Re: [Patch, Fortran] PR57697 - Fix an issue with defined assignment
Thomas Koenig wrote: Build and regtested on x86-64-gnu-linux. OK? The patch is OK, also for 4.8. Please add a test case which also checks for the ForTrilinos failure. Thanks for the review. I have now committed the current patch as Rev. 202609. I will later try to create a test case, which fails before 202609 and works with it. Additionally, I will work on the 4.8 backport. (Unfortunately, it does not simply apply on gcc-4_8-branch has the gfc_build_intrinsic_call has slightly changed.) Tobias
Re: [RFC] Offloading Support in libgomp
On Sun, Sep 15, 2013 at 07:41:24PM +0400, Michael V. Zolotukhin wrote: > > Libgomp will start N-1 new threads, and all of them would want to look up > > mappings for i1,i2,...iK in the splay tree. The first one wouldn't find > > anything and would map and insert all the values to the tree. But the > > following > > ones would look-up these addresses in the exactly same order, which will > > lead to > > totally unbalanced tree. > > > > Am I missing anything or is it a real problem? > On second thought, this access order doesn't necessarily mean accessing in > ascending/descending keys order, so there is no problem here. Yes, splay tree can get totally unbalanced and you can have a linear lookup time, the O(log n) lookup time is amortized. But, if you e.g. really do lookup sorted keys (which is not given, the compiler puts vars into the clauses based on the user order or in the order references to those vars are discovered, plus for array sections pointer kinds which usually have different addresses go immediately after the data ones), you really can have one O(n) lookup if you've looked e.g. the highest address last time and now you're looking up the lowest and the tree is totally unbalanced, but then won't the following lookups be all O(1), because the keys you are looking up will be always immediately in the right child? Anyway, if the splay trees ever cause issues in real-world, it is not hard to just replace them by something else (R-B trees, AVL trees or similar). Jakub
Re: [Patch, Fortran] PR57697 - Fix an issue with defined assignment
Hi Tobias, > As testing showed, it didn't fix the real-world code: ForTrilinos's > ForTrilinos_ADT_3D_Burgers_6th_Pade did still fail as it has: > > *_F.DA65 = matrix_diff_x (&parm.621); >_F.DA66 = ax->epetra_rowmatrix.universal; // Deref of "ax"! > > Build and regtested on x86-64-gnu-linux. > OK? The patch is OK, also for 4.8. Please add a test case which also checks for the ForTrilinos failure. Thomas
Re: [ping][PATCH][1 of 2] Add value range info to SSA_NAME for zero sign extension elimination in RTL
Hi,
Updated the patch to the latest changes in trunk that splits tree.h. I
also noticed an error in printing double_int and fixed it.
Is this OK?
Thanks,
Kugan
+2013-09-12 Kugan Vivekanandarajah
+
+ * cfgexpand.c (maybe_dump_rtl_for_gimple_stmt) : Add range to dump.
+ * gimple-pretty-print.c (print_double_int) : New function.
+ * gimple-pretty-print.c (dump_gimple_phi) : Dump range info.
+ * (pp_gimple_stmt_1) : Likewise.
+ * tree-ssa-alias.c (dump_alias_info) : Check pointer type.
+ * tree-ssa-copy.c (fini_copy_prop) : Check pointer type and copy
+ range info.
+ * tree-ssanames.c (make_ssa_name_fn) : Check pointer type in
+ initialize.
+ * (set_range_info) : New function.
+ * (get_range_info) : Likewise.
+ * (duplicate_ssa_name_range_info) : Likewise.
+ * (duplicate_ssa_name_fn) : Check pointer type and call correct
+ duplicate function.
+ * tree-vrp.c (vrp_finalize): Call set_range_info to upddate
+ value range of SSA_NAMEs.
+ * tree.h (SSA_NAME_PTR_INFO) : changed to access via union
+ * tree.h (SSA_NAME_RANGE_INFO) : New macro
+
diff --git a/gcc/cfgexpand.c b/gcc/cfgexpand.c
index 88e48c2..302188e 100644
--- a/gcc/cfgexpand.c
+++ b/gcc/cfgexpand.c
@@ -1820,7 +1820,7 @@ maybe_dump_rtl_for_gimple_stmt (gimple stmt, rtx since)
{
fprintf (dump_file, "\n;; ");
print_gimple_stmt (dump_file, stmt, 0,
- TDF_SLIM | (dump_flags & TDF_LINENO));
+ TDF_SLIM | TDF_RANGE | (dump_flags & TDF_LINENO));
fprintf (dump_file, "\n");
print_rtl (dump_file, since ? NEXT_INSN (since) : since);
diff --git a/gcc/dumpfile.h b/gcc/dumpfile.h
index ddc770a..8896d89 100644
--- a/gcc/dumpfile.h
+++ b/gcc/dumpfile.h
@@ -83,9 +83,10 @@ enum tree_dump_index
#define TDF_CSELIB (1 << 23) /* Dump cselib details. */
#define TDF_SCEV (1 << 24) /* Dump SCEV details. */
#define TDF_COMMENT (1 << 25) /* Dump lines with prefix ";;" */
-#define MSG_OPTIMIZED_LOCATIONS (1 << 26) /* -fopt-info optimized sources */
-#define MSG_MISSED_OPTIMIZATION (1 << 27) /* missed opportunities */
-#define MSG_NOTE (1 << 28) /* general optimization info */
+#define TDF_RANGE (1 << 26) /* Dump range information. */
+#define MSG_OPTIMIZED_LOCATIONS (1 << 27) /* -fopt-info optimized sources */
+#define MSG_MISSED_OPTIMIZATION (1 << 28) /* missed opportunities */
+#define MSG_NOTE (1 << 29) /* general optimization info */
#define MSG_ALL (MSG_OPTIMIZED_LOCATIONS | MSG_MISSED_OPTIMIZATION \
| MSG_NOTE)
diff --git a/gcc/gimple-pretty-print.c b/gcc/gimple-pretty-print.c
index 01a1ab5..6531010 100644
--- a/gcc/gimple-pretty-print.c
+++ b/gcc/gimple-pretty-print.c
@@ -1600,6 +1600,25 @@ dump_gimple_asm (pretty_printer *buffer, gimple gs, int spc, int flags)
}
}
+/* Dumps double_int CST to BUFFER. */
+
+static void
+print_double_int (pretty_printer *buffer, double_int cst)
+{
+ tree node = double_int_to_tree (integer_type_node, cst);
+ if (TREE_INT_CST_HIGH (node) == 0)
+pp_printf (buffer, HOST_WIDE_INT_PRINT_UNSIGNED, TREE_INT_CST_LOW (node));
+ else if (TREE_INT_CST_HIGH (node) == -1
+ && TREE_INT_CST_LOW (node) != 0)
+pp_printf (buffer, "-" HOST_WIDE_INT_PRINT_UNSIGNED,
+ -TREE_INT_CST_LOW (node));
+ else
+sprintf (pp_buffer (buffer)->digit_buffer,
+ HOST_WIDE_INT_PRINT_DOUBLE_HEX,
+ (unsigned HOST_WIDE_INT) TREE_INT_CST_HIGH (node),
+ (unsigned HOST_WIDE_INT) TREE_INT_CST_LOW (node));
+}
+
/* Dump a PHI node PHI. BUFFER, SPC and FLAGS are as in pp_gimple_stmt_1.
