[PR rtl-optimization/122321][RISC-V] Bounds check another access to ira_reg_equiv array

So another case where we're indexing into the ira_reg_equiv array without
checking bounds.  I sincerely hope we're not playing wack-a-mole here, but two
failures in a couple months for the same core problem is worrisome.

Bootstrapped and regression tested on x86_64 and riscv64 as well as run through
all the embedded targets in my tester without regressions.

PR rtl-optimization/122321
gcc/
* lra-constraints.cc (update_equiv): Make sure REGNO is in
ira_reg_equiv before trying to update ira_reg_equiv.

gcc/testsuite/
* gcc.target/riscv/rvv/autovec/pr122321.c: New test.

Fortran: implement constraint F2018:C1585 on pure function results [PR78640]

PR fortran/78640

gcc/fortran/ChangeLog:

* resolve.cc (resolve_fl_procedure): Check function result of a
pure function against F2018:C1585.

gcc/testsuite/ChangeLog:

* gfortran.dg/pure_result.f90: New test.

c++/modules: Stream BASELINK_OPTYPE [PR122381]

This is used in template conversion operators to determine what type the
user requested.

PR c++/122381

gcc/cp/ChangeLog:

* module.cc (trees_out::core_vals): Write BASELINK_OPTYPE (aka
TREE_CHAIN).
(trees_in::core_vals): Read it.

gcc/testsuite/ChangeLog:

* g++.dg/modules/convop-2_a.H: New test.
* g++.dg/modules/convop-2_b.C: New test.

Signed-off-by: Nathaniel Shead <nathanieloshead@gmail.com>

c: GNU extension allowing compound literals of variable size

This patch implements a GNU extension by allowing compound literals
to be VLAs which then can be initialized with an empty initializer.
This addresses a use case where one would now need to use alloca, but
this also has limitations (e.g. allocated memory accumulates in a loop).
The error for a compound literal with variable size is changed to a
pedwarn, and a new error for static and constexpr is added.

gcc/c/ChangeLog:
* c-decl.cc (build_compound_literal): Add error.
* c-parser.cc (c_parser_braced_init): Take bool argument for
variable size instead of DECL.
(c_parser_initializer,c_parser_initval): Adapt.
(c_parser_postfix_expression_after_paren_type): Change
error to pedwarn.
* c-typeck.cc (process_init_element): Add error for
variable-size compound literal with static or constexpr.

gcc/ChangeLog:
* doc/extend.texi: Document new extension.

gcc/testsuite/ChangeLog:
* gcc.dg/gnu-compoundlit-1.c: New test.
* gcc.dg/gnu-compoundlit-2.c: New test.
* gcc.dg/pr68090.c: Adapt.
* gcc.dg/vla-init-4.c: Adapt.
* gcc.dg/vla-init-5.c: Adapt.

c: Fix ICE when using va_arg with arrays [PR97986]

When array type is passed to va_arg, this is undefined behavior.
Emit a warning, and insert a run-time trap after evaluating side effects,
but return the correct type for sizeof / typeof.  For C90 a VLA is an
error.

PR c/97986

gcc/c/ChangeLog:
* c-parser.cc (c_parser_postfix_expression): Adapt.
* c-tree.h (c_build_va_arg): Adapt prototype.
* c-typeck.cc (c_build_va_arg): Handle UB.

gcc/testsuite/ChangeLog:
* gcc.dg/pr97986-1.c: New test.
* gcc.dg/pr97986-2.c: New test.

xtensa: Add alternative negsf2 insn pattern

If both the source and destination are address (GP) registers, emitting
instructions that invert the MSB of the address register is two bytes
shorter if TARGET_DENSITY is enabled than emitting a NEG.S machine inst-
ruction that uses hardware FP registers with two reloads.

     /* example */
     float test(float a) {
       return -a;
     }

     ;; before
     test:
      entry sp, 32
      wfr f0, a2
      neg.s f0, f0
      rfr a2, f0
      retw.n

     ;; after
     test:
      entry sp, 32
      movi.n a8, 1
      slli a8, a8, 31
      add.n a2, a2, a8
      retw.n

By the way, in configurations that do not use hardware FP register, the
RTL expansion pass will emit such insns by default.

gcc/ChangeLog:

* config/xtensa/xtensa.md (negsf2):
Add another insn pattern that is valid when TARGET_DENSITY is
enabled and both the source and destination are address registers.

xtensa: Make use of ROUND.S instruction

Due to inconsistencies in the behavior of rounding half, making this
machine instruction available was retracted in a previous commit (5f3b5b0616fe883e86e95d9476371cf87059ca7f),
but it may be useful to have it available if strict implementation of
floating-point arithmetic is not required.

gcc/ChangeLog:

* config/xtensa/xtensa.md
(c_enum "unspec", int_iterator ANY_ROUND): Add UNSPEC_ROUND.
(int_attr m_round): Add a pair of UNSPEC_ROUND and "round".
(int_attr c_round): New integer iterator attribute, that expands
to "flag_unsafe_math_optimizations" in the case of UNSPEC_ROUND,
and to "1" otherwise.
(l<m_round>sfsi2, *l<m_round>sfsi2_2x, *l<m_round>sfsi2_scaled):
Append " && <c_round>" to the conditions.

xtensa: Remove redundant use of 'F'-constraint

Because it is redundant to specify 'F'-constraints on operands in single-
alternative match templates whose predicates imply CONST_DOUBLE_P().

gcc/ChangeLog:

* config/xtensa/xtensa.md (*fix<s_fix>_truncsfsi2_scaled,
*float<s_float>sisf2_scaled, *l<m_round>sfsi2_scaled):
Remove 'F'-constraint.

testsuite: Fix fold-vecperm-1.c for targets that don't have vectors

This testcase is testing at optimization but with targets that don't
have vectors it will fail because there will be zero VEC_PERM_EXPR.
So instead let's check earlier in say forwprop3.

Pushed as obvious after a test on x86_64-linux-gnu with -mno-sse.

gcc/testsuite/ChangeLog:

* gcc.dg/fold-vecperm-1.c: Test at forwprop3.

Signed-off-by: Andrew Pinski <andrew.pinski@oss.qualcomm.com>

ipa: Fix pritting of symtab_node type [PR122512]

The problem here r16-4747-g529c25ed6e0a06 changed
symtab_type to toplevel_type and then added 2 types
at the begining but didn't update symtab_type_names.
This obvious patch changes symtab_type_names to toplevel_type_names,
adds a _MAX type to toplevel_type and a comment about keeping them in sync.
Also adds an static assert to make sure if someone adds a toplevel_type
that toplevel_type_names get updated too.

Pushed as obvious after a build and test.

PR ipa/122512
gcc/ChangeLog:

* cgraph.h (enum toplevel_type): Add TOPLEVEL_MAX. Add comment
about keeping toplevel_type_names in sync.
* symtab.cc (symtab_type_names): Rename to ...
(toplevel_type_names): Add new types and add an assert about
the size.
(symtab_node::get_symtab_type_string): Use toplevel_type_names
instead of symtab_type_names.
(symtab_node::dump_base): Likewise.

Signed-off-by: Andrew Pinski <andrew.pinski@oss.qualcomm.com>

Daily bump.

fold-const: Remove div_if_zero_remainder declaration.

I missed the declaration of div_if_zero_remainder in fold-const.h
when it was removed in r16-4907-ga4be88ef0f2669. This removes it.

Pushed as obvious after a quick build and test.

PR tree-optimization/122437
gcc/ChangeLog:

* fold-const.h (div_if_zero_remainder): Remove.

Signed-off-by: Andrew Pinski <andrew.pinski@oss.qualcomm.com>

AArch64: support bf16 to sf extensions [PR121853]

It looks like during the upstreaming of BF16 we didn't implement the extend
optab for it.

As a result we go through soft-float emulation which results in massive
performance drop in projects using BF16.

As an example, for

float convert(__bf16 value) {
    return (float)value;
}

we generate:

convert(__bf16):
        stp     x29, x30, [sp, -16]!
        mov     x29, sp
        bl      __extendbfsf2
        ldp     x29, x30, [sp], 16
        ret

and after this patch

convert:
        movi    v31.4s, 0
        ext     v0.16b, v31.16b, v0.16b, #14
        ret

We generate an ext with movi because this has same latency as a shift however
it has twice the throughput.  The zero vector is zero latency as such in real
workloads this codegen is much better than using shifts.

As a reminder, BF16 -> FP32 is just shifting left 16 bits.

The expand pattern has to rely on generating multiple subregs due to a
restriction that subregs can't chang floating point size and type at the same
time.

I've tried alternative approaches like using the EXT as SF mode, but the
paradoxical subreg of BF -> SF isn't allowed and using an extend doesn't work
because extend is what we're defining.

gcc/ChangeLog:

PR target/121853
* config/aarch64/aarch64-simd.md (extendbfsf2): New.

gcc/testsuite/ChangeLog:

PR target/121853
* gcc.target/aarch64/pr121853_1.c: New test.
* gcc.target/aarch64/pr121853_2.c: New test.

fortran: remove redundant code related to constant pointer in atomic_cas

This part is unreachable after r16-4474-g2c1949bf152f8f.

gcc/fortran/ChangeLog:

* intrinsic.texi: Fix typo.
* trans-intrinsic.cc (conv_intrinsic_atomic_cas): Remove unreachable
code.

