* tree-ssa-loop-ivopts.c (get_computation_aff_1): New.
(get_computation_aff): Reorder parameters. Use get_computation_aff_1.
(get_computation_at, rewrite_use_address): Update use of
get_computation_aff.
Martin Jambor [Tue, 2 May 2017 12:49:55 +0000 (14:49 +0200)]
[PR 78687] Set SRA grp_write lazily
2017-05-02 Martin Jambor <mjambor@suse.cz>
PR tree-optimization/78687
* tree-sra.c (access): New field parent.
(process_subtree_disqualification): New function.
(disqualify_candidate): Call it.
(build_accesses_from_assign): Reset write flag if creating an
assighnment link.
(build_access_subtree): Fill in parent field and also prpagate
down grp_write flag.
(create_artificial_child_access): New parameter set_grp_write, set
grp_write to its value.
(propagate_subaccesses_across_link): Also propagate grp_write flag
values.
(propagate_all_subaccesses): Push the closest parent back to work
queue if add_access_to_work_queue returned true.
Richard Biener [Tue, 2 May 2017 12:43:47 +0000 (12:43 +0000)]
common.opt (fstrict-overflow): Alias negative to fwrapv.
2017-05-02 Richard Biener <rguenther@suse.de>
* common.opt (fstrict-overflow): Alias negative to fwrapv.
* doc/invoke.texi (fstrict-overflow): Remove all traces of
-fstrict-overflow documentation.
* tree.h (TYPE_OVERFLOW_UNDEFINED): Do not test flag_strict_overflow.
(POINTER_TYPE_OVERFLOW_UNDEFINED): Test !flag_wrapv instead of
flag_strict_overflow.
* ipa-inline.c (can_inline_edge_p): Do not test flag_strict_overflow.
* lto-opts.c (lto_write_options): Do not stream it.
* lto-wrapper.c (merge_and_complain): Do not handle it.
* opts.c (default_options_table): Do not set -fstrict-overflow.
(finish_options): Likewise do not clear it when sanitizing.
* simplify-rtx.c (simplify_const_relational_operation): Do not
test flag_strict_overflow.
ada/
* gcc-interface/misc.c (gnat_post_options): Do not set
-fstrict-overflow.
* c-c++-common/Wlogical-op-1.c: Add -fwrapv to restore previous
behavior.
* gcc.target/i386/pr46253.c: Make i unsigned to avoid warning.
Uros Bizjak [Tue, 2 May 2017 11:56:35 +0000 (13:56 +0200)]
alpha.md (*add<mode>3_ieee): Merge to add<mode>3 using enabled attribute.
* config/alpha/alpha.md (*add<mode>3_ieee): Merge to add<mode>3
using enabled attribute.
(*sub<mode>3_ieee): Merge to sub<mode>3 using enabled attribute.
(*mul<mode>3_ieee): Merge to mul<mode>3 using enabled attribute.
(*div<mode>3_ieee): Merge to div<mode>3 using enabled attribute.
(*sqrt<mode>2_ieee): Merge to sqrt<mode>2 using enabled attribute.
(*fix_truncdfdi_ieee): Merge to *fix_truncdfdi2 using enabled attribute.
(*fix_truncsfdi_ieee): Merge to *fix_truncsfdi2 using enabled attribute.
(*floatdisf_ieee): Merge to floatdisf2 using enabled attribute.
(*floatdidf_ieee): Merge to floatdidf2 using enabled attribute.
(*truncdfsf2_ieee): Merge to truncdfsf2 using enabled attribute.
(*cmpdf_ieee): Merge to *cmpdf_internal using enabled attribute.
Richard Biener [Tue, 2 May 2017 11:24:52 +0000 (11:24 +0000)]
re PR tree-optimization/80549 (wrong code at -O2 and -O3 in both 32-bit and 64-bit modes on x86_64-linux-gnu (executable does not terminate))
2017-05-02 Richard Biener <rguenther@suse.de>
PR tree-optimization/80549
* tree-cfgcleanup.c (mfb_keep_latches): New helper.
(cleanup_tree_cfg_noloop): Create forwarders to known loop
headers if they do not have a preheader.
Eric Botcazou [Tue, 2 May 2017 09:21:19 +0000 (09:21 +0000)]
trans.c (assoc_to_constructor): Make sure Corresponding_Discriminant is only called on discriminants.