The caller is responsible for calling pp_flush on BUFFER to finalize
@@ -1628,6 +1647,27 @@ dump_gimple_phi (pretty_printer *buffer, gimple phi, int spc, int flags)
pp_string (buffer, "# ");
}
+ if ((flags & TDF_RANGE)
+ && !POINTER_TYPE_P (TREE_TYPE (lhs))
+ && SSA_NAME_RANGE_INFO (lhs))
+{
+ double_int min, max;
+ value_range_type range_type;
+ get_range_info (lhs, min, max, range_type);
+ if (range_type == VR_VARYING)
+pp_printf (buffer, "# RANGE VR_VARYING");
+ else if (range_type == VR_RANGE || range_type == VR_ANTI_RANGE)
+ {
+pp_printf (buffer, "# RANGE ");
+pp_printf (buffer, "%s[", range_type == VR_RANGE ? "" : "~");
+print_double_int (buffer, min);
+pp_printf (buffer, ", ");
+print_double_int (buffer, max);
+pp_printf (buffer, "]");
+newline_and_indent (buffer, spc);
+ }
+}
+
if (flags & TDF_RAW)
dump_gimple_fmt (buffer, spc, flags, "%G <%T, ", phi,
gimple_phi_result (phi));
@@ -1930,6 +1970,32 @@ pp_gimple_stmt_1 (pretty_printer *buffer, gimple gs, int spc, int flags)
}
}
+ if ((flags & TDF_RANGE)
+ && gimple_has_lhs (gs))
+{
+ tree lhs = gimple_get_lhs (gs);
+
Re: Generic lambda and implicit function template commits reverted
Looks good. Jason
Re: Generic lambda and implicit function template commits reverted
On 15.09.2013 15:45, Jason Merrill wrote:
On 09/15/2013 06:22 AM, Adam Butcher wrote:
[PATCH 1/5] Fix uninitialized variables causing breakage with
-Werror.
[PATCH 2/5] Don't accept 'auto' as the 'type' of a template
parameter.
OK.
I've also added a case for rejecting 'auto' in a catch parameter.
[PATCH 3/5] Fix location diagnostics by returning to the deprecated
'input_location' global; must be a better fix for this.
Don't know why 'location_of (type)' gave ":" rather
than
"file:line:col:". My current workaround is to return to using
'input_location'. This gives the correct result but I doubt it
is
acceptable.
This seems to be because make_auto_1 sets the location of the auto
type to BUILTINS_LOCATION; I don't remember why I did that. Changing
it to use input_location seems appropriate.
Thanks. Doing that makes this patch unnecessary.
[PATCH 4/5] Lift CALL_FROM_THUNK_P setting to above the potential
'build_cplus_new' call to prevent ICE due to unexpected
tree type.
Rather than this, I've moved the call to 'build_cplus_new' back down to
after 'start_preparsed_function' as I needed to call
'set_flags_from_callee' prior to it but within function scope to prevent
regression of 49260 and 47263.
Deltas below. No regressions in g++.dg with these updates.
Cheers,
Adam
Subject: [PATCH] * pt.c (make_auto_1): Use input_location rather than
BUILTINS_LOCATION.
---
diff --git a/gcc/cp/pt.c b/gcc/cp/pt.c
index 58f920e..70f13bb 100644
--- a/gcc/cp/pt.c
+++ b/gcc/cp/pt.c
@@ -20925,7 +20925,7 @@ static tree
make_auto_1 (tree name)
{
tree au = cxx_make_type (TEMPLATE_TYPE_PARM);
- TYPE_NAME (au) = build_decl (BUILTINS_LOCATION,
+ TYPE_NAME (au) = build_decl (input_location,
TYPE_DECL, name, au);
TYPE_STUB_DECL (au) = TYPE_NAME (au);
TEMPLATE_TYPE_PARM_INDEX (au) = build_template_parm_index
--
[PATCH] Move 'build_cplus_new' call to after 'start_preparsed_function'
and call 'set_flags_from_callee' prior to prevent ICE due to unexpected
tree type and fix exception handling.
---
diff --git a/gcc/cp/lambda.c b/gcc/cp/lambda.c
index 0da22fd..c9118d8 100644
--- a/gcc/cp/lambda.c
+++ b/gcc/cp/lambda.c
@@ -885,13 +885,10 @@ maybe_add_lambda_conv_op (tree type)
}
}
else
-{
call = build_call_a (callop,
direct_argvec->length (),
direct_argvec->address ());
- if (MAYBE_CLASS_TYPE_P (TREE_TYPE (call)))
- call = build_cplus_new (TREE_TYPE (call), call, tf_warning_or_error);
-}
+
CALL_FROM_THUNK_P (call) = 1;
tree stattype = build_function_type (fn_result, FUNCTION_ARG_CHAIN
(callop));
@@ -987,6 +984,12 @@ maybe_add_lambda_conv_op (tree type)
}
tree body = begin_function_body ();
tree compound_stmt = begin_compound_stmt (0);
+ if (!generic_lambda_p)
+{
+ set_flags_from_callee (call);
+ if (MAYBE_CLASS_TYPE_P (TREE_TYPE (call)))
+ call = build_cplus_new (TREE_TYPE (call), call, tf_warning_or_error);
+}
call = convert_from_reference (call);
finish_return_stmt (call);
--
[PATCH] Don't allow 'auto' in type of catch parameter.