Fortran: Use specific PDT constructors from a generic list [PR122452]

2025-10-31  Paul Thomas  <pault@gcc.gnu.org>

gcc/fortran
PR fortran/122452
* primary.cc (gfc_match_rvalue): Give priority to specific
procedures in a generic interface with the same name as a
PDT template. If found, use as the procedure instead of the
constructor generated from the PDT template.

gcc/testsuite/
PR fortran/122452
* gfortran.dg/pdt_65.f03: New test.

tree-optimization/122502 - avoid folding during imm use walk with active ranger

The following works around an unfortunate interaction with ranger
and immediate use walking.  An actual solution needs more thoughts.

PR tree-optimization/122502
* tree-scalar-evolution.cc (final_value_replacement_loop):
Avoid folding from within FOR_EACH_IMM_USE_STMT due to active
ranger.

* gcc.dg/torture/pr122502.c: New testcase.

fold: Remove div_if_zero_remainder  [PR122437]

While looking into something related to fold-const.cc, I noticed
that div_if_zero_remainder was unused. The last use of it was
removed in r5-3374-gf65586dcd19846 (via the removal of try_move_mult_to_index),
over 11 years ago.  So it is time to remove this unused function too.
Plus this function is just a small wrapper around wi::multiple_of_p but
creating trees which can get expensive so folks should use wi::multiple_of_p
directly instead.

Bootstrapped and tested on x86_64-linux-gnu.

PR tree-optimization/122437
gcc/ChangeLog:

* fold-const.cc (div_if_zero_remainder): Remove.

Signed-off-by: Andrew Pinski <andrew.pinski@oss.qualcomm.com>

Fix pr122478.c testcase

The problem here is the dump file and the number of times the match
should happen is swapped.

Pushed as obvious.

gcc/testsuite/ChangeLog:

* gcc.dg/tree-ssa/pr122478.c: Swap `1` and `"optimized"`.

Signed-off-by: Andrew Pinski <andrew.pinski@oss.qualcomm.com>

sccp: Fix ICE during folding after proping const in some cases [PR122497]

In this case we originally had:
```
  # g_4.3_21 = PHI <g_4.3_12(3)>
  _1 = g_4.3_21 + 2;
  _2 = g_2[_1][g_4.3_21];
```

SCCP figures out g_4.3_12/g_4.3_21 is 1.
final_value_replacement_loop would remove the phi defining _12.
do the constant prop of the const 1 but that would ICE as we try
to fold the reference `g_2[_1][g_4.3_21]` but that would try to see
the range of `_1` but since `_1` refers back to _21 there would be an
ICE as the phi is already been removed.

The obvious fix is to move the constant prop until after the new statement
for _21 is created.

This fixes the change done by r14-6010-g2dde9f326ded84 and r14-6114-gde0ab339a79535.

This does not fix gcc.dg/graphite/pr82819.c nor tr2/dynamic_bitset/pr92059.cc though;
I will look into those issues in a few.

Pushed as obvious after bootstrap/test.

PR tree-optimization/122497

gcc/ChangeLog:

* tree-scalar-evolution.cc (final_value_replacement_loop): Call replace_uses_by
only after the replacement statement was created.

gcc/testsuite/ChangeLog:

* gcc.dg/torture/pr122497-1.c: New test.

Signed-off-by: Andrew Pinski <andrew.pinski@oss.qualcomm.com>

LoongArch: Remove the "-mlasx" compilation option from the test cases.

Because LoongArch previously lacked support for 128-bit vectors
(vec_widen_{add/sub/mult}_{hi/lo}), the following test cases
needed to be enabled for 256-bit vectors to pass.
Now that 128-bit support has been added, "-mlasx" has been removed.

gcc/testsuite/ChangeLog:

* gcc.dg/vect/slp-widen-mult-half.c: Remove '-mlasx'.
* gcc.dg/vect/tree-vect.h: Check whether the runtime
environment supports LSX instructions.
* gcc.dg/vect/vect-widen-mult-const-s16.c: Dito.
* gcc.dg/vect/vect-widen-mult-const-u16.c: Dito.
* gcc.dg/vect/vect-widen-mult-half-u8.c: Dito.
* gcc.dg/vect/vect-widen-mult-half.c: Dito.
* gcc.dg/vect/vect-widen-mult-u16.c: Dito.
* gcc.dg/vect/vect-widen-mult-u8-s16-s32.c: Dito.
* gcc.dg/vect/vect-widen-mult-u8-u32.c: Dito.
* gcc.dg/vect/vect-widen-mult-u8.c: Dito.

LoongArch: Support vec_widen_[s]{add/sub/mult}_{hi/lo}_{v4si/v8si}.

gcc/ChangeLog:

* config/loongarch/lasx.md: Support.
* config/loongarch/loongarch.cc
(loongarch_expand_vec_widen_hilo): Remove unused code.
* config/loongarch/lsx.md: Support.

LoongArch: Implement the vector dot product operation with quadruple width.

The logic for the vector dot product operation, where the destination
elements are quadruple the width of the source elements, is as
follows (Take sdot_prodv4siv16qi as an example.):

v16i8 src1, src2;
v4i32 src3, dest;

dest[0] = src1[0] * src2[0] + src1[1] * src2[1]
  + src1[2] * src2[2] + src1[3] * src2[3]
  + src3[0]
dest[1] = src1[4] * src2[4] + src1[5] * src2[5]
  + src1[6] * src2[6] + src1[7] * src2[7]
  + src3[1]
dest[2] = src1[8] * src2[8] + src1[9] * src2[9]
  + src1[10] * src2[10] + src1[11] * src2[11]
  + src3[2]
dest[3] = src1[12] * src2[12] + src1[13] * src2[13]
  + src1[14] * src2[14] + src1[15] * src2[15]
  + src3[3]

gcc/ChangeLog:

* config/loongarch/lasx.md (ILASX_HB): Move to ...
* config/loongarch/lsx.md (ILSX_HB): Move to ...
* config/loongarch/simd.md (ILSX_HB): ... here.
(ILASX_HB): ... here.
(IVEC_HB): New iterator.
(WVEC_QUARTER): New attr.
(wvec_quarter): Likewise.
(simdfmt_qw): Likewise.
(<su>dot_prod<wvec_quarter><mode>): New template.

LoongArch: Implement vec_widen_<su>{add/sub/mult}_{hi/lo}_{v16qi/v8hi}.

This patch will cause gcc.dg/vect/bb-slp-35.c test to fail.

gcc/ChangeLog:

* config/loongarch/lasx.md (hi_lo): Move to ...
* config/loongarch/simd.md (hi_lo): ... here.
* config/loongarch/loongarch.cc
(loongarch_expand_vec_widen_hilo): Add 128-bit data processing.
* config/loongarch/lsx.md
(vec_widen_<su><optab>_<hi_lo>_<mode>): New define_expand.
(vec_widen_<su>mult_<hi_lo>_<mode>): Likewise.

LoongArch: Simplify vec_widen_<su>{add/sub/mult}_{hi/lo}_m describe.

gcc/ChangeLog:

* config/loongarch/lasx.md (vec_widen_<su>add_hi_<mode>):
Move.
(vec_widen_<su>add_lo_<mode>): Move.
(vec_widen_<su>sub_hi_<mode>): Move.
(vec_widen_<su>sub_lo_<mode>): Move.
(vec_widen_<su>mult_hi_<mode>): Move.
(vec_widen_<su>mult_lo_<mode>): Move.
(hi_lo): New define_int_attr.
(vec_widen_<su><optab>_<hi_lo>_<mode>): New define_expand.
(vec_widen_<su>mult_<hi_lo>_<mode>): Likewise.
* config/loongarch/loongarch-protos.h
(loongarch_expand_vec_widen_hilo): Modify the function
parameter list.
* config/loongarch/loongarch.cc
(loongarch_expand_vec_widen_hilo): Optimized.

Daily bump.

c++: Fix up and tweak std-name-hint.gperf

When converting reflection metafunction evaluation into gperf, I had
a look at the other *.gperf files in gcc/cp/.
cfns.gperf doesn't have any hash conflicts, but std-name-hint.gperf
has 7, which I believe means that for 7 of the strings it falls back to some
slow loop over not sure how many entries.

Apparently, 2 out of the 7 conflicts are just because two lines are
identically duplicated in the file, the patch included below fixes that.

2025-10-30  Jakub Jelinek  <jakub@redhat.com>

* std-name-hint.gperf: Remove duplicate entries for ispanstream
and istringstream.
* std-name-hint.h: Regenerate.

gccrs: Improve ErrorPropagationExpr handling

gcc/rust/ChangeLog:

* ast/rust-ast-pointer-visitor.cc (PointerVisitor::visit):
Handle the expression contained in an ErrorPropagationExpr.
* ast/rust-ast-visitor.cc (DefaultASTVisitor::visit): Likewise.

Signed-off-by: Owen Avery <powerboat9.gamer@gmail.com>

gccrs: Attempt to resolve pattern items in tuple pattern with a rest pattern

gcc/rust/ChangeLog:

* typecheck/rust-hir-type-check-pattern.cc (visit(TuplePattern)): Update HAS_REST
case to continue to attempt to resolve pattern items after emitting size error.