2017-05-02 Eric Botcazou <ebotcazou@adacore.com>
* gcc-interface/trans.c (assoc_to_constructor): Make sure
Corresponding_Discriminant is only called on discriminants.
Skip the saving of the result only for them.
(gnat_to_gnu) <N_Selected_Component>: Likewise.
<N_Unchecked_Type_Conversion>: Translate the result type first.
(gigi): Set TREE_NOTHROW on Begin_Handler.
(stmt_list_cannot_raise_p): New predicate.
(Exception_Handler_to_gnu_gcc): Emit a simple final call instead of
a cleanup if the statements of the handler cannot raise.
(process_freeze_entity): Use Is_Record_Type.
(process_type): Likewise.
The declaration for memcpy had the size parameter declared as
unsigned long. For avr, __SIZE_TYPE__ is unsigned int, and
this caused a builtin-declaration-mismatch warning, resulting
in a couple of FAILs.
Fixed by typedef'ing __SIZE_TYPE__ to size_t and using size_t as the
type for memcpy's third parameter.
Arnaud Charlet [Tue, 2 May 2017 09:17:13 +0000 (11:17 +0200)]
[multiple changes]
2017-05-02 Eric Botcazou <ebotcazou@adacore.com>
* einfo.ads (Corresponding_Record_Component): New alias
for Node21 used for E_Component and E_Discriminant.
* einfo.adb (Corresponding_Record_Component): New function.
(Set_Corresponding_Record_Component): New procedure.
(Write_Field21_Name): Handle Corresponding_Record_Component.
* sem_ch3.adb (Inherit_Component): Set
Corresponding_Record_Component for every component in
the untagged case. Clear it afterwards for non-girder
discriminants.
* gcc-interface/decl.c (gnat_to_gnu_entity)
<E_Record_Type>: For a derived untagged type with discriminants
and constraints, apply the constraints to the layout of the
parent type to deduce the layout.
(field_is_aliased): Delete.
(components_to_record): Test DECL_ALIASED_P directly.
(annotate_rep): Check that fields are present except for
an extension.
(create_field_decl_from): Add DEBUG_INFO_P
parameter and pass it in recursive and other calls. Add guard
for the manual CSE on the size.
(is_stored_discriminant): New predicate.
(copy_and_substitute_in_layout): Consider only
stored discriminants and check that original fields are present
in the old type. Deal with derived types. Adjust call to
create_variant_part_from.
2017-05-02 Ed Schonberg <schonberg@adacore.com>
* exp_ch6.adb (Expand_Call_Helper): When locating the
accessibility entity created for an access parameter, handle
properly a reference to a formal of an enclosing subprogram. if
the reference appears in an inherited class-wide condition, it
is the rewriting of the reference in the ancestor expression,
but the accessibility entity must be that of the current formal.
Eric Botcazou [Tue, 2 May 2017 09:08:28 +0000 (09:08 +0000)]
sem_ch3.adb (Build_Derived_Private_Type): If the parent type has discriminants...
2017-05-02 Eric Botcazou <ebotcazou@adacore.com>
* sem_ch3.adb (Build_Derived_Private_Type): If the parent type
has discriminants, do not override the Stored_Constraint list of
the full view of the derived type with that of the derived type.
* gnatls.adb: Merge and refactor code from Prj.Env and remove
this deprecated dependency.
2017-05-02 Ed Schonberg <schonberg@adacore.com>
* exp_util.ads: minor comment addition.
2017-05-02 Eric Botcazou <ebotcazou@adacore.com>
* sem_ch3.adb (Build_Derived_Record_Type): Fix a few typos and
pastos in part #3 of the head comment.
2017-05-02 Ed Schonberg <schonberg@adacore.com>
* exp_ch3.adb (Freeze_Type): Do not generate an invariant
procedure body for a local (sub)type declaration within a
predicate function. Invariant checks do not apply to these, and
the expansion of the procedure will happen in the wrong scope,
leading to misplaced freeze nodes.
2017-05-02 Ed Schonberg <schonberg@adacore.com>
* exp_util.adb (Insert_Library_Level_Action): Use proper scope
to analyze generated actions. If the main unit is a body,
the required scope is that of the corresponding unit declaration.
2017-05-02 Arnaud Charlet <charlet@adacore.com>
* einfo.adb (Declaration_Node): flip branches of
an IF statement to avoid repeated negations in its condition;
no change in semantics, only to improve readability.