---
diff --git a/gcc/cp/decl.c b/gcc/cp/decl.c
index 6a4e863..80ceca1 100644
--- a/gcc/cp/decl.c
+++ b/gcc/cp/decl.c
@@ -10328,6 +10328,11 @@ grokdeclarator (const cp_declarator
*declarator,
error ("template parameter declared %");
type = error_mark_node;
}
+ else if (decl_context == CATCHPARM)
+ {
+ error ("catch parameter declared %");
+ type = error_mark_node;
+ }
else if (current_class_type && LAMBDA_TYPE_P (current_class_type))
{
if (cxx_dialect < cxx1y)
--
[PATCH] cpp0x/auto9.C: Downgrade expected error to expected pedwarn.
---
diff --git a/gcc/testsuite/g++.dg/cpp0x/auto9.C
b/gcc/testsuite/g++.dg/cpp0x/auto9.C
index 190bfa6..f357f2b 100644
--- a/gcc/testsuite/g++.dg/cpp0x/auto9.C
+++ b/gcc/testsuite/g++.dg/cpp0x/auto9.C
@@ -117,8 +117,8 @@ template struct G {}; // { dg-error
"auto" }
template struct H { H (); ~H (); };
H h; // { dg-error "invalid" }
-void qq (auto);// { dg-error "auto" }
-void qr (auto*); // { dg-error "auto" }
+void qq (auto);// { dg-warning "auto" }
+void qr (auto*); // { dg-warning "auto" }
// PR c++/46145
typedef auto autot;// { dg-error "auto" }
--
C++ PATCH for c++/41933 (variadic lambda capture)
This patch fixes one of the few remaining holes in GCC's C++11 support:
lambda capture of a variadic function parameter pack. The
implementation involves introducing the internal notion of a FIELD_DECL
pack, even though such a thing can not be written explicitly in C++11.
Tested x86_64-pc-linux-gnu, applying to trunk.
commit 85cb3c43f2b8aa7897b499e125a5d77e7a8a5a3c
Author: Jason Merrill
Date: Sun Sep 15 08:56:42 2013 -0700
Core DR 904
PR c++/41933
* parser.c (cp_parser_lambda_introducer): Handle variadic capture.
* lambda.c (add_capture): Handle variadic capture.
(add_default_capture, lambda_capture_field_type): Likewise.
(build_capture_proxy, register_capture_members): Likewise.
* pt.c (register_specialization): Allow FIELD_DECL.
(retrieve_specialization): Likewise.
(find_parameter_packs_r): Handle FIELD_DECL and VAR_DECL.
(tsubst_pack_expansion): Handle FIELD_DECL packs.
(gen_elem_of_pack_expansion_instantiation): Likewise.
(instantiate_class_template_1): Likewise.
(tsubst_decl, tsubst_copy): Likewise.
(tsubst_expr) [DECL_EXPR]: Handle capture proxy packs.
(tsubst_copy_and_build) [VAR_DECL]: Likewise.
* semantics.c (finish_non_static_data_member): Don't try to represent
the type of a COMPOUND_REF of a FIELD_DECL pack.
diff --git a/gcc/cp/lambda.c b/gcc/cp/lambda.c
index bf75834..1af301d 100644
--- a/gcc/cp/lambda.c
+++ b/gcc/cp/lambda.c
@@ -215,7 +215,8 @@ lambda_capture_field_type (tree expr, bool explicit_init_p)
}
else
type = non_reference (unlowered_expr_type (expr));
- if (!type || WILDCARD_TYPE_P (type) || type_uses_auto (type))
+ if (!type || WILDCARD_TYPE_P (type) || type_uses_auto (type)
+ || DECL_PACK_P (expr))
{
type = cxx_make_type (DECLTYPE_TYPE);
DECLTYPE_TYPE_EXPR (type) = expr;
@@ -320,15 +321,21 @@ tree
lambda_proxy_type (tree ref)
{
tree type;
+ if (ref == error_mark_node)
+return error_mark_node;
if (REFERENCE_REF_P (ref))
ref = TREE_OPERAND (ref, 0);
+ gcc_assert (TREE_CODE (ref) == COMPONENT_REF);
type = TREE_TYPE (ref);
- if (type && !WILDCARD_TYPE_P (non_reference (type)))
-return type;
- type = cxx_make_type (DECLTYPE_TYPE);
- DECLTYPE_TYPE_EXPR (type) = ref;
- DECLTYPE_FOR_LAMBDA_PROXY (type) = true;
- SET_TYPE_STRUCTURAL_EQUALITY (type);
+ if (!type || WILDCARD_TYPE_P (non_reference (type)))
+{
+ type = cxx_make_type (DECLTYPE_TYPE);
+ DECLTYPE_TYPE_EXPR (type) = ref;
+ DECLTYPE_FOR_LAMBDA_PROXY (type) = true;
+ SET_TYPE_STRUCTURAL_EQUALITY (type);
+}
+ if (DECL_PACK_P (TREE_OPERAND (ref, 1)))
+type = make_pack_expansion (type);
return type;
}
@@ -341,6 +348,9 @@ build_capture_proxy (tree member)
{
tree var, object, fn, closure, name, lam, type;
+ if (PACK_EXPANSION_P (member))
+member = PACK_EXPANSION_PATTERN (member);
+
closure = DECL_CONTEXT (member);
fn = lambda_function (closure);
lam = CLASSTYPE_LAMBDA_EXPR (closure);
@@ -422,12 +432,20 @@ vla_capture_type (tree array_type)
and return it. */
tree
-add_capture (tree lambda, tree id, tree initializer, bool by_reference_p,
+add_capture (tree lambda, tree id, tree orig_init, bool by_reference_p,
bool explicit_init_p)
{
char *buf;
tree type, member, name;
bool vla = false;
+ bool variadic = false;
+ tree initializer = orig_init;
+
+ if (PACK_EXPANSION_P (initializer))
+{
+ initializer = PACK_EXPANSION_PATTERN (initializer);
+ variadic = true;
+}
if (TREE_CODE (initializer) == TREE_LIST)
initializer = build_x_compound_expr_from_list (initializer, ELK_INIT,
@@ -498,6 +516,9 @@ add_capture (tree lambda, tree id, tree initializer, bool by_reference_p,
IDENTIFIER_MARKED (name) = true;
}
+ if (variadic)
+type = make_pack_expansion (type);
+
/* Make member variable. */
member = build_decl (input_location, FIELD_DECL, name, type);
DECL_VLA_CAPTURE_P (member) = vla;
@@ -518,8 +539,14 @@ add_capture (tree lambda, tree id, tree initializer, bool by_reference_p,
&& current_class_type == LAMBDA_EXPR_CLOSURE (lambda))
finish_member_declaration (member);
+ tree listmem = member;
+ if (variadic)
+{
+ listmem = make_pack_expansion (member);
+ initializer = orig_init;
+}
LAMBDA_EXPR_CAPTURE_LIST (lambda)
-= tree_cons (member, initializer, LAMBDA_EXPR_CAPTURE_LIST (lambda));
+= tree_cons (listmem, initializer, LAMBDA_EXPR_CAPTURE_LIST (lambda));
if (LAMBDA_EXPR_CLOSURE (lambda))
return build_capture_proxy (member);
@@ -538,9 +565,14 @@ register_capture_members (tree captures)
return;
register_capture_members (TREE_CHAIN (captures));
+
+ tree field = TREE_PURPOSE (captures);
+ if (PACK_EXPANSION_P (field))
+field = PACK_EXPANSION_PATTERN (field);
+
/* We set this in add_capture to avoid duplicates. */
- IDENTIFIER_MARKED (DECL_NAME (TREE_PURPOSE (captures))) = false;