Signed-off-by: Yap Zhi Heng <yapzhhg@gmail.com>

gccrs: Fix ICE on enum in tuple struct pattern

When we typecheck a tuple struct pattern and the type of its path is
an enum, it may refer to the enum itself and not a variant. Emit an
E0532 error on such cases.

Fixes Rust-GCC/gccrs#3917
Fixes Rust-GCC/gccrs#3918
Fixes Rust-GCC/gccrs#3926

gcc/rust/ChangeLog:

* typecheck/rust-hir-type-check-pattern.cc (TypeCheckPattern::visit): Emit
an error when the path refers to an enum itself rather than its variant.

gcc/testsuite/ChangeLog:

* rust/compile/match-tuplestructpattern-non-variant.rs: New test.

Signed-off-by: Ryo Yoshida <low.ryoshida@gmail.com>

gccrs: Fix ICE for repr attribute malformation

gcc/rust/ChangeLog:

* typecheck/rust-hir-type-check-base.cc (parse_repr_options): Changed TOKEN_TREE
assert into error, update malformed repr attribute error message to be inline
with other attribute error messages.

Signed-off-by: Yap Zhi Heng <yapzhhg@gmail.com>

gccrs: fix inner attr derive doesn't show error

gcc/rust/ChangeLog:

* ast/rust-ast.cc (Attribute::is_derive):
Change string derive to its definition.
* util/rust-attribute-values.h:
Add attribute definition derive.
* util/rust-attributes.cc (AttributeChecker::visit):
Add method check_inner_attributes.
(AttributeChecker::check_inner_attributes):
Check if there is a bad inner attribute.
* util/rust-attributes.h:
Add method check_inner_attributes in .h.

gcc/testsuite/ChangeLog:

* rust/compile/issue-4212.rs: New test.

Signed-off-by: Lucas Ly Ba <lucas.ly-ba@outlook.fr>

gccrs: fix ICE in TyVar constructor

gcc/rust/ChangeLog:

* typecheck/rust-tyty-util.cc (TyVar::TyVar): Add null check to avoid ICE.
(TyVar::get_tyty): Return nullptr when lookup fails.
(TyVar::clone): Handle null base type safely.
(TyVar::monomorphized_clone): Add fallback for error types.

gcc/testsuite/ChangeLog:

* rust/compile/issue-3556.rs: New test.

Signed-off-by: lishin <lishin1008@gmail.com>

gccrs: Emit error when tuple-indexing on non-tuples

Fixes Rust-GCC#3927

gcc/rust/ChangeLog:

* typecheck/rust-hir-type-check-expr.cc (TypeCheckExpr::visit): check
and properly emit an error when the tuple index is on a non-tuple-struct.

gcc/testsuite/ChangeLog:

* rust/compile/tuple_index_on_non_tuple.rs: New test.

Signed-off-by: Ryo Yoshida <low.ryoshida@gmail.com>

gccrs: Refactor TyTy::ConstType into separate types

This patch refactors the const generic type system to follow the same pattern
as regular type parameters. The monolithic ConstType is split into four
distinct types:

  ConstParamType (generic parameter placeholder)
  ConstValueType (resolved constant value)
  ConstInferType (inference variable)
  ConstErrorType (error sentinel)

gcc/rust/ChangeLog:

* backend/rust-compile-expr.cc (CompileExpr::array_copied_expr): refactor to new classes
* backend/rust-compile-pattern.cc (CompilePatternCheckExpr::visit): likewise
(CompilePatternBindings::visit): likewise
* backend/rust-compile-type.cc (TyTyResolveCompile::visit): likewise
* backend/rust-compile-type.h: likewise
* typecheck/rust-hir-type-check-base.cc (TypeCheckBase::resolve_literal): likewise
* typecheck/rust-hir-type-check-expr.cc (TypeCheckExpr::visit): likewise
* typecheck/rust-hir-type-check-pattern.cc (TypeCheckPattern::visit): likewise
* typecheck/rust-hir-type-check-type.cc (TypeCheckType::visit): likewise
* typecheck/rust-substitution-mapper.cc (SubstMapperInternal::visit): likewise
* typecheck/rust-substitution-mapper.h: likewise
* typecheck/rust-type-util.cc (unify_site_and): remove bad delete (alpine32)
* typecheck/rust-tyty-call.h: refactor to new classes
* typecheck/rust-tyty-subst.cc (SubstitutionParamMapping::clone): likewise
(SubstitutionRef::infer_substitions): likewise
* typecheck/rust-tyty-util.cc (TyVar::get_implicit_const_infer_var): likewise
* typecheck/rust-tyty-util.h: likewise
* typecheck/rust-tyty-variance-analysis-private.h: likewise
* typecheck/rust-tyty-visitor.h: likewise
* typecheck/rust-tyty.cc (BaseType::destructure): likewise
(BaseType::monomorphized_clone): likewise
(BaseType::is_concrete): likewise
(VariantDef::clone): likewise
(VariantDef::monomorphized_clone): likewise
(ArrayType::as_string): likewise
(ArrayType::get_capacity): likewise
(ArrayType::handle_substitions): likewise
(generate_tree_str): likewise
(ConstType::ConstType): likewise
(ConstParamType::ConstParamType): likewise
(ConstType::accept_vis): likewise
(ConstParamType::const_kind): likewise
(ConstParamType::get_symbol): likewise
(ConstParamType::can_resolve): likewise
(ConstParamType::resolve): likewise
(ConstParamType::accept_vis): likewise
(ConstType::set_value): likewise
(ConstType::as_string): likewise
(ConstParamType::as_string): likewise
(ConstType::clone): likewise
(ConstParamType::clone): likewise
(ConstType::get_symbol): likewise
(ConstParamType::get_name): likewise
(ConstType::can_resolve): likewise
(ConstParamType::is_equal): likewise
(ConstType::resolve): likewise
(ConstValueType::ConstValueType): likewise
(ConstValueType::const_kind): likewise
(ConstValueType::accept_vis): likewise
(ConstValueType::as_string): likewise
(ConstValueType::clone): likewise
(ConstValueType::get_name): likewise
(ConstValueType::is_equal): likewise
(ConstValueType::get_value): likewise
(ConstInferType::ConstInferType): likewise
(ConstInferType::const_kind): likewise
(ConstInferType::accept_vis): likewise
(ConstType::get_name): likewise
(ConstInferType::as_string): likewise
(ConstInferType::clone): likewise
(ConstInferType::get_name): likewise
(ConstType::is_equal): likewise
(ConstInferType::is_equal): likewise
(ConstErrorType::ConstErrorType): likewise
(ConstErrorType::const_kind): likewise
(ConstType::handle_substitions): likewise
(ConstErrorType::accept_vis): likewise
(ConstErrorType::as_string): likewise
(ConstErrorType::clone): likewise
(ConstErrorType::get_name): likewise
(ConstErrorType::is_equal): likewise
* typecheck/rust-tyty.h (class BaseConstType): likewise
(class ConstType): likewise
(class ConstParamType): likewise
(class ConstValueType): likewise
(class ConstInferType): likewise
(class ConstErrorType): likewise
* typecheck/rust-unify.cc (UnifyRules::commit): likewise
(UnifyRules::go): likewise
(UnifyRules::expect_array): likewise
(UnifyRules::expect_const): likewise
* typecheck/rust-unify.h: likewise

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: test: add a regression test for issue Rust-GCC/gccrs#4145

gcc/testsuite/ChangeLog:

* rust/compile/issue-4145.rs: New test.

Signed-off-by: Elle Rhumsaa <elle@weathered-steel.dev>

gccrs: fix: add early return for empty module file

Converts an assert into an early return during AST parsing.

Resolves: Rust-GCC/gccrs#4145

gcc/rust/ChangeLog:

* ast/rust-ast.cc (Module::process_file_path): empty module early return

Signed-off-by: Elle Rhumsaa <elle@weathered-steel.dev>

gccrs: Add doc attribute checking on traits

Doc attribute checking was not performed on traits.

gcc/rust/ChangeLog:

* util/rust-attributes.cc (check_doc_attribute): Change message.
(AttributeChecker::visit): Add doc attribute checking on traits.

gcc/testsuite/ChangeLog:

* rust/compile/attr_malformed_doc.rs: New test.

Signed-off-by: Pierre-Emmanuel Patry <pierre-emmanuel.patry@embecosm.com>

gccrs: Avoid malformed attribute conversion

Add location for malformed attribute errors and avoid processing of
malformed attributes.

gcc/rust/ChangeLog:

* ast/rust-ast.cc (AttrInputMetaItemContainer::separate_cfg_attrs):
Avoid malformed attributes.
* util/rust-attributes.cc (AttributeChecker::visit): Change location.

gcc/testsuite/ChangeLog:

* rust/compile/attr_malformed_path.rs: New test.

Signed-off-by: Pierre-Emmanuel Patry <pierre-emmanuel.patry@embecosm.com>

gccrs: Force crash when retrieving meta item location

We still don't know which location should be preferred over the other,
this means that nobody should rely on this function's return value.

gcc/rust/ChangeLog:

* ast/rust-expr.h: Force crash when retrieving locus.

Signed-off-by: Pierre-Emmanuel Patry <pierre-emmanuel.patry@embecosm.com>

gccrs: Add override modifier

gcc/rust/ChangeLog:

* ast/rust-ast.h: Add missing override modifier.
* ast/rust-path.h: Likewise.