Arnaud Charlet [Tue, 2 May 2017 09:00:48 +0000 (11:00 +0200)]
[multiple changes]
2017-05-02 Justin Squirek <squirek@adacore.com>
* sem_ch4.adb (Analyze_Case_Expression): Add check for valid
alternative expression.
* sem_res.adb (Resolve_Case_Expression): Ditto.
2017-05-02 Ed Schonberg <schonberg@adacore.com>
* exp_disp.adb (Set_All_DT_Position, In_Predef_Prim_DT):
Refine predicate for the case where the primitive operation
is a renaming of equality. An overriding operation that is
a user-defined renaming of predefined equality inherits its
slot from the overridden operation. Otherwise it is treated
as a predefined op and occupies the same predefined slot as
equality. A call to it is transformed into a call to its alias,
which is the predefined equality. A dispatching call thus uses
the proper slot if operation is further inherited and called
with class-wide arguments.
2017-05-02 Justin Squirek <squirek@adacore.com>
* errout.adb (Set_Msg_Text): Add a case to switch the message
type when the character '[' is detected signifying a warning
about a run-time exception.
* opt.ads Add a new Warning_Mode value for new switch
* switch-b.adb (Scan_Binder_Switches): Add case for the binder
to handle new warning mode
* usage.adb (Usage): Add usage entry for -gnatwE
* warnsw.adb (Set_Warning_Switch): Add case for the new switch
2017-05-02 Ed Schonberg <schonberg@adacore.com>
* sem_prag.adb (Process_Conversion): Reject an intrinsic operator
declaration that operates on some fixed point type.
Arnaud Charlet [Tue, 2 May 2017 08:55:03 +0000 (10:55 +0200)]
[multiple changes]
2017-05-02 Ed Schonberg <schonberg@adacore.com>
* exp_ch3.adb (Build_Initialization_Call): Generate a null
statement if the initialization call is a null procedure, as
can happen with a controlled type with no explicit Initialize
procedure, or an array of such.
* exp_ch7.adb (Process_Object_Declaration): For a type with
controlled components that has a trivial Initialize procedure,
insert declaration for finalization counter after object
declaration itself.
(Make_Deep_Array_Body, Build_Initialize_statements): Do not create
finalization block and attendant declarations if component has
a trivial Initialize procedure.
(Make_Init_Call): Do not generate a call if controlled type has
a trivial Initialize procedure.
2017-05-02 Eric Botcazou <ebotcazou@adacore.com>
* g-forstr.ads (Data): Move Format component last.
* g-forstr.adb ("+"): Adjust for above change.
* g-rewdat.ads (Buffer): Move Buffer, Current, Pattern and Value last.
* g-sechas.ads (Context): Move Key last.
* g-socket.ads (Service_Entry_Type): Move Aliases last.
* s-fileio.adb (Temp_File_Record): Move Name last.
* s-regexp.adb (Regexp_Value): Move Case_Sensitive last.
* xr_tabls.ads (Project_File): Move Src_Dir and Obj_Dir last.
2017-05-02 Jerome Lambourg <lambourg@adacore.com>
* bindusg.adb, bindgen.adb, gnatbind.adb, opt.ads: Remove the -nognarl
switch introduced recently. finally not needed.
* checks.ads (Null_Exclusion_Static_Checks): Add Boolean
parameter Array_Comp to indicate the case of an array object
with null-excluding components.
* checks.adb (Null_Exclusion_Static_Checks):
Call Compile_Time_Constraint_Error instead of
Apply_Compile_Time_Constraint_Error in the component case. Also
call that when Array_Comp is True, with an appropriate warning for
the array component case. Only create an explicit initialization
by null in the case of an object of a null-excluding access type
(and no longer do that in the component case).
* sem_ch3.adb (Check_Component): Add a Boolean parameter
Array_Comp defaulted to False. Pass Empty for the Comp
actual when calling Null_Exclusion_Static_Checks in the case
where Comp_Decl matches Object_Decl, because we don't have a
component in that case. In the case of an object or component
of an array type, pass True for Array_Comp on the recursive call
to Check_Component.
Arnaud Charlet [Tue, 2 May 2017 08:47:29 +0000 (10:47 +0200)]
[multiple changes]
2017-05-02 Bob Duff <duff@adacore.com>
* s-taprop-linux.adb (Prio_To_Linux_Prio): New function to correctly
compute the linux priority from the Ada priority. Call this everywhere
required. In particular, the previous version was not doing this
computation when setting the ceiling priority in various places. It
was just converting to C.int, which results in a ceiling that is off
by 1.