Re: Recent IPA regression with internal functions
> 2013-09-13 Jakub Jelinek > > * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early > for internal calls. That seems resonable. I wonder if we want to consider internal calls to form callgarph edges at all: perhaps we can just modify cgraph builder+verifier to skip them and consider them to be normal instruction Honza > > --- gcc/ipa-prop.c.jj 2013-09-13 16:48:54.0 +0200 > +++ gcc/ipa-prop.c2013-09-13 17:28:28.086058903 +0200 > @@ -1551,6 +1551,8 @@ ipa_compute_jump_functions_for_edge (str > return; >vec_safe_grow_cleared (args->jump_functions, arg_num); > > + if (gimple_call_internal_p (call)) > +return; >if (ipa_func_spec_opts_forbid_analysis_p (cs->caller)) > return; > > > > Jakub
Fwd: Re: [Patch, Fortran] PR57697 - Fix an issue with defined assignment [fwd: bur...@net-b.de]
Yet another try to send this email - this time from a different
server. For completeness:
* The original email didn't made it, nor a repost. But the mail
server didn't bounce back.
* For another email, only the reply made it - but not the original
email: http://gcc.gnu.org/ml/fortran/2013-09/msg00025.html
Locally, it works as I BCC'ed myself to the emails.
--- Begin Message ---
--- Begin Message ---
Re-sent as it didn't show up in the archive. (I wonder why this and
another email didn't made it, but the follow-up to that email did.)
Tobias Burnus wrote:
Hi Thomas, hello all,
As it turned out, my patch wasn't working for the real-world code. I
created a follow-up patch. See below.
* * *
Thomas Koenig wrote:
the patch is OK, also for 4.8. Thanks a lot for fixing this.
Thanks for the review!
Just a couple of nits:
- You may want to remove the output from the test case.
Done. (Well, I missed one print line.)
- The two consecutive ifs in
if (left != 0B)
{
if (_F.DA0 != 0B) goto L.2;
_F.DA0 = (struct parent *) __builtin_malloc (4);
L.2:;
*_F.DA0 = *left;
}
L.1:;
if (left != 0B) goto L.4;
are a little bit inelegant. It is not really important, because
they will be merged on optimization, but if you find an easy
way to do this in the FE code, you might want to consider doing
so. I would advise against spending a lot of work on this, though :-)
That's a bit difficult - part of the "if"s are generated at resolution
time (resolve.c, like my patch) others are generated in trans-expr.c
for realloc on assignment. I don't see a simple way to avoid the two
conditions, unfortunately.
Committed to the trunk as Rev. 202601. (By the way, the automatic
addition of the committal to the PR now works again :-)
* * *
As testing showed, it didn't fix the real-world code: ForTrilinos's
ForTrilinos_ADT_3D_Burgers_6th_Pade did still fail as it has:
*_F.DA65 = matrix_diff_x (&parm.621);
_F.DA66 = ax->epetra_rowmatrix.universal; // Deref of "ax"!
The reason for the failure is that ax == NULL but only "ax" is
allocatable while universal isn't. That's now fixed by the attached
patch. With that patch, ForTrilions's
ForTrilinos_ADT_3D_Burgers_6th_Pade and
ForTrilinos_concrete_burgers_solver now pass (instead of segfault).
Additionally, I changed ISYM_ASSOCIATED to ISYM_ALLOCATED which
matches the internal name and is a bit more consistent. As either one
boils down to a null-pointer check, it shouldn't lead to any code-gen
difference on tree level.
Build and regtested on x86-64-gnu-linux.
OK?
Tobias
2013-09-15 Tobias Burnus
PR fortran/57697
* resolve.c (generate_component_assignments): Correctly handle the
case that the LHS is not allocated.
2013-09-15 Tobias Burnus
PR fortran/57697
* gfortran.dg/defined_assignment_10.f90: Comment print statement.
diff --git a/gcc/fortran/resolve.c b/gcc/fortran/resolve.c
index f2892e2..fbd9a6a 100644
--- a/gcc/fortran/resolve.c
+++ b/gcc/fortran/resolve.c
@@ -9547,17 +9547,20 @@ generate_component_assignments (gfc_code **code, gfc_namespace *ns)
t1, (*code)->expr1,
NULL, NULL, (*code)->loc);
- /* For allocatable LHS, check whether it is allocated. */
- if (gfc_expr_attr((*code)->expr1).allocatable)
+ /* For allocatable LHS, check whether it is allocated. Note
+ that allocatable components with defined assignment are
+ not yet support. See PR 57696. */
+ if ((*code)->expr1->symtree->n.sym->attr.allocatable)
{
gfc_code *block;
+ gfc_expr *e =
+ gfc_lval_expr_from_sym ((*code)->expr1->symtree->n.sym);
block = gfc_get_code (EXEC_IF);
block->block = gfc_get_code (EXEC_IF);
block->block->expr1
= gfc_build_intrinsic_call (ns,
-GFC_ISYM_ASSOCIATED, "allocated",
-(*code)->loc, 2,
-gfc_copy_expr ((*code)->expr1), NULL);
+GFC_ISYM_ALLOCATED, "allocated",
+(*code)->loc, 1, e);
block->block->next = temp_code;
temp_code = block;
}
@@ -9570,9 +9573,11 @@ generate_component_assignments (gfc_code **code, gfc_namespace *ns)
this_code->ext.actual->expr = gfc_copy_expr (t1);
add_comp_ref (this_code->ext.actual->expr, comp1);
- /* If the LHS is not allocated, we pointer-assign the LHS address
- to the temporary - after the LHS has been allocated. */
- if (gfc_expr_attr((*code)->expr1).allocatable)
+ /* If the LHS variable is allocatable and wasn't allocated and
+ the temporary is allocatable, pointer assign the address of
+ the freshly allocated LHS to the temporary. */
+ if ((*code)->expr1->symtree->n.sym->attr.allocatable
+ && gfc_expr_attr ((*code)->expr1).allocatable)
{
gfc_code *block;
gfc_expr *cond;
@@ -9583,9 +9588,8 @@ generate_component_assignments (gfc_code **code, gfc_namespace *n
Re: [Patch, Fortran] PR43366 - add invalid-diagnostic for poly assignment
Now with attachment…
Tobias Burnus wrote:
Fortran 2008 permits assignment to polymorphic variables with some
constraints. The patch, which was sitting in my tree, adds diagnostic
to reject invalid use. For valid code, it runs into the existing
not-yet-implemented error.