Signed-off-by: Pierre-Emmanuel Patry <pierre-emmanuel.patry@embecosm.com>

gccrs: Emit an error on malformed path

Path must be made of a single literal item, otherwise an error should be
thrown.

gcc/rust/ChangeLog:

* util/rust-attributes.cc (AttributeChecker::check_attribute): Recurse
within attr input for additional attribute checking.
(AttributeChecker::visit): Remove empty definition in favor of default
ast visitor definition.
* util/rust-attributes.h: Remove now unused prototypes.

Signed-off-by: Pierre-Emmanuel Patry <pierre-emmanuel.patry@embecosm.com>

gccrs: Use rich_location for TupleStructPattern type check num field error

gcc/rust/ChangeLog:

* typecheck/rust-hir-type-check-pattern.cc (visit(TupleStructPattern)):
Update error for mismatched number of fields to use rich_location.

Signed-off-by: Yap Zhi Heng <yapzhhg@gmail.com>

gccrs: Implement compilation support for HIR::TupleStructItemsHasRest

gcc/rust/ChangeLog:

* backend/rust-compile-pattern.cc: Add compilation support for TupleStructItemsHasRest
in CompilePatternCheckExpr(TupleStructPattern) & CompilePatternBindings(TupleStructPattern)

Signed-off-by: Yap Zhi Heng <yapzhhg@gmail.com>

gccrs: Implement lowering of AST::TupleStructItemsHasRest

gcc/rust/ChangeLog:

* hir/rust-ast-lower-pattern.cc (ASTLoweringPattern::visit(TupleStructPattern)):
Implement lowering of AST::TupleStructItemsHasRest to HIR.
* typecheck/rust-hir-type-check-pattern.cc (TypeCheckPattern::visit(TupleStructPattern)):
Add the respective type checking for AST::TupleStructItemsHasRest
* checks/errors/rust-hir-pattern-analysis.cc (lower_tuple_pattern):
Add respective pattern for HAS_REST case.

Signed-off-by: Yap Zhi Heng <yapzhhg@gmail.com>

gccrs: tests: added extra parens compile test

Adds a test to verify extra parentheses are discarded, and the source is
compiled as expected.

Resolves: Rust-GCC/gccrs#2886

gcc/testsuite/ChangeLog:

* rust/compile/issue-4148.rs: New test.

Signed-off-by: Elle Rhumsaa <elle@weathered-steel.dev>

gccrs: Test libcore

This adds compilation of libcore up to the AST validation phase to the
testsuite.

gcc/testsuite/ChangeLog:

* rust/core/core.exp: New test.

Signed-off-by: Owen Avery <powerboat9.gamer@gmail.com>

gccrs: Fix StructPattern type checking rejecting tuple struct scrutinee

Previously, type checking of StructPattern will throw an error if it is used
to match against a tuple struct, even though it is possible to do so in rustc.

gcc/rust/ChangeLog:

* typecheck/rust-hir-type-check-pattern.cc (visit(StructPattern)): Add type check
support for StructPatterns matching against a TUPLE type ADT.
* backend/rust-compile-pattern.cc(CompilePatternBindings::visit(StructPattern)):
Update assert to allow TUPLE type ADTs.
* hir/tree/rust-hir.cc (StructPatternField::as_string()): Improve info dumped.

Signed-off-by: Yap Zhi Heng <yapzhhg@gmail.com>

gccrs: Add HIR lowering support for rest pattern in struct patterns' AST

gcc/rust/ChangeLog:

* ast/rust-pattern.h (StructPatternElements): Rename has_struct_pattern_etc
to has_rest_pattern, and has_etc to has_rest to signify presense of rest
patterns more clearly.
* ast/rust-pattern.cc (StructPatternElements::as_string): Rename variables
accordingly.
* ast/rust-ast-collector.cc: Rename variables accordingly.
* expand/rust-cfg-strip.cc: Rename variables accordingly.
* parse/rust-parse-impl.h: Rename variable accordingly.
* hir/tree/rust-hir-pattern.h (StructPatternElements): Add a boolean to track
presense of rest pattern.
* hir/rust-ast-lower-pattern.cc (visit(StructPattern)): Add support for
lowering rest pattern to HIR.
* typecheck/rust-hir-type-check-pattern.cc (visit(StructPattern)): Remove
size check when rest pattern is present.

Signed-off-by: Yap Zhi Heng <yapzhhg@gmail.com>

gccrs: Handle link_name attribute

gcc/rust/ChangeLog:

* backend/rust-compile-extern.h: Add includes.
(CompileExternItem::visit): Use get_link_name.
(CompileExternItem::get_link_name): New static member function.
* util/rust-attribute-values.h (Attributes::LINK_NAME): New
static constexpr member variable.
* util/rust-attributes.cc (__definitions): New entry for
LINK_NAME.
* util/rust-ggc.cc: Include "rust-ast.h".
(Ident::Ident): Add overload for Rust::Identifier.
* util/rust-ggc.h (class Identifier): Forward declare.
(Ident::Ident): Add overload for Rust::Identifier.

gcc/testsuite/ChangeLog:

* rust/execute/torture/link-name.rs: New test.

Signed-off-by: Owen Avery <powerboat9.gamer@gmail.com>

gccrs: expand: Inherit from PointerVisitor

gcc/rust/ChangeLog:

* expand/rust-expand-visitor.cc: Remove old visitors.
* expand/rust-expand-visitor.h: Likewise, plus inherit from PointerVisitor.

gccrs: Add origin location to expanded tokens

gcc/rust/ChangeLog:

* expand/rust-macro-expand.cc: Forward invocation tree locus to
substitution context.
* expand/rust-macro-substitute-ctx.cc: Use origin location for expanded
tokens.
* expand/rust-macro-substitute-ctx.h (class SubstituteCtx): Save
invocation location.

gcc/testsuite/ChangeLog:

* rust/compile/macros/mbe/macro58.rs: New test.

Signed-off-by: Pierre-Emmanuel Patry <pierre-emmanuel.patry@embecosm.com>

gccrs: Emit errors from the transcriber when they occur

Emitting the errors later means some error could be emitted multiple
times.

gcc/rust/ChangeLog:

* expand/rust-macro-expand.cc (transcribe_expression): Emit error
early.

Signed-off-by: Pierre-Emmanuel Patry <pierre-emmanuel.patry@embecosm.com>

gccrs: Prevent segfault on builtin metavars

Builtin metavars are not in the fragment, the compiler should not emit
an error message.

gcc/rust/ChangeLog:

* expand/rust-macro-substitute-ctx.cc (is_builtin_metavariable): Add
function to check builtin metavars knowledge.
(SubstituteCtx::check_repetition_amount): Do not process missing
fragment.

Signed-off-by: Pierre-Emmanuel Patry <pierre-emmanuel.patry@embecosm.com>

gccrs: Fix ICE where we expect a num enum variant

This changes the assertion into a valid error diagnostic.

Fixes Rust-GCC#3538

gcc/rust/ChangeLog:

* backend/rust-compile-resolve-path.cc: add error diag

gcc/testsuite/ChangeLog:

* rust/compile/issue-3538.rs: New test.

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: Add error diag for self params on plain functions

self params dont have a type unless used within impl blocks. Rustc as far as I
can tell in this senario generics a synthetic param of type Self in this senario
so that it keeps consistent error diagnostic for number of parameters but the
logic for what the parameter typpe should be seems unclear.

Fixes Rust-GCC#3592

gcc/rust/ChangeLog:

* hir/rust-ast-lower-item.cc (ASTLoweringItem::visit): add error diagnostic

gcc/testsuite/ChangeLog:

* rust/compile/issue-3592.rs: New test.

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: Fix ICE on copied array expressions

We need to check for errors on the number of copies expression before
trying to const fold it otherwise it will just fail in the const evaluator.

Fixes Rust-GCC#4165

gcc/rust/ChangeLog:

* typecheck/rust-hir-type-check-expr.cc (TypeCheckExpr::visit): check for error

gcc/testsuite/ChangeLog:

* rust/compile/issue-4165.rs: New test.

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: Remove old hack from resolve operator overload

This was an old chunk of code to try and sort out the Self type on resolving
methods for impl blocks. But this is not required when our unify code works
properly.

gcc/rust/ChangeLog:

* typecheck/rust-hir-type-check-expr.cc: remove old hack

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: fnptr types can hold onto generic params so it needs to handle substs

This patch adds support to recursively walk fnptr types to do generic
substitutions.

Fixes Rust-GCC/gccrs#4090

gcc/rust/ChangeLog:

* typecheck/rust-substitution-mapper.cc (SubstMapperInternal::visit): handle fnptr
* typecheck/rust-tyty.cc (FnPtr::handle_substitions): new
* typecheck/rust-tyty.h: likewise

gcc/testsuite/ChangeLog:

* rust/compile/issue-4090-1.rs: New test.
* rust/compile/issue-4090-2.rs: New test.

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: Improve FFIOpt

Also fixes https://github.com/Rust-GCC/gccrs/issues/4171.

gcc/rust/ChangeLog:

* ast/rust-fmt.h (class FFIOpt): Adjust internal structure to
match a repr(C) rust enum.

libgrust/ChangeLog:

* libformat_parser/src/lib.rs (struct FFIOpt): Likewise and
remove some now-redundant methods.