Arnaud Charlet [Tue, 2 May 2017 08:44:41 +0000 (10:44 +0200)]
[multiple changes]
2017-05-02 Tristan Gingold <gingold@adacore.com>
* s-trasym.ads: Add comment.
2017-05-02 Bob Duff <duff@adacore.com>
* sem_elab.adb, sem_elab.ads: Minor comment fixes.
* sem_ch4.adb: Minor reformatting.
* s-taprop-linux.adb, s-taspri-posix.ads: Code refactoring.
* s-taspri-posix-noaltstack.ads: Minor refactoring.
* sinput.ads: Minor typo fix.
2017-05-02 Ed Schonberg <schonberg@adacore.com>
* exp_ch9.adb (Discriminated_Size): Moved to sem_util.
* sem_util.ads, sem_util.adb (Discriminated_Size): Predicate moved
here from exp_ch9, to recognize objects whose creation requires
dynamic allocation, so that the proper warning can be emitted
when restriction No_Implicit_Heap_Allocation is in effect.
* sem_ch3.adb (Analyze_Object_Declaration): Use Discriminated_Size
to emit proper warning when an object that requires dynamic
allocation is declared.
Arnaud Charlet [Tue, 2 May 2017 08:39:29 +0000 (10:39 +0200)]
[multiple changes]
2017-05-02 Ed Schonberg <schonberg@adacore.com>
* sem_ch4.adb (Find_Equality_Types, Try_One_Interp): The same relaxed
visibility rules for equality operators that apply within an
instantiation apply within an inlined body.
* sem_type.adb (Add_One_Interp): ditto.
* sem_prag.adb (Analyze_Pragma): Forbid pragma Contract_Cases on null
procedures.
2017-05-02 Eric Botcazou <ebotcazou@adacore.com>
* snames.ads-tmpl
(Name_Assume, Name_Attribute_Definition, Name_Loop_Optimize,
Name_No_Tagged_Streams): Move to regular pragmas. Add
placeholders for Default_Scalar_Storage_Order, Dispatching_Domain,
and Secondary_Stack_Size.
(Pragma_Id): Move Pragma_Assume,
Pragma_Attribute_Definition, Pragma_Loop_Optimize and
Pragma_No_Tagged_Streams to second part.
* sem_ch4.adb (Analyze_Selected_Component): Improve error
detection for illegal references to private components or
operations of a protected type in the body of the type.
Arnaud Charlet [Tue, 2 May 2017 08:31:12 +0000 (10:31 +0200)]
[multiple changes]
2017-05-02 Eric Botcazou <ebotcazou@adacore.com>
* opt.ads: Add missing GNAT markers in comments.
* opt.adb (Set_Opt_Config_Switches): Do not override earlier
settings of Optimize_Alignment at the end.
* checks.adb (Apply_Constraint_Check): Do not apply
a discriminant check when the associated type is a constrained
subtype created for an unconstrained nominal type.
* exp_attr.adb: Minor reformatting.
2017-05-02 Bob Duff <duff@adacore.com>
* sem_ch3.adb (OK_For_Limited_Init_In_05): Handle correctly
the N_Raise_Expression case.
* sem_ch6.adb (Check_Limited_Return): Minor: clarify comment,
and add assertions.
2017-05-02 Yannick Moy <moy@adacore.com>
* exp_ch4.adb (Expand_N_Op_Ne): Do not bump parenthese level and
optimize length comparison in GNATprove mode.
* exp_spark.adb (Expand_SPARK_Op_Ne): New function to rewrite
operator /= into negation of operator = when needed.
(Expand_SPARK): Call new
function to expand operator /=.
2017-05-02 Ed Schonberg <schonberg@adacore.com>
* exp_fixd.adb (Expand_Divide_Fixed_By_Fixed_Giving_Fixed):
Simplify the expression for a fixed-fixed division to remove
divisions by constants whenever possible, as an optimization
for restricted targets.
* checks.adb, sem_ch3.adb, sem_ch6.adb: Minor reformatting.
2017-05-02 Bob Duff <duff@adacore.com>
* exp_attr.adb (Callable, Identity, Terminated): Use Find_Prim_Op
to find primitive ops, instead of using an Identifier that will
later be looked up. This is necessary because these ops are not
necessarily visible at all places where we need to call them.