Build + regtested on x86-64-gnu-linux.
OK for the trunk?
2013-09-15 Tobias Burnus
PR fortran/43366
* resolve.c (resolve_ordinary_assign): Add invalid-diagnostic for
polymorphic assignment.
diff --git a/gcc/fortran/resolve.c b/gcc/fortran/resolve.c
index f2892e2..1157f28 100644
--- a/gcc/fortran/resolve.c
+++ b/gcc/fortran/resolve.c
@@ -9010,14 +9010,15 @@ resolve_ordinary_assign (gfc_code *code, gfc_namespace *ns)
bool rval = false;
gfc_expr *lhs;
gfc_expr *rhs;
int llen = 0;
int rlen = 0;
int n;
gfc_ref *ref;
+ symbol_attribute attr;
if (gfc_extend_assign (code, ns))
{
gfc_expr** rhsptr;
if (code->op == EXEC_ASSIGN_CALL)
{
@@ -9174,15 +9175,34 @@ resolve_ordinary_assign (gfc_code *code, gfc_namespace *ns)
gfc_current_ns->proc_name->attr.implicit_pure = 0;
/* Fortran 2008, C1283. */
if (gfc_is_coindexed (lhs))
gfc_current_ns->proc_name->attr.implicit_pure = 0;
}
- /* F03:7.4.1.2. */
+ /* F2008, 7.2.1.2. */
+ attr = gfc_expr_attr (lhs);
+ if (lhs->ts.type == BT_CLASS && attr.allocatable)
+{
+ if (attr.codimension)
+ {
+ gfc_error ("Assignment to polymorphic coarray at %L is not "
+ "permitted", &lhs->where);
+ return false;
+ }
+ if (!gfc_notify_std (GFC_STD_F2008, "Assignment to an allocatable "
+ "polymorphic variable at %L", &lhs->where))
+ return false;
+ if (!gfc_option.flag_realloc_lhs)
+ {
+ gfc_error ("Assignment to an allocatable polymorphic variable at %L "
+ "requires -frealloc-lhs", &lhs->where);
+ return false;
+ }
+}
/* FIXME: Valid in Fortran 2008, unless the LHS is both polymorphic
and coindexed; cf. F2008, 7.2.1.2 and PR 43366. */
if (lhs->ts.type == BT_CLASS)
{
gfc_error ("Variable must not be polymorphic in intrinsic assignment at "
"%L - check that there is a matching specific subroutine "
"for '=' operator", &lhs->where);
RE: [PATCH, PR 57748] Check for out of bounds access
Hello Richard, attached is my second attempt at fixing PR 57748. This time the movmisalign path is completely removed and a similar bug in the read handling of misaligned structures with a non-BLKmode is fixed too. There are several new test cases for the different possible failure modes. This patch was boot-strapped and regression tested on x86_64-unknown-linux-gnu and i686-pc-linux-gnu. Additionally I generated eCos and an eCos-application (on ARMv5 using packed structures) with an arm-eabi cross compiler, and looked for differences in the disassembled code with and without this patch, but there were none. OK for trunk? Regards Bernd.2013-09-15 Bernd Edlinger PR middle-end/57748 * expr.c (expand_assignment): Remove misalignp code path. Check for bitregion in offset arithmetic. (expand_expr_real_1): Use EXAND_MEMORY on base object. testsuite: PR middle-end/57748 * gcc.dg/torture/pr57748-1.c: New test. * gcc.dg/torture/pr57748-2.c: New test. * gcc.dg/torture/pr57748-3.c: New test. * gcc.dg/torture/pr57748-3a.c: New test. * gcc.dg/torture/pr57748-4.c: New test. * gcc.dg/torture/pr57748-4a.c: New test. patch-pr57748.diff Description: Binary data
Re: [RFC] Offloading Support in libgomp
> Libgomp will start N-1 new threads, and all of them would want to look up > mappings for i1,i2,...iK in the splay tree. The first one wouldn't find > anything and would map and insert all the values to the tree. But the > following > ones would look-up these addresses in the exactly same order, which will lead > to > totally unbalanced tree. > > Am I missing anything or is it a real problem? On second thought, this access order doesn't necessarily mean accessing in ascending/descending keys order, so there is no problem here. Thanks, Michael
Re: Generic lambda and implicit function template commits reverted
On 09/15/2013 06:22 AM, Adam Butcher wrote: [PATCH 1/5] Fix uninitialized variables causing breakage with -Werror. [PATCH 2/5] Don't accept 'auto' as the 'type' of a template parameter. OK. [PATCH 3/5] Fix location diagnostics by returning to the deprecated 'input_location' global; must be a better fix for this. Don't know why 'location_of (type)' gave ":" rather than "file:line:col:". My current workaround is to return to using 'input_location'. This gives the correct result but I doubt it is acceptable. This seems to be because make_auto_1 sets the location of the auto type to BUILTINS_LOCATION; I don't remember why I did that. Changing it to use input_location seems appropriate. [PATCH 4/5] Lift CALL_FROM_THUNK_P setting to above the potential 'build_cplus_new' call to prevent ICE due to unexpected tree type. [PATCH 5/5] Handle forward declaration of implicit function templates. Previously kept template parameter types in scope. OK. Jason
Re: Generic lambda and implicit function template commits reverted
Hi Jason,
Could you cast your eyes over these changes please? I intend to roll
them up into the appropriate patches. I will make sure I bootstrap
and "before-and-after" the g++.dg testsuite before pushing next time!