Signed-off-by: Owen Avery <powerboat9.gamer@gmail.com>

gccrs: Implement TokenCollector::visit(AST::FormatArgs&)

This allows visiting format_args! during token collection.

gcc/rust/ChangeLog:

* ast/rust-ast-collector.cc (TokenCollector::visit): Implement
FormatArgs visitor to reconstruct format_args! macro syntax.
* ast/rust-builtin-ast-nodes.h (FormatArguments): Add get_args(),
size(), and empty() accessor methods.

Signed-off-by: Magnus-Mage <romescha123@gmail.com>

gccrs: Fix ICE for invalid const capacity expression handling

When we have an invalid capacity expression we can't try to then also
const fold it as GCC will assert on invalid conversions.

Fixes Rust-GCC#4168

gcc/rust/ChangeLog:

* typecheck/rust-hir-type-check-type.cc (TypeCheckType::visit): check for invalid capacity

gcc/testsuite/ChangeLog:

* rust/compile/issue-4168.rs: New test.

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: Import stdarch 1.49.0

This commit imports stdarch 1.49.0 into libgrust/rustc-lib/stdarch. This
is necessary for compiling libcore, as libcore attempts to reuse some
files from stdarch.

libgrust/ChangeLog:

* rustc-lib/stdarch/.cirrus.yml: New file.
* rustc-lib/stdarch/.github/workflows/main.yml: New file.
* rustc-lib/stdarch/.gitignore: New file.
* rustc-lib/stdarch/CONTRIBUTING.md: New file.
* rustc-lib/stdarch/Cargo.toml: New file.
* rustc-lib/stdarch/LICENSE-APACHE: New file.
* rustc-lib/stdarch/LICENSE-MIT: New file.
* rustc-lib/stdarch/README.md: New file.
* rustc-lib/stdarch/ci/android-install-ndk.sh: New file.
* rustc-lib/stdarch/ci/android-install-sdk.sh: New file.
* rustc-lib/stdarch/ci/android-sysimage.sh: New file.
* rustc-lib/stdarch/ci/docker/aarch64-linux-android/Dockerfile: New file.
* rustc-lib/stdarch/ci/docker/aarch64-unknown-linux-gnu/Dockerfile: New file.
* rustc-lib/stdarch/ci/docker/arm-linux-androideabi/Dockerfile: New file.
* rustc-lib/stdarch/ci/docker/arm-unknown-linux-gnueabihf/Dockerfile: New file.
* rustc-lib/stdarch/ci/docker/armv7-unknown-linux-gnueabihf/Dockerfile: New file.
* rustc-lib/stdarch/ci/docker/i586-unknown-linux-gnu/Dockerfile: New file.
* rustc-lib/stdarch/ci/docker/i686-unknown-linux-gnu/Dockerfile: New file.
* rustc-lib/stdarch/ci/docker/mips-unknown-linux-gnu/Dockerfile: New file.
* rustc-lib/stdarch/ci/docker/mips64-unknown-linux-gnuabi64/Dockerfile: New file.
* rustc-lib/stdarch/ci/docker/mips64el-unknown-linux-gnuabi64/Dockerfile: New file.
* rustc-lib/stdarch/ci/docker/mipsel-unknown-linux-musl/Dockerfile: New file.
* rustc-lib/stdarch/ci/docker/nvptx64-nvidia-cuda/Dockerfile: New file.
* rustc-lib/stdarch/ci/docker/powerpc-unknown-linux-gnu/Dockerfile: New file.
* rustc-lib/stdarch/ci/docker/powerpc64-unknown-linux-gnu/Dockerfile: New file.
* rustc-lib/stdarch/ci/docker/powerpc64le-unknown-linux-gnu/Dockerfile: New file.
* rustc-lib/stdarch/ci/docker/s390x-unknown-linux-gnu/Dockerfile: New file.
* rustc-lib/stdarch/ci/docker/wasm32-wasi/Dockerfile: New file.
* rustc-lib/stdarch/ci/docker/x86_64-linux-android/Dockerfile: New file.
* rustc-lib/stdarch/ci/docker/x86_64-unknown-linux-gnu-emulated/Dockerfile: New file.
* rustc-lib/stdarch/ci/docker/x86_64-unknown-linux-gnu/Dockerfile: New file.
* rustc-lib/stdarch/ci/dox.sh: New file.
* rustc-lib/stdarch/ci/gba.json: New file.
* rustc-lib/stdarch/ci/run-docker.sh: New file.
* rustc-lib/stdarch/ci/run.sh: New file.
* rustc-lib/stdarch/ci/runtest-android.rs: New file.
* rustc-lib/stdarch/ci/style.sh: New file.
* rustc-lib/stdarch/crates/assert-instr-macro/Cargo.toml: New file.
* rustc-lib/stdarch/crates/assert-instr-macro/build.rs: New file.
* rustc-lib/stdarch/crates/assert-instr-macro/src/lib.rs: New file.
* rustc-lib/stdarch/crates/core_arch/Cargo.toml: New file.
* rustc-lib/stdarch/crates/core_arch/LICENSE-APACHE: New file.
* rustc-lib/stdarch/crates/core_arch/LICENSE-MIT: New file.
* rustc-lib/stdarch/crates/core_arch/README.md: New file.
* rustc-lib/stdarch/crates/core_arch/avx512f.md: New file.
* rustc-lib/stdarch/crates/core_arch/build.rs: New file.
* rustc-lib/stdarch/crates/core_arch/rustfmt.toml: New file.
* rustc-lib/stdarch/crates/core_arch/src/aarch64/crc.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/aarch64/crypto.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/aarch64/mod.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/aarch64/neon/generated.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/aarch64/neon/mod.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/aarch64/prefetch.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/aarch64/test_support.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/aarch64/tme.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/aarch64/v8.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/acle/barrier/common.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/acle/barrier/cp15.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/acle/barrier/mod.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/acle/barrier/not_mclass.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/acle/barrier/v8.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/acle/dsp.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/acle/ex.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/acle/hints.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/acle/mod.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/acle/registers/aarch32.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/acle/registers/mod.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/acle/registers/v6m.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/acle/registers/v7m.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/acle/sat.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/acle/simd32.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/arm/armclang.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/arm/crc.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/arm/mod.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/arm/neon/generated.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/arm/neon/mod.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/arm/neon/table_lookup_tests.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/arm/test_support.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/arm/v6.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/arm/v7.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/core_arch_docs.md: New file.
* rustc-lib/stdarch/crates/core_arch/src/lib.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/macros.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/mips/mod.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/mips/msa.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/mips/msa/macros.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/mod.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/nvptx/mod.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/powerpc/altivec.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/powerpc/mod.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/powerpc/vsx.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/powerpc64/mod.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/simd.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/simd_llvm.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/v64.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/wasm32/atomic.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/wasm32/memory.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/wasm32/mod.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/wasm32/simd128.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/abm.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/adx.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/aes.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/avx.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/avx2.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/avx512f.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/avx512ifma.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/bmi1.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/bmi2.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/bswap.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/bt.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/cpuid.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/eflags.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/f16c.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/fma.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/fxsr.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/macros.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/mod.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/pclmulqdq.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/rdrand.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/rdtsc.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/rtm.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/sha.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/sse.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/sse2.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/sse3.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/sse41.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/sse42.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/sse4a.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/ssse3.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/tbm.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/test.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86/xsave.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86_64/abm.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86_64/adx.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86_64/avx.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86_64/avx2.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86_64/avx512f.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86_64/bmi.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86_64/bmi2.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86_64/bswap.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86_64/bt.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86_64/cmpxchg16b.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86_64/fxsr.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86_64/mod.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86_64/rdrand.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86_64/sse.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86_64/sse2.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86_64/sse41.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86_64/sse42.rs: New file.
* rustc-lib/stdarch/crates/core_arch/src/x86_64/xsave.rs: New file.
* rustc-lib/stdarch/crates/core_arch/tests/cpu-detection.rs: New file.
* rustc-lib/stdarch/crates/simd-test-macro/Cargo.toml: New file.
* rustc-lib/stdarch/crates/simd-test-macro/src/lib.rs: New file.
* rustc-lib/stdarch/crates/std_detect/Cargo.toml: New file.
* rustc-lib/stdarch/crates/std_detect/LICENSE-APACHE: New file.
* rustc-lib/stdarch/crates/std_detect/LICENSE-MIT: New file.
* rustc-lib/stdarch/crates/std_detect/README.md: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/arch/aarch64.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/arch/arm.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/arch/mips.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/arch/mips64.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/arch/powerpc.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/arch/powerpc64.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/arch/x86.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/bit.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/cache.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/error_macros.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/macros.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/mod.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/os/aarch64.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/os/freebsd/aarch64.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/os/freebsd/arm.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/os/freebsd/auxvec.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/os/freebsd/mod.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/os/freebsd/powerpc.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/os/linux/aarch64.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/os/linux/arm.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/os/linux/auxvec.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/os/linux/cpuinfo.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/os/linux/mips.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/os/linux/mod.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/os/linux/powerpc.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/os/other.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/os/windows/aarch64.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/os/x86.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/test_data/linux-rpi3.auxv: New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/test_data/linux-x64-i7-6850k.auxv:
New file.
* rustc-lib/stdarch/crates/std_detect/src/detect/test_data/macos-virtualbox-linux-x86-4850HQ.auxv:
New file.
* rustc-lib/stdarch/crates/std_detect/src/lib.rs: New file.
* rustc-lib/stdarch/crates/std_detect/src/mod.rs: New file.
* rustc-lib/stdarch/crates/std_detect/tests/cpu-detection.rs: New file.
* rustc-lib/stdarch/crates/std_detect/tests/macro_trailing_commas.rs: New file.
* rustc-lib/stdarch/crates/std_detect/tests/x86-specific.rs: New file.
* rustc-lib/stdarch/crates/stdarch-gen/Cargo.toml: New file.
* rustc-lib/stdarch/crates/stdarch-gen/README.md: New file.
* rustc-lib/stdarch/crates/stdarch-gen/neon.spec: New file.
* rustc-lib/stdarch/crates/stdarch-gen/src/main.rs: New file.
* rustc-lib/stdarch/crates/stdarch-test/Cargo.toml: New file.
* rustc-lib/stdarch/crates/stdarch-test/src/disassembly.rs: New file.
* rustc-lib/stdarch/crates/stdarch-test/src/lib.rs: New file.
* rustc-lib/stdarch/crates/stdarch-test/src/wasm.rs: New file.
* rustc-lib/stdarch/crates/stdarch-verify/.gitattributes: New file.
* rustc-lib/stdarch/crates/stdarch-verify/Cargo.toml: New file.
* rustc-lib/stdarch/crates/stdarch-verify/arm-intrinsics.html: New file.
* rustc-lib/stdarch/crates/stdarch-verify/build.rs: New file.
* rustc-lib/stdarch/crates/stdarch-verify/mips-msa.h: New file.
* rustc-lib/stdarch/crates/stdarch-verify/src/lib.rs: New file.
* rustc-lib/stdarch/crates/stdarch-verify/tests/arm.rs: New file.
* rustc-lib/stdarch/crates/stdarch-verify/tests/mips.rs: New file.
* rustc-lib/stdarch/crates/stdarch-verify/tests/x86-intel.rs: New file.
* rustc-lib/stdarch/crates/stdarch-verify/x86-intel.xml: New file.
* rustc-lib/stdarch/examples/Cargo.toml: New file.
* rustc-lib/stdarch/examples/hex.rs: New file.
* rustc-lib/stdarch/examples/wasm.rs: New file.
* rustc-lib/stdarch/triagebot.toml: New file.
* rustc-lib/stdarch/vendor.yml: New file.