* exp_util.ads: Minor comment fix.
Arnaud Charlet [Tue, 2 May 2017 08:26:12 +0000 (10:26 +0200)]
[multiple changes]
2017-05-02 Ed Schonberg <schonberg@adacore.com>
* sem_ch6.adb (Fully_Conformant_Expressions): Two entity
references are fully conformant if they are both expansions
of the discriminant of a protected type, within one of the
protected operations. One occurrence may be expanded into a
constant declaration while the other is an input parameter to
the corresponding generated subprogram.
2017-05-02 Justin Squirek <squirek@adacore.com>
* sem_ch3.adb (Check_For_Null_Excluding_Components): Created for
recursivly searching composite-types for null-excluding access
types and verifying them.
(Analyze_Object_Declaration): Add a
call to Check_Null_Excluding_Components for static verification
of non-initialized objects.
* checks.adb, checks.ads (Null_Exclusion_Static_Checks): Added
a parameter for a composite-type's component and an extra case
for printing component information.
2017-05-02 Yannick Moy <moy@adacore.com>
* sem_ch10.adb (Analyze_Subunit): Take
configuration pragma into account when restoring appropriate
pragma for analysis of subunit.
2017-05-02 Justin Squirek <squirek@adacore.com>
* s-tasren.adb, s-tasini.adb, s-taprop-linux.adb,
s-mudido-affinity.adb,, a-exetim-posix.adb, a-direio.adb,
g-socket.adb, s-taenca.adb, s-fileio.adb: Remove unused use-type
clauses from the runtime.
Arnaud Charlet [Tue, 2 May 2017 08:18:17 +0000 (10:18 +0200)]
[multiple changes]
2017-05-02 Eric Botcazou <ebotcazou@adacore.com>
* freeze.adb (Check_Component_Storage_Order): Do not treat bit-packed
array components specially.
2017-05-02 Ed Schonberg <schonberg@adacore.com>
* sem_ch8.adb (Premature_Usage): If the premature usage of
an entity is as the expression in its own object decaration,
rewrite the reference as Any_Id to prevent cascaded errors or
compiler loops when such an entity is used in an address clause.
Iain Buclaw [Mon, 1 May 2017 23:31:24 +0000 (23:31 +0000)]
d-demangle.c (strtol): Remove declaration.
libiberty/ChangeLog:
* d-demangle.c (strtol): Remove declaration.
Updated all callers to use dlang_number.
(dlang_number): New function.
(dlang_value): Moved check for ISDIGIT into dlang_parse_integer.
* testsuite/d-demangle-expected: Add tests.
* d-demangle.c (dlang_parse_symbol): Remove function.
(dlang_parse_qualified): New function.
(dlang_parse_mangle): New function.
(dlang_type): Update to call dlang_parse_qualified.
(dlang_identifier): Update to call either dlang_parse_qualified or
dlang_parse_mangle.
(dlang_type_modifier_p): Remove function.
(dlang_call_convention_p): Don't allow type modifiers in mangle.
(dlang_template_args): Update to call dlang_identifier.
(dlang_demangle): Update to call dlang_parse_mangle.
* testsuite/d-demangle-expected: Add tests.
David Malcolm [Mon, 1 May 2017 19:15:36 +0000 (19:15 +0000)]
Eliminate fixit_hint class hierarchy
The original implementation of fix-it hints (r230674) had an abstract
base class "fixit_hint" and three subclasses, representing
each of insertions, replacements, and deletions.
Having multiple classes for fix-it hints was a nuisance, as it required
per-class logic everywhere that the hints were handled.
In r239632 I eliminated the deletion subclass in favor of replacement
with the empty string (two subclasses are easier than three).
This patch eliminates the class hierarchy altogether by implementing
insertion in terms of replacement, by representing replacements via
a half-open interval (so that for an insertion, start == next location,
and we're effectively replacing an empty range at the insertion point
with the new string).
This greatly simplifies the code for handling fix-it hints; for example
it allows removal of a parallel class hierarchy of line_event within
edit-context.c.
It also improves consolidation of hints: we can now consolidate
insertions at the same location, affecting a couple of tests
(selftest::test_one_liner_many_fixits and
gcc.dg/Wmissing-braces-fixits.c).
gcc/ChangeLog:
* diagnostic-show-locus.c (layout::get_expanded_location): Rewrite
to use new fixit_hint representation, using the "replace" logic.