[PATCH 1/5] Fix uninitialized variables causing breakage with -Werror.
Not actually sure that this should be necessary. Initialization of
'direct_argvec' and 'call' and subsequent references are behind
'!generic_lambda_p' and 'generic_lambda_p' tests respectively.
'generic_lambda_p' is declared 'const bool'.
To pacify -Wmaybe-uninitialized, I have initialized them to 0.
[PATCH 2/5] Don't accept 'auto' as the 'type' of a template parameter.
The implicit function template code was incorrectly firing in a
template
parameter list.
[PATCH 3/5] Fix location diagnostics by returning to the deprecated
'input_location' global; must be a better fix for this.
Don't know why 'location_of (type)' gave ":" rather than
"file:line:col:". My current workaround is to return to using
'input_location'. This gives the correct result but I doubt it is
acceptable.
[PATCH 4/5] Lift CALL_FROM_THUNK_P setting to above the potential
'build_cplus_new' call to prevent ICE due to unexpected tree type.
Plain old bug which I introduced when reorganizing the conversion op
code.
[PATCH 5/5] Handle forward declaration of implicit function templates.
Previously kept template parameter types in scope.
Another bug. Template parameter list erroneously left in scope.
Cheers,
Adam
[PATCH 1/5] Fix uninitialized variables causing breakage with -Werror.
---
diff --git a/gcc/cp/lambda.c b/gcc/cp/lambda.c
index 2d20333..0da22fd 100644
--- a/gcc/cp/lambda.c
+++ b/gcc/cp/lambda.c
@@ -792,8 +792,8 @@ maybe_add_lambda_conv_op (tree type)
particular, parameter pack expansions are marked
PACK_EXPANSION_LOCAL_P in
the body CALL, but not in DECLTYPE_CALL. */
- vec *direct_argvec;
- tree decltype_call = 0, call;
+ vec *direct_argvec = 0;
+ tree decltype_call = 0, call = 0;
tree fn_result = TREE_TYPE (TREE_TYPE (callop));
if (generic_lambda_p)
--
[PATCH 2/5] Don't accept 'auto' as the 'type' of a template parameter.
---
diff --git a/gcc/cp/decl.c b/gcc/cp/decl.c
index e6e24f8..6a4e863 100644
--- a/gcc/cp/decl.c
+++ b/gcc/cp/decl.c
@@ -10323,7 +10323,12 @@ grokdeclarator (const cp_declarator
*declarator,
if (type_uses_auto (type))
{
- if (current_class_type && LAMBDA_TYPE_P (current_class_type))
+ if (template_parm_flag)
+ {
+ error ("template parameter declared %");
+ type = error_mark_node;
+ }
+ else if (current_class_type && LAMBDA_TYPE_P (current_class_type))
{
if (cxx_dialect < cxx1y)
pedwarn (location_of (type), 0,
--
[PATCH 3/5] Fix location diagnostics by returning to the deprecated
'input_location' global; must be a better fix for this.
Using 'location_of (type)' yields ": " rather than
"file:line:col: "
---
diff --git a/gcc/cp/decl.c b/gcc/cp/decl.c
index 6a4e863..a948580 100644
--- a/gcc/cp/decl.c
+++ b/gcc/cp/decl.c
@@ -10331,18 +10331,18 @@ grokdeclarator (const cp_declarator
*declarator,
else if (current_class_type && LAMBDA_TYPE_P (current_class_type))
{
if (cxx_dialect < cxx1y)
- pedwarn (location_of (type), 0,
+ pedwarn (input_location, 0,
"use of % in lambda parameter declaration "
"only available with "
"-std=c++1y or -std=gnu++1y");
}
else if (cxx_dialect < cxx1y)
- pedwarn (location_of (type), 0,
+ pedwarn (input_location, 0,
"use of % in parameter declaration "
"only available with "
"-std=c++1y or -std=gnu++1y");
else
- pedwarn (location_of (type), OPT_Wpedantic,
+ pedwarn (input_location, OPT_Wpedantic,
"ISO C++ forbids use of % in parameter "
"declaration");
}
--
[PATCH 4/5] Lift CALL_FROM_THUNK_P setting to above the potential
'build_cplus_new' call to prevent ICE due to unexpected tree type.
---
diff --git a/gcc/cp/lambda.c b/gcc/cp/lambda.c
index 0da22fd..0154840 100644
--- a/gcc/cp/lambda.c
+++ b/gcc/cp/lambda.c
@@ -883,16 +883,17 @@ maybe_add_lambda_conv_op (tree type)
tf_warning_or_error);
--processing_template_decl;
}
+ CALL_FROM_THUNK_P (call) = 1;
}
else
{
call = build_call_a (callop,
direct_argvec->length (),
direct_argvec->address ());
+ CALL_FROM_THUNK_P (call) = 1;
if (MAYBE_CLASS_TYPE_P (TREE_TYPE (call)))
call = build_cplus_new (TREE_TYPE (call), call, tf_warning_or_error);
}
- CALL_FROM_THUNK_P (call
[Patch, Fortran] PR58356 - fix ICE with finalizer in type extension
A rather obvious patch.
Build and regtested on x86-64-gnu-linux.
OK for the trunk?
Tobias
2013-09-15 Tobias Burnus
PR fortran/58356
* class.c (generate_finalization_wrapper): Init proc_tree if
not yet resolved.
2013-09-15 Tobias Burnus
PR fortran/58356
* gfortran.dg/finalize_19.f90: New.
diff --git a/gcc/fortran/class.c b/gcc/fortran/class.c
index 629b052..7117e83 100644
--- a/gcc/fortran/class.c
+++ b/gcc/fortran/class.c
@@ -1881,6 +1881,8 @@ generate_finalization_wrapper (gfc_symbol *derived, gfc_namespace *ns,
for (fini = derived->f2k_derived->finalizers; fini; fini = fini->next)
{
+ if (!fini->proc_tree)
+ fini->proc_tree = gfc_find_sym_in_symtree (fini->proc_sym);
if (fini->proc_tree->n.sym->attr.elemental)
{
fini_elem = fini;
diff --git a/gcc/testsuite/gfortran.dg/finalize_19.f90 b/gcc/testsuite/gfortran.dg/finalize_19.f90
new file mode 100644
index 000..1eeb6af
--- /dev/null
+++ b/gcc/testsuite/gfortran.dg/finalize_19.f90
@@ -0,0 +1,21 @@
+! { dg-do compile }
+!