Signed-off-by: Owen Avery <powerboat9.gamer@gmail.com>

gccrs: Fix ICE handling division by zero in const eval

This pulls over more constexpr support from C++ front-end.

Fixes Rust-GCC#4146

gcc/rust/ChangeLog:

* backend/rust-constexpr.cc (eval_constant_expression): port over missing bits
(eval_store_expression): likewise
(eval_call_expression): likewise
(eval_binary_expression): likewise
(eval_bit_field_ref): likewise
(eval_check_shift_p): likewise
(fold_pointer_plus_expression): likewise
(maybe_fold_addr_pointer_plus): likewise
(fold_expr): likewise
(union_active_member): likewise
(fold_indirect_ref_1): likewise
(rs_fold_indirect_ref): likewise
(rs_eval_indirect_ref): likewise
(eval_logical_expression): likewise
(eval_vector_conditional_expression): likewise
(eval_bare_aggregate): likewise
(cxx_eval_trinary_expression): likewise
(eval_and_check_array_index): likewise
(eval_array_reference): likewise
(eval_component_reference): likewise
(rs_bind_parameters_in_call): likewise
(eval_builtin_function_call): likewise
(constexpr_fn_retval): likewise
(verify_constant): likewise
(get_array_or_vector_nelts): likewise
(eval_conditional_expression): likewise
(eval_switch_expr): likewise
(eval_unary_expression): likewise
(cxx_eval_outermost_constant_expr): likewise
(potential_constant_expression_1): likewise

gcc/testsuite/ChangeLog:

* rust/compile/issue-2394.rs: Update test case
* rust/compile/issue-4146.rs: New test.

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: Fix ICE with invalid const expression

This patch handles the overflowed var expression in the const block, so that
we error properly in the const expr code. It was missing some stuff from the
c++ implementation in how this should be handled properly.

Fixes Rust-GCC#4139

gcc/rust/ChangeLog:

* backend/rust-compile-expr.cc (CompileExpr::compile_integer_literal): cleanup
* backend/rust-constexpr.cc (struct constexpr_global_ctx): port over c++ helpers
(decl_really_constant_value): likewise
(eval_constant_expression): likewise
(non_const_var_error): likewise

gcc/testsuite/ChangeLog:

* rust/compile/issue-4139.rs: New test.

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: Protect from errors in alternate pattern parsing

Fixes https://github.com/Rust-GCC/gccrs/issues/4155.

gcc/rust/ChangeLog:

* parse/rust-parse-impl.h (Parser::parse_pattern): Ignore
inner patterns which fail to parse.

gcc/testsuite/ChangeLog:

* rust/compile/issue-4155.rs: New test.

Signed-off-by: Owen Avery <powerboat9.gamer@gmail.com>

gccrs: Remove AST::TraitItemConst

gcc/rust/ChangeLog:

* ast/rust-ast-collector.cc (TokenCollector::visit): Remove
overload for TraitItemConst.
* ast/rust-ast-collector.h (TokenCollector::visit): Likewise.
* ast/rust-ast-pointer-visitor.cc (PointerVisitor::visit):
Likewise.
* ast/rust-ast-pointer-visitor.h (PointerVisitor::visit):
Likewise.
* ast/rust-ast-visitor.cc (DefaultASTVisitor::visit): Likewise.
* ast/rust-ast-visitor.h (ASTVisitor::visit): Likewise.
(DefaultASTVisitor::visit): Likewise.
* expand/rust-cfg-strip.cc (CfgStrip::visit): Likewise.
* expand/rust-cfg-strip.h (CfgStrip::visit): Likewise.
* expand/rust-derive.h (DeriveVisitor::visit): Likewise.
* expand/rust-expand-visitor.cc (ExpandVisitor::visit):
Likewise.
* expand/rust-expand-visitor.h (ExpandVisitor::visit): Likewise.
* hir/rust-ast-lower-base.cc (ASTLoweringBase::visit): Likewise.
* hir/rust-ast-lower-base.h (ASTLoweringBase::visit): Likewise.
* util/rust-attributes.cc (AttributeChecker::visit): Likewise.
* util/rust-attributes.h (AttributeChecker::visit): Likewise.
* ast/rust-ast.cc (ConstantItem::as_string): Handle missing
expression.
(TraitItemConst::as_string): Remove function definition.
(TraitItemConst::accept_vis): Remove function definition.
* ast/rust-item.h (ConstantItem::mark_for_strip): Adjust
comment.
(ConstantItem::has_expr): Make const.
(class TraitItemConst): Remove definition.
* ast/rust-ast-full-decls.h (class TraitItemConst): Remove
declaration.
* hir/rust-ast-lower-implitem.cc (ASTLowerTraitItem::visit):
Handle ConstantItem instead of TraitItemConst.
* hir/rust-ast-lower-implitem.h (ASTLowerTraitItem::visit):
Likewise.
* parse/rust-parse-impl.h (Parser::parse_trait_const): Return
pointer to ConstantItem instead of TraitItemConst.
* parse/rust-parse.h (Parser::parse_trait_const): Likewise.

Signed-off-by: Owen Avery <powerboat9.gamer@gmail.com>

gccrs: Create subdirectory checks/errors/feature

gcc/rust/ChangeLog:

* Make-lang.in: Handle source files in checks/errors/feature.
* checks/errors/rust-feature-gate.cc: Move to...
* checks/errors/feature/rust-feature-gate.cc: ...here.
* checks/errors/rust-feature-gate.h: Move to...
* checks/errors/feature/rust-feature-gate.h: ...here.
* checks/errors/rust-feature.cc: Move to...
* checks/errors/feature/rust-feature.cc: ...here.
* checks/errors/rust-feature.h: Move to...
* checks/errors/feature/rust-feature.h: ...here.

Signed-off-by: Owen Avery <powerboat9.gamer@gmail.com>

gccrs: nr1.0: Remove EarlyNameResolver

gcc/rust/ChangeLog:

* Make-lang.in (GRS_OBJS): Remove rust-early-name-resolver.o.
* expand/rust-macro-builtins-helpers.h: Remove inclusion of
rust-early-name-resolver.h.
* expand/rust-macro-builtins.cc: Likewise.
* expand/rust-macro-expand.cc: Likewise.
* expand/rust-macro-expand.h: Likewise.
* resolve/rust-early-name-resolver.cc: Removed.
* resolve/rust-early-name-resolver.h: Removed.

Signed-off-by: Owen Avery <powerboat9.gamer@gmail.com>

gccrs: Import libcore 1.49.0

This commit imports libcore 1.49.0 into a new directory,
"libgrust/rustc-lib/core". LICENSE-* files are taken from the rustc
1.49.0 repository root.

libgrust/ChangeLog:

* rustc-lib/LICENSE-APACHE: New file.
* rustc-lib/LICENSE-MIT: New file.
* rustc-lib/version-info: New file.