(get_line_span_for_fixit_hint): Likewise.
(layout::print_any_fixits): Likewise.
(selftest::test_one_liner_many_fixits): Rename to...
(selftest::test_one_liner_many_fixits_1): ...this, and update
comment and expected output to reflect that the multiple fix-it
hints are now consolidated into one insertion.
(selftest::test_one_liner_many_fixits_2): New test.
(selftest::test_diagnostic_show_locus_one_liner): Update for
above.
(selftest::test_fixit_consolidation): Update for fix-it API
change.
* diagnostic.c (print_parseable_fixits): Likewise.
* edit-context.c (edited_line::m_line_events): Convert from
auto_vec <line_event *> to auto_vec <line_event>.
(class line_event): Convert from abstract base class to a concrete
class, taking over the role of replace_event.
(class insert_event): Delete.
(class replace_event): Rename to class line_event. Convert to
half-open range.
(edit_context::add_fixits): Reimplement.
(edit_context::apply_insert): Delete.
(edit_context::apply_replace): Rename to...
(edit_context::apply_fixit): ...this. Convert to half-open range.
(edited_file::apply_insert): Delete.
(edited_file::apply_replace): Rename to...
(edited_file::apply_fixit): ...this.
(edited_line::~edited_line): Drop deletion of events.
(edited_line::apply_insert): Delete.
(edited_line::apply_replace): Rename to...
(edited_line::apply_fixit): ...this. Convert to half-open range.
Update for change to type of m_line_events.
* edit-context.h (edit_context::apply_insert): Delete.
(edit_context::apply_replace): Rename to...
(edit_context::apply_fixit): ...this.
gcc/testsuite/ChangeLog:
* gcc.dg/Wmissing-braces-fixits.c: Update expected output to
reflect insertion fix-it hints at the same location now being
consolidated.
libcpp/ChangeLog:
* include/line-map.h (source_range::intersects_line_p): Delete.
(rich_location::add_fixit): Delete.
(rich_location::maybe_add_fixit): New method.
(class fixit_hint): Reimplement in terms of...
(class fixit_replace): ...this.
(class fixit_insert): Delete.
* line-map.c (linemap_position_for_loc_and_offset): Drop overzealous
linemap_assert_fails.
(source_range::intersects_line_p): Rename to...
(fixit_hint::affects_line_p): New function.
(rich_location::add_fixit_insert_before): Reimplement in terms of
maybe_add_fixit, moving validation there.
(rich_location::add_fixit_insert_after): Likewise.
(column_before_p): Delete.
(rich_location::add_fixit_replace): Reimplement in terms of
maybe_add_fixit, moving validation there. Convert closed input range
to half-open range.
(rich_location::add_fixit): Delete.
(rich_location::maybe_add_fixit): New function.
(fixit_insert::fixit_insert): Delete.
(fixit_insert::~fixit_insert): Delete.
(fixit_insert::affects_line_p): Delete.
(fixit_insert::maybe_append_replace): Delete.
(fixit_replace::fixit_replace): Rename to...
(fixit_hint::fixit_hint): ...this, rewriting as necessary.
(fixit_replace::~fixit_replace): Delete.
(fixit_replace::affects_line_p): Delete.
(fixit_replace::maybe_append_replace): Rename to...
(fixit_hint::maybe_append): ...this, rewriting as necessary.
Tom de Vries [Mon, 1 May 2017 18:31:07 +0000 (18:31 +0000)]
Require linker plugin for pr78768.c
The test-case has an xfail-ed line. For linkers without plugin support, that
line happens to xpass. Require linker with plugin support, such that the line
is no longer xpass-ing, but unsupported.
* d-demangle.c (dlang_identifier): Handle template constraint symbols.
(dlang_parse_template): Only advance if template symbol prefix is
followed by a digit.
* testsuite/d-demangle-expected: Add tests.
Jakub Jelinek [Mon, 1 May 2017 09:50:59 +0000 (11:50 +0200)]
re PR target/79430 (action of statement incorrectly optimised away)
PR target/79430
* rtlanal.c (reg_set_p): If reg is a stack_pointer_rtx, also
check for stack push/pop autoinc.
* config/i386/i386.c (ix86_agi_dependent): Return false
if the only reason why modified_in_p returned true is that
addr is SP based and set_insn is a push or pop.
Eric Botcazou [Mon, 1 May 2017 08:12:31 +0000 (08:12 +0000)]
decl.c (components_to_record): Add more comments.