+! PR fortran/58356
+!
+! Contributed by Andrew Benson
+!
+module ct
+ type :: cfl
+ contains
+ final :: cfld
+ end type cfl
+ type, extends(cfl) :: cfde
+ contains
+ end type cfde
+contains
+ subroutine cfld(self)
+implicit none
+type(cfl), intent(inout) :: self
+return
+ end subroutine cfld
+end module ct
Re: [PATCH 2/6] Andes nds32: machine description of nds32 porting (3).
Some comments about the final part:
Chung-Ju Wu writes:
> +(define_constraint "Ibms"
> + "The immediate value with power of 2"
> + (and (match_code "const_int")
> + (match_test "(TARGET_ISA_V3 || TARGET_ISA_V3M)
> + && (floor_log2 (ival) < 8)
> + && (ival > 0)
> + && (ival == (1 << floor_log2 (ival)))")))
I think the last three lines are equivalent to:
IN_RANGE (exact_log2 (ival), 0, 7)
> +(define_constraint "Ifex"
> + "The immediate value with power of 2 minus 1"
> + (and (match_code "const_int")
> + (match_test "(TARGET_ISA_V3 || TARGET_ISA_V3M)
> + && (ival < 256)
> + && (ival > 0)
> + && (floor_log2 (ival + 1) - 1 < 8)
> + && ((ival + 1) == (1 << floor_log2 (ival + 1)))")))
And here:
IN_RANGE (exact_log2 (ival + 1), 1, 8)
> +#ifndef nds32_OPTS_H
> +#define nds32_OPTS_H
Very minor, but the mixture of case looks odd.
> +;; -
> +;; Boolean DImode instructions.
> +;; -
> +
> +;; Boolean and,ior,xor insns.
> +
> +;; 'and' operation.
> +
> +(define_expand "anddi3"
> + [(set (match_operand:DI 0 "register_operand" "")
> + (and:DI (match_operand:DI 1 "register_operand" "")
> + (match_operand:DI 2 "register_operand" "")))]
> + ""
> + ""
> +)
> +
> +; Use '#' to split instruction.
> +(define_insn "*anddi3_insn"
> + [(set (match_operand:DI 0 "register_operand" "=&r, &r")
> + (and:DI (match_operand:DI 1 "register_operand" " %0, r")
> + (match_operand:DI 2 "register_operand" " r, r")))]
> + ""
> + "#"
> + [(set_attr "length" "8")])
> +
> +; Use '#' to split instruction.
> +; The zero extend of operand 2 clears the high word of the output operand.
> +(define_insn_and_split "*anddi_zesidi_di"
> + [(set (match_operand:DI 0 "register_operand"
> "=&r, &r")
> + (and:DI (zero_extend:DI (match_operand:SI 2 "register_operand" " r,
> r"))
> + (match_operand:DI 1 "register_operand" " 0,
> r")))]
> + ""
> + "#"
> + "reload_completed"
> + [(set (match_dup 0) (and:SI (match_dup 1) (match_dup 2)))
> + (set (match_dup 3) (const_int 0))]
> +{
> + operands[3] = gen_highpart (SImode, operands[0]);
> + operands[0] = gen_lowpart (SImode, operands[0]);
> + operands[1] = gen_lowpart (SImode, operands[1]);
> +}
> + [(set_attr "length" "8")])
> +
> +; Use '#' to split instruction.
> +(define_insn "*anddi_sesidi_di"
> + [(set (match_operand:DI 0 "register_operand"
> "=&r, &r")
> + (and:DI (sign_extend:DI (match_operand:SI 2 "register_operand" " r,
> r"))
> + (match_operand:DI 1 "register_operand" " 0,
> r")))]
> + ""
> + "#"
> + [(set_attr "length" "8")])
> +
> +
> +;; 'ior' operation.
> +
> +(define_expand "iordi3"
> + [(set (match_operand:DI 0 "register_operand" "")
> + (ior:DI (match_operand:DI 1 "register_operand" "")
> + (match_operand:DI 2 "register_operand" "")))]
> + ""
> + ""
> +)
> +
> +; Use '#' to split instruction.
> +(define_insn "*iordi3_insn"
> + [(set (match_operand:DI 0 "register_operand" "=&r, &r")
> + (ior:DI (match_operand:DI 1 "register_operand" " %0, r")
> + (match_operand:DI 2 "register_operand" " r, r")))]
> + ""
> + "#"
> + [(set_attr "length" "8")])
> +
> +; Use '#' to split instruction.
> +(define_insn "*iordi_zesidi_di"
> + [(set (match_operand:DI 0 "register_operand"
> "=&r, &r")
> + (ior:DI (zero_extend:DI (match_operand:SI 2 "register_operand" " r,
> r"))
> + (match_operand:DI 1 "register_operand" " 0,
> ?r")))]
> + ""
> + "#"
> + [(set_attr "length" "8")])
> +
> +; Use '#' to split instruction.
> +(define_insn "*iordi_sesidi_di"
> + [(set (match_operand:DI 0 "register_operand"
> "=&r, &r")
> + (ior:DI (sign_extend:DI (match_operand:SI 2 "register_operand" " r,
> r"))
> + (match_operand:DI 1 "register_operand" " 0,
> r")))]
> + ""
> + "#"
> + [(set_attr "length" "8")])
> +
> +
> +;; 'xor' operation.
> +
> +(define_expand "xordi3"
> + [(set (match_operand:DI 0 "register_operand" "")
> + (xor:DI (match_operand:DI 1 "register_operand" "")
> + (match_operand:DI 2 "register_operand" "")))]
> + ""
> + ""
> +)
> +
> +; Use '#' to split instruction.
> +(define_insn "*xordi3_insn"
> + [(set (match_operand:DI 0 "register_operand" "=&r, &r")
> + (xor:DI (match_operand:DI 1 "register_operand" " %0, r")
> + (match_operand:DI 2 "register_operand" " r, r")))]
> + ""
> + "#"
> + [(set_attr "length" "8")])
> +
> +; Use '#' to split instruction.
> +(define_insn "*xordi_zesidi_di"
> + [(set (match_operand:DI 0 "register_operand"
Re: [Patch, Fortran] PR57697 - Fix an issue with defined assignment
Hi Tobias,
the patch is OK, also for 4.8. Thanks a lot for fixing this.
Just a couple of nits:
- You may want to remove the output from the test case.