Signed-off-by: Owen Avery <powerboat9.gamer@gmail.com>

gccrs: Add libcore to the list of ignored prefixes

contrib/ChangeLog:

* gcc-changelog/git_commit.py (ignored_prefixes): Add
'libgrust/rustc-lib/core/'.

Signed-off-by: Owen Avery <powerboat9.gamer@gmail.com>

gccrs: track abi and unsafety on fnptrs for proper type checking

We need to distinguish between abi's and unsafety on fnptrs. There is
a commented out check because there is a regression in:

  rust/compile/try-catch-unwind-{new/old}.rs

But i think this is because the test case should be taking an FnOnce
from reading std::panic in rust 1.49. Where as we are passing an fnptr
which is probably because we didnt support fnonce at all then.

Addresses Rust-GCC/gccrs#4090

gcc/rust/ChangeLog:

* hir/tree/rust-hir-item.h: add unsafe helper
* typecheck/rust-hir-type-check-type.cc (TypeCheckType::visit): pass in abi and unsafe
* typecheck/rust-tyty.cc (BaseType::monomorphized_clone): likewise
(FnPtr::as_string): emit more info
(FnPtr::clone): update ctor call
* typecheck/rust-tyty.h: new ctor params
* typecheck/rust-unify.cc (UnifyRules::expect_fnptr): check abi and unsafe

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: get rid of more clones in unify path

gcc/rust/ChangeLog:

* typecheck/rust-unify.cc (UnifyRules::expect_reference): remove clone
(UnifyRules::expect_pointer): likewise
(UnifyRules::expect_slice): likewise
(UnifyRules::expect_tuple): likewise
(UnifyRules::expect_int): likewise
(UnifyRules::expect_uint): likewise
(UnifyRules::expect_float): likewise

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: Add test to confirm Rust-GCC/gccrs#3922 fix

gcc/testsuite/ChangeLog:

* rust/compile/issue-3922.rs: New test.

Signed-off-by: Pierre-Emmanuel Patry <pierre-emmanuel.patry@embecosm.com>

gccrs: Add test to confirm Rust-GCC/gccrs#3924 fix

gcc/testsuite/ChangeLog:

* rust/compile/issue-3924.rs: New test.

Signed-off-by: Pierre-Emmanuel Patry <pierre-emmanuel.patry@embecosm.com>

gccrs: nr1.0: Remove chunks of resolution phase

This removes large chunks of the 1.0 resolution phase, not the nr1.0
resolver itself quite yet.

gcc/rust/ChangeLog:

* Make-lang.in (GRS_OBJS): Remove entries.
* rust-lang.cc: Remove inclusion of "rust-ast-resolve-item.h".
(run_rust_tests): Remove call to rust_simple_path_resolve_test.
* resolve/rust-ast-resolve-base.cc: Removed.
* resolve/rust-ast-resolve-base.h: Removed.
* resolve/rust-ast-resolve-expr.cc: Removed.
* resolve/rust-ast-resolve-expr.h: Removed.
* resolve/rust-ast-resolve-implitem.h: Removed.
* resolve/rust-ast-resolve-item.cc: Removed.
* resolve/rust-ast-resolve-item.h: Removed.
* resolve/rust-ast-resolve-path.cc: Removed.
* resolve/rust-ast-resolve-path.h: Removed.
* resolve/rust-ast-resolve-pattern.cc: Removed.
* resolve/rust-ast-resolve-pattern.h: Removed.
* resolve/rust-ast-resolve-stmt.cc: Removed.
* resolve/rust-ast-resolve-stmt.h: Removed.
* resolve/rust-ast-resolve-struct-expr-field.cc: Removed.
* resolve/rust-ast-resolve-struct-expr-field.h: Removed.
* resolve/rust-ast-resolve-toplevel.h: Removed.
* resolve/rust-ast-resolve-type.cc: Removed.
* resolve/rust-ast-resolve-type.h: Removed.
* resolve/rust-ast-resolve.cc: Removed.

Signed-off-by: Owen Avery <powerboat9.gamer@gmail.com>

gccrs: Unify path-to-type handling in Late resolver

gcc/rust/ChangeLog:

* resolve/rust-late-name-resolver-2.0.cc (Late::visit): Use
resolve_type_path_like in overloads for TypePath,
StructExprStruct, StructExprStructBase, and
StructExprStructFields.
(resolve_type_path_like): New static function based off
Late::visit overload for TypePath.
* util/rust-unwrap-segment.h (unwrap_segment_error_string): New
inline static function definitions.

Signed-off-by: Owen Avery <powerboat9.gamer@gmail.com>

gccrs: Fix crash on break outside of loop context

We need to add a guard to catch the case when there is no loop context
for break outside of loop.

Fixes Rust-GCC#3969

gcc/rust/ChangeLog:

* backend/rust-compile-expr.cc (CompileExpr::visit): add guard

gcc/testsuite/ChangeLog:

* rust/compile/issue-3969.rs: New test.

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: Fix segv when handling invalid array capacities

We need to catch the error node for the array capacity and return early.
Otherwise we try to const evaluate something thats just silly. Also
when compiling array expressions we can simply reuse the array capacity
expression we already have cons folded.

Fixes Rust-GCC#3965

gcc/rust/ChangeLog:

* backend/rust-compile-context.h: add assertions for context peeks
* backend/rust-compile-expr.cc (CompileExpr::visit): check for valid loop context
(CompileExpr::array_copied_expr): just reuse array tyty capacity value
* typecheck/rust-hir-type-check-expr.cc (TypeCheckExpr::visit): catch error

gcc/testsuite/ChangeLog:

* rust/compile/issue-3965-1.rs: New test.
* rust/compile/issue-3965-2.rs: New test.

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: Implement missing read-only checker case for TuplePatternItemsHasRest

gcc/rust/ChangeLog:

* checks/errors/rust-readonly-check.cc (collect_assignment_tuple): Implement
read-only checker for tuple patterns with rest pattern.

Signed-off-by: Yap Zhi Heng <yapzhhg@gmail.com>

gccrs: Implement missing var decl case for TuplePatternItemsHasRest

GIMPLE dump from compiling issue-3930.rs:

...
  const i32 b;
  const i32 a;

  D.114.__0 = 2;
  D.114.__1 = 3;
  RUSTTMP.1 = D.114;
  a = RUSTTMP.1.__0;
  b = RUSTTMP.1.__1;
...

gcc/rust/ChangeLog:

* backend/rust-compile-var-decl.h (CompileVarDecl::visit(TuplePattern)): Implement
variable declaration bindings for tuple patterns with rest pattern (i.e.
TuplePatternItemsHasRest).

Signed-off-by: Yap Zhi Heng <yapzhhg@gmail.com>

gccrs: Recognize rustc_allow_const_fn_unstable

This doesn't actually handle the attribute, although it does allow us to
recognize it as valid.

gcc/rust/ChangeLog:

* util/rust-attribute-values.h
(Attributes::RUSTC_ALLOW_CONST_FN_UNSTABLE): New static
constexpr member variable.
* util/rust-attributes.cc (__definitions): Add entry for
RUSTC_ALLOW_CONST_FN_UNSTABLE.

Signed-off-by: Owen Avery <powerboat9.gamer@gmail.com>

gccrs: Add implicit infer support for unify on const types

gcc/rust/ChangeLog:

* typecheck/rust-unify.cc (UnifyRules::commit): commit hook update
(UnifyRules::go): insert implicit infer const types

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: allow unifications against non const types

When type resolving a function which returns a const generic for example
this means its unifying the ConstType vs the the specified type so this
mean unwrapping the type of the const.

gcc/rust/ChangeLog:

* typecheck/rust-unify.cc (UnifyRules::go): unwrap the const type

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: Fix bug with unconstrained const generic

gcc/rust/ChangeLog:

* typecheck/rust-hir-type-check-base.cc (walk_types_to_constrain): track the ref as well

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: track generic const generics properly

gcc/rust/ChangeLog:

* typecheck/rust-tyty-subst.cc: const generic arguments dont have a value yet

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: Fix formatting of tyty code

gcc/rust/ChangeLog:

* typecheck/rust-tyty.cc (VariantDef::clone): fix formatting
(VariantDef::monomorphized_clone): likewise
* typecheck/rust-tyty.h: likewise

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: Remove fully the can_eq TyTy::BaseType::can_eq interface

This was an initial helper from back in the day which was not maintained
but was a nice const simple way to check if types are compatable. But
reusing our unify code is much much more acurate and single source of
truth on the type system.

gcc/rust/ChangeLog:

* typecheck/rust-tyty.cc (InferType::can_eq): remove
(ErrorType::can_eq): likewise
(ADTType::can_eq): likewise
(TupleType::can_eq): likewise
(FnType::can_eq): likewise
(FnPtr::can_eq): likewise
(ClosureType::can_eq): likewise
(ArrayType::can_eq): likewise
(SliceType::can_eq): likewise
(BoolType::can_eq): likewise
(IntType::can_eq): likewise
(UintType::can_eq): likewise
(FloatType::can_eq): likewise
(USizeType::can_eq): likewise
(ISizeType::can_eq): likewise
(CharType::can_eq): likewise
(ReferenceType::can_eq): likewise
(PointerType::can_eq): likewise
(ParamType::can_eq): likewise
(ConstType::can_eq): likewise
(OpaqueType::can_eq): likewise
(StrType::can_eq): likewise
(NeverType::can_eq): likewise
(PlaceholderType::can_eq): likewise
(ProjectionType::can_eq): likewise
(DynamicObjectType::can_eq): likewise
* typecheck/rust-tyty.h: remove can_eq
* typecheck/rust-tyty-cmp.h: Removed.

