceptions): New subprogram.
Minor refactoring to reduce the size of
Process_Sources_In_Multi_Language_Mode.
Avoid extra copied of Source_Data, which we found in the past could be
quite slow.
(Mark_Excluded_Sources): new subprogram.
(Remove_Locally_Removed_Files_From_Units): merged into the above
Refactors Process_Sources_In_Multi_Language_Mode to reduce its size,
and allow better sharing of code between multi_lang and ada_only modes
(Project_Extends): removed, since exact duplicate of Prj.Is_Extending
projects / thirdparty / gcc.git / log