* gcc-interface/decl.c (components_to_record): Add more comments.
Put back pending fields onto the regular list if the misalignment
happens to cancel itself.
Jan Hubicka [Sun, 30 Apr 2017 15:02:11 +0000 (17:02 +0200)]
re PR tree-optimization/79224 (Large C-Ray slowdown)
PR ipa/79224
* ipa-inline-analysis.c (dump_predicate): Add optional parameter NL.
(account_size_time): Use two predicates - exec_pred and
nonconst_pred_ptr.
(evaluate_conditions_for_known_args): Compute both clause and
nonspec_clause.
(evaluate_properties_for_edge): Evaulate both clause and nonspec_clause.
(inline_summary_t::duplicate): Update.
(estimate_function_body_sizes): Caluculate exec and nonconst predicates
separately.
(compute_inline_parameters): Likewise.
(estimate_edge_size_and_time): Update caluclation of time.
(estimate_node_size_and_time): Compute both time and nonspecialized
time.
(estimate_ipcp_clone_size_and_time): Update.
(inline_merge_summary): Update.
(do_estimate_edge_time): Update.
(do_estimate_edge_size): Update.
(do_estimate_edge_hints): Update.
(inline_read_section, inline_write_summary): Stream both new predicates.
* ipa-inline.c (compute_uninlined_call_time): Take uninlined_call_time
as argument.
(compute_inlined_call_time): Cleanup.
(big_speedup_p): Update.
(edge_badness): Update.
* ipa-inline.h (INLINE_TIME_SCALE): Remove.
(size_time_entry): Replace predicate by exec_predicate and
nonconst_predicate.
(edge_growth_cache_entry): Cache both time nad nonspecialized time.
(estimate_edge_time): Return also nonspec_time.
(reset_edge_growth_cache): Update.
Jakub Jelinek [Sat, 29 Apr 2017 16:18:11 +0000 (18:18 +0200)]
re PR rtl-optimization/80491 (Compiler regression for long-add case.)
PR rtl-optimization/80491
* ifcvt.c (noce_process_if_block): When looking for x setter
with missing else_bb, don't check only the insn right before
cond_earliest, but look for the last insn that x is modified in
within the same bb.
PR tree-optimization/79697
* tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Check if callee
is BUILT_IN_STRDUP, BUILT_IN_STRNDUP, BUILT_IN_REALLOC.
(propagate_necessity): Check if def_callee is BUILT_IN_STRDUP or
BUILT_IN_STRNDUP.
* gimple-fold.c (gimple_fold_builtin_realloc): New function.
(gimple_fold_builtin): Call gimple_fold_builtin_realloc.
PR tree-optimization/80523
* gimple-ssa-sprintf.c (target_to_host_charmap): New global variable.
(init_target_to_host_charmap, target_to_host, target_strtol10): New
functions.
(maybe_warn, format_directive, parse_directive): Use new functions.
(pass_sprintf_length::execute): Call init_target_to_host_charmap.
gcc/testsuite/ChangeLog:
PR tree-optimization/80523
* gcc.dg/tree-ssa/builtin-sprintf-warn-18.c: New test.
* configure.ac (SYSTEM_HEADER_DIR, BUILD_SYSTEM_HEADER_DIR,
target_header_dir): Set correctly.
* configure: Regenerated.
* Makefile.in (BUILD_SYSTEM_HEADER_DIR): New make variabe.
(LIMITS_H_TEST, if_multiarch, stmp-fixinc): Use BUILD_SYSTEM_HEADER_DIR
instead of SYSTEM_HEADER_DIR.
* sem_ch4.adb (Complete_Object_Operation): When rewriting the
controlling actual in a prefixed call, preserve the original node
information if the prefix itself has been rewritten, for ASIS use.
* exp_ch6.adb (Insert_Post_Call_Actions):
Code clean up. Insert the post-call actions after an enclosing
procedure call when N denotes a function call and appears as an
actual parameter in the procedure call.
2017-04-28 Eric Botcazou <ebotcazou@adacore.com>
* freeze.adb (Check_Component_Storage_Order): If there is a clause
for the component, also reject the attribute if the component
doesn't end on a byte boundary and its scalar storage order is
different from that of the enclosing record type.
* sem_util.ads, sem_util.adb (Might_Raise): New function
that replaces Is_Exception_Safe, but has the opposite
sense. Is_Exception_Safe was missing various cases -- calls inside
a pragma Debug, calls inside an 'if' or assignment statement,
etc. Might_Raise now walks the entire subtree looking for things
that can raise.