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: get rid of last can_eq usage finally

gcc/rust/ChangeLog:

* typecheck/rust-tyty.cc (ParamType::is_equal): uses types_compatable

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: reuse an error mark node instead of new ones each failure

During unification we denote failures using a new error mark node. This
keeps a static one and reuses it instead.

gcc/rust/ChangeLog:

* typecheck/rust-unify.cc (unify_error_type_node): new static node
(UnifyRules::go): likewise
(UnifyRules::expect_inference_variable): likewise
(UnifyRules::expect_adt): likewise
(UnifyRules::expect_str): likewise
(UnifyRules::expect_reference): likewise
(UnifyRules::expect_pointer): likewise
(UnifyRules::expect_param): likewise
(UnifyRules::expect_array): likewise
(UnifyRules::expect_slice): likewise
(UnifyRules::expect_fndef): likewise
(UnifyRules::expect_fnptr): likewise
(UnifyRules::expect_tuple): likewise
(UnifyRules::expect_bool): likewise
(UnifyRules::expect_char): likewise
(UnifyRules::expect_int): likewise
(UnifyRules::expect_uint): likewise
(UnifyRules::expect_float): likewise
(UnifyRules::expect_isize): likewise
(UnifyRules::expect_usize): likewise
(UnifyRules::expect_never): likewise
(UnifyRules::expect_placeholder): likewise
(UnifyRules::expect_projection): likewise
(UnifyRules::expect_dyn): likewise
(UnifyRules::expect_closure): likewise
(UnifyRules::expect_opaque): likewise
(UnifyRules::expect_const): likewise

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: simplify cloning of types during unify

We should only clone when we commit and it was successful. With no
cloning there is a regression in min specialization 2. Probably because
it needs to know the unify site hirid and ensure we have it set there.

gcc/rust/ChangeLog:

* typecheck/rust-unify.cc (UnifyRules::Resolve): check for success
(UnifyRules::expect_inference_variable): dont clone
(UnifyRules::expect_adt): likewise
(UnifyRules::expect_str): likewise
(UnifyRules::expect_reference): likewise
(UnifyRules::expect_pointer): likewise
(UnifyRules::expect_param): likewise
(UnifyRules::expect_array): likewise
(UnifyRules::expect_slice): likewise
(UnifyRules::expect_fndef): likewise
(UnifyRules::expect_fnptr): likewise
(UnifyRules::expect_tuple): likewise
(UnifyRules::expect_bool): likewise
(UnifyRules::expect_char): likewise
(UnifyRules::expect_int): likewise
(UnifyRules::expect_uint): likewise
(UnifyRules::expect_float): likewise
(UnifyRules::expect_isize): likewise
(UnifyRules::expect_usize): likewise
(UnifyRules::expect_never): likewise
(UnifyRules::expect_placeholder): likewise
(UnifyRules::expect_projection): likewise
(UnifyRules::expect_dyn): likewise
(UnifyRules::expect_closure): likewise

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: remove 2nd last usage of can_eq

gcc/rust/ChangeLog:

* typecheck/rust-tyty.cc (BaseType::satisfies_bound): use types_compatable instead

gcc/testsuite/ChangeLog:

* rust/compile/issue-1725-2.rs: remove bad error message

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: Add check bounds flag to unify rules for compatability checks

We need to make the type bounds check a flag because it can turn into a
recursive type bounds check. This allows us to remove another can_eq usage

gcc/rust/ChangeLog:

* typecheck/rust-type-util.cc (types_compatable):  add check bounds flag
(unify_site_and): likewise
* typecheck/rust-type-util.h (types_compatable): likewise
(unify_site_and): likewise
* typecheck/rust-tyty-bounds.cc: likewise
* typecheck/rust-unify.cc (UnifyRules::UnifyRules): likewise
(UnifyRules::Resolve): likewise
(UnifyRules::resolve_subtype): likewise
(UnifyRules::go): likewise
* typecheck/rust-unify.h: likewise

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: Update can_eq to be a types_compatable interface

Getting close to getting rid of can_eq and tyty-cmp soon.

gcc/rust/ChangeLog:

* typecheck/rust-hir-path-probe.cc (PathProbeType::process_impl_item_candidate):
refactor to types_compatable

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: respect the unify rules commit flag

We use the types compatable interface for unify here and so if
we dont respect the commit flag the interface can have unintended
side effects with infer type hints throwing things off down the line.

gcc/rust/ChangeLog:

* typecheck/rust-unify.cc (UnifyRules::expect_inference_variable): dont commit
(UnifyRules::expect_adt): likewise
(UnifyRules::expect_bool): likewise
(UnifyRules::expect_char): likewise
(UnifyRules::expect_int): likewise
(UnifyRules::expect_uint): likewise
(UnifyRules::expect_float): likewise
(UnifyRules::expect_isize): likewise
(UnifyRules::expect_usize): likewise

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: Remove some const usage so we can get rid of more can_eq usage

I needed to remove som usage of const so we can get rid of can_eq in
type bounds probe. This means we can use the types_compatable interface
instead. Which is much better.

gcc/rust/ChangeLog:

* backend/rust-compile-base.h: remove const
* backend/rust-compile-expr.cc: likewise
* backend/rust-compile.cc (HIRCompileBase::coerce_to_dyn_object): likewise
* typecheck/rust-hir-type-bounds.h: likewise
* typecheck/rust-type-util.cc (lookup_associated_impl_block): likewise
* typecheck/rust-type-util.h (lookup_associated_impl_block): likewise
* typecheck/rust-tyty-bounds.cc (TypeBoundsProbe::TypeBoundsProbe): likewise
(TypeBoundsProbe::Probe): likewise
* typecheck/rust-tyty-cmp.h: likewise
* typecheck/rust-tyty-subst.cc (SubstitutionRef::monomorphize): likewise
* typecheck/rust-tyty.cc (BaseType::satisfies_bound): likewise
(BaseType::bounds_compatible): likewise
(VariantDef::clone): likewise
(VariantDef::monomorphized_clone): likewise
(OpaqueType::is_equal): likewise
(DynamicObjectType::is_equal): likewise
* typecheck/rust-tyty.h: likewise

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: Make the coerce borrowed pointer consistent with try flag

This interface is not respecting the coercion api commit is always false,
we need to ensure this is respected via the try flag like the rest of the
coercion logic.

gcc/rust/ChangeLog:

* typecheck/rust-coercion.cc (TypeCoercionRules::select): respect try flag

Signed-off-by: Philip Herron <herron.philip@googlemail.com>

gccrs: Error out on unsuffixed self list use declaration

gcc/rust/ChangeLog:

* resolve/rust-early-name-resolver-2.0.cc (Early::finalize_rebind_import):
Replace assert with early break and remove early return.
(Early::visit): Check for unsuffixed lower self list.
* resolve/rust-early-name-resolver-2.0.h: Add visit function prototype.

gcc/testsuite/ChangeLog:

* rust/compile/use_self_alone_in_list.rs: New test.

Signed-off-by: Pierre-Emmanuel Patry <pierre-emmanuel.patry@embecosm.com>

gccrs: Emit error on lonely self use declaration

gcc/rust/ChangeLog:

* resolve/rust-early-name-resolver-2.0.cc (Early::visit): Emit an error
on top level rebind self use declaration.

gcc/testsuite/ChangeLog:

* rust/compile/use_self_alone.rs: New test.

Signed-off-by: Pierre-Emmanuel Patry <pierre-emmanuel.patry@embecosm.com>

gccrs: Allow early resolution to use the language prelude

gcc/rust/ChangeLog:

* Make-lang.in (GRS_OBJS): Add "rust-resolve-builtins.o".
* resolve/rust-late-name-resolver-2.0.cc: Include
"rust-resolve-builtins.h".
(next_node_id): Remove function.
(next_hir_id): Likewise.
(Late::setup_builtin_types): Likewise.
(Late::go): Use Builtins::setup_type_ctx instead of
Late::setup_builtin_types.
* resolve/rust-late-name-resolver-2.0.h
(Late::setup_builtin_types): Remove function.
* rust-session-manager.cc: Include "rust-resolve-builtins.h".
(Session::expansion): Call Builtins::setup_lang_prelude.
* resolve/rust-resolve-builtins.cc: New file.
* resolve/rust-resolve-builtins.h: New file.

gcc/testsuite/ChangeLog:

* rust/compile/primitive-import.rs: New test.

Signed-off-by: Owen Avery <powerboat9.gamer@gmail.com>

gccrs: Remove Parser::parse_tuple_index_expr_float

Unlike in C, floating point literals can't start with a '.', and
therefore could never be split into a '.' followed by an integer.

gcc/rust/ChangeLog:

* parse/rust-parse-impl.h (Parser::left_denotation): Remove
usage of parse_tuple_index_expr_float.
(Parser::parse_closure_expr_pratt): Remove function.

gcc/testsuite/ChangeLog:

* rust/compile/parse_float_dot.rs: New test.

Signed-off-by: Owen Avery <powerboat9.gamer@gmail.com>