public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
* [pushed] c++: xvalue array subscript [PR103185]
@ 2023-12-20 17:31 Jason Merrill
  0 siblings, 0 replies; only message in thread
From: Jason Merrill @ 2023-12-20 17:31 UTC (permalink / raw)
  To: gcc-patches

Tested x86_64-pc-linux-gnu, applying to trunk.

-- 8< --

Normally we handle xvalue array subscripting with ARRAY_REF, but in this
case we weren't doing that because the operands were reversed.  Handle that
case better.

	PR c++/103185

gcc/cp/ChangeLog:

	* typeck.cc (cp_build_array_ref): Handle swapped operands.

gcc/testsuite/ChangeLog:

	* g++.dg/cpp1z/array-prvalue2.C: New test.
	* g++.dg/cpp1z/eval-order3.C: Test swapped operands.
---
 gcc/cp/typeck.cc                            | 16 +++++++++++++---
 gcc/testsuite/g++.dg/cpp1z/array-prvalue2.C |  5 +++++
 gcc/testsuite/g++.dg/cpp1z/eval-order3.C    |  5 +++++
 3 files changed, 23 insertions(+), 3 deletions(-)
 create mode 100644 gcc/testsuite/g++.dg/cpp