- The two consecutive ifs in
>
> if (left != 0B)
> {
> if (_F.DA0 != 0B) goto L.2;
> _F.DA0 = (struct parent *) __builtin_malloc (4);
> L.2:;
> *_F.DA0 = *left;
> }
> L.1:;
> if (left != 0B) goto L.4;
are a little bit inelegant. It is not really important, because
they will be merged on optimization, but if you find an easy
way to do this in the FE code, you might want to consider doing
so. I would advise against spending a lot of work on this, though :-)
Thomas
Re: [RFC] Offloading Support in libgomp
Hi Jakub,
This patch looks ok for me in general, but I am a bit worried about using
splay-trees. Couldn't we end up with their worst case linear performance
instead desired log?
Imagine the following scenario:
#pragma parallel ... // to produce N-threads
{
# pragma target map (i1, i2, ...iK)
{
// some code to offload using i1, i2, ... iK
}
}
Libgomp will start N-1 new threads, and all of them would want to look up
mappings for i1,i2,...iK in the splay tree. The first one wouldn't find
anything and would map and insert all the values to the tree. But the following
ones would look-up these addresses in the exactly same order, which will lead to
totally unbalanced tree.
Am I missing anything or is it a real problem?
Thanks, Michael
> Jakub
Re: [v3] More noexcept for vectors
I had to separate the constructor that takes an allocator from the default
constructor in debug/profile, since in release the noexcept only applies
to one of them (and the testsuite asserts that release and debug agree on
this). An alternative would be to make the release vector default
constructor conditionally noexcept (depending on the allocator). Or to use
explicit vector(const Allocator& = Allocator()); also in normal mode
instead of splitting it in two.
I also removed a noexcept in profile mode where the release mode doesn't
have noexcept.
Tested, no regression.
2013-09-15 Marc Glisse
PR libstdc++/58338
* include/bits/stl_vector.h
(_Vector_impl::_Vector_impl(_Tp_alloc_type const&),
_Vector_impl::_Vector_impl(_Tp_alloc_type&&),
_Vector_impl::_M_swap_data,
_Vector_base::_Vector_base(const allocator_type&),
_Vector_base::_Vector_base(allocator_type&&),
_Vector_base::_Vector_base(_Vector_base&&),
vector::vector(const allocator_type&), vector::operator[],
vector::operator[] const, vector::front, vector::front const,
vector::back, vector::back const, vector::pop_back,
vector::_M_erase_at_end): Mark as noexcept.
(vector::~vector): Remove useless noexcept.
* include/debug/vector (vector::vector()): Split.
(vector::vector(const _Allocator&), vector::operator[],
vector::operator[] const, vector::front, vector::front const,
vector::back, vector::back const, vector::pop_back,
_M_requires_reallocation, _M_update_guaranteed_capacity,
_M_invalidate_after_nth): Mark as noexcept.
(vector::~vector): Remove useless noexcept.
* include/profile/vector (vector::vector()): Split.
(vector::vector(const _Allocator&), vector::operator[],
vector::operator[] const, vector::front, vector::front const,
vector::back, vector::back const): Mark as noexcept.
(vector::vector(vector&&, const _Allocator&)): Remove wrong noexcept.
(vector::~vector): Remove useless noexcept.
--
Marc GlisseIndex: include/bits/stl_vector.h
===
--- include/bits/stl_vector.h (revision 202588)
+++ include/bits/stl_vector.h (working copy)
@@ -80,32 +80,32 @@ _GLIBCXX_BEGIN_NAMESPACE_CONTAINER
: public _Tp_alloc_type
{
pointer _M_start;
pointer _M_finish;
pointer _M_end_of_storage;
_Vector_impl()
: _Tp_alloc_type(), _M_start(0), _M_finish(0), _M_end_of_storage(0)
{ }
- _Vector_impl(_Tp_alloc_type const& __a)
+ _Vector_impl(_Tp_alloc_type const& __a) _GLIBCXX_NOEXCEPT
: _Tp_alloc_type(__a), _M_start(0), _M_finish(0), _M_end_of_storage(0)
{ }
#if __cplusplus >= 201103L
- _Vector_impl(_Tp_alloc_type&& __a)
+ _Vector_impl(_Tp_alloc_type&& __a) noexcept
: _Tp_alloc_type(std::move(__a)),
_M_start(0), _M_finish(0), _M_end_of_storage(0)
{ }
#endif
- void _M_swap_data(_Vector_impl& __x)
+ void _M_swap_data(_Vector_impl& __x) _GLIBCXX_NOEXCEPT
{
std::swap(_M_start, __x._M_start);
std::swap(_M_finish, __x._M_finish);
std::swap(_M_end_of_storage, __x._M_end_of_storage);
}
};
public:
typedef _Alloc allocator_type;
@@ -117,36 +117,36 @@ _GLIBCXX_BEGIN_NAMESPACE_CONTAINER
_M_get_Tp_allocator() const _GLIBCXX_NOEXCEPT
{ return *static_cast(&this->_M_impl); }
allocator_type
get_allocator() const _GLIBCXX_NOEXCEPT
{ return allocator_type(_M_get_Tp_allocator()); }
_Vector_base()
: _M_impl() { }
- _Vector_base(const allocator_type& __a)
+ _Vector_base(const allocator_type& __a) _GLIBCXX_NOEXCEPT
: _M_impl(__a) { }
_Vector_base(size_t __n)
: _M_impl()
{ _M_create_storage(__n); }
_Vector_base(size_t __n, const allocator_type& __a)
: _M_impl(__a)
{ _M_create_storage(__n); }
#if __cplusplus >= 201103L
- _Vector_base(_Tp_alloc_type&& __a)
+ _Vector_base(_Tp_alloc_type&& __a) noexcept
: _M_impl(std::move(__a)) { }
- _Vector_base(_Vector_base&& __x)
+ _Vector_base(_Vector_base&& __x) noexcept
: _M_impl(std::move(__x._M_get_Tp_allocator()))
{ this->_M_impl._M_swap_data(__x._M_impl); }
_Vector_base(_Vector_base&& __x, const allocator_type& __a)
: _M_impl(__a)
{
if (__x.get_allocator() == __a)
this->_M_impl._M_swap_data(__x._M_impl);
else
{
@@ -246,21 +246,21 @@ _GLIBCXX_BEGIN_NAMESPACE_CONTAINER
* @brief Default constructor creates no elements.
*/
vector()
: _Base() { }
/**
* @brief Creates a %vector with no elements.
* @param __a An allocator object.
*/
explicit
- vector(const alloc