* exp_ch9.adb (Is_Exception_Safe): Remove.
(Build_Protected_Subprogram_Body): Replace call to
Is_Exception_Safe with "not Might_Raise". Misc cleanup (use
constants where possible).
* exp_ch7.adb: Rename Is_Protected_Body -->
Is_Protected_Subp_Body. A protected_body is something different
in the grammar.
* checks.adb (Install_Primitive_Elaboration_Check):
Do not generate an elaboration check if all checks have been
suppressed.
2017-04-28 Ed Schonberg <schonberg@adacore.com>
* sem_ch13.adb (Analyze_Aspect_Specifications, case
Interrupt_Handler and Attach_Handler): Generate reference
to protected operation to prevent spurious warnings about
unreferenced entities. Previous scheme failed with style checks
enabled.
2017-04-28 Ed Schonberg <schonberg@adacore.com>
* sem_prag.adb (Relocate_Pragmas_To_Body): A pragma Warnings
that follows an expression function must not be relocated to
the generated body, because it applies to the code that follows.
* sem_util.adb (Is_Dependent_Component_Of_Mutable_Object): Test
Relaxed_RM_Semantics to avoid having CodePeer issue errors on
code that might violate the more stringent checking for 'Access
introduced in Ada 2005.
* exp_ch9.adb (Build_Lock_Free_Unprotected_Subprogram_Body): Also
initialize Block_Decls variable.
(Expand_Entry_Barrier): Add pragma Warnings on Func_Body variable.
(Build_Dispatching_Requeue): Add pragma Warnings on Op variable.
* exp_disp.adb (Expand_Interface_Actuals): Initialize
Formal_DDT and Actual_DDT variables.
(Expand_Interface_Thunk): Initialize Iface_Formal.
(Make_DT): Initialize Size_Comp.
(Make_Select_Specific_Data_Table): Initialize Decls.
* exp_dist.adb (Add_RACW_Primitive_Declarations_And_Bodies):
Also initialize more RPC_Receiver_* variables.
(Build_To_Any_Function): Initialize Cstr_Formal.
* exp_prag.adb (Expand_Pragma_Contract_Cases): Initialize Msg_Str.
2017-04-28 Ed Schonberg <schonberg@adacore.com>
* sem_ch6.adb (Freeze_Type_Refs): For an interface conversion
node between access types, freeze the designated type as well,
so that dispatch table pointers are created in the proper scope,
and not in the constructed body of the expression function.
2017-04-28 Bob Duff <duff@adacore.com>
* alloc.ads (Nodes_Initial): Go back to previous value. The large
value makes large compilations faster, but small compilations slower.
2017-04-28 Arnaud Charlet <charlet@adacore.com>
* sem_util.adb: minor typos in Is_Child_Or_Sibling.
* erroutc.adb (Compilation_Errors): Do not consider info messages
as suitable warnings when warnings must be treated as errors.
* sem_ch7.adb (Analyze_Package_Specification): Do not consider
internally-generated packages when outputting completion
information.
* errout.adb (Output_Messages): Do not consider info messages as
suitable warnings when warnings must be treated as errors.
* errutil.adb (Finalize): Do not consider info messages as
suitable warnings when warnings must be treated as errors.
2017-04-28 Eric Botcazou <ebotcazou@adacore.com>
* warnsw.ads: Minor fix for incorrect wording in comment.
2017-04-28 Ed Schonberg <schonberg@adacore.com>
* sem_res.adb (In_Instance_Code): New predicate in
Valid_Conversion, to determine whether a type conversion appears
as (or within) an actual for a formal object. Type conversions
in instances are not rechecked in Valid_Conversion because
visibility changes between generic location andi instance may
lead to spurious errors, but conversions within an actual must be
fully checked, and they are not fully resolved when pre-analyzing
the actuals.
* exp_ch6.adb (Expand_N_Extended_Return_Statement): Use
New_Copy_Tree instead of Relocate_Node as any subsequent copies
of the relocated node will have mangled Parent pointers.
* sem_util.adb (Build_NCT_Hash_Tables): Reset both hash
tables used in conjunction with entity and itype replication.
(Visit_Entity): Rewrite the restriction on which entities
require duplication. The restriction now includes all types.
projects / thirdparty / gcc.git / log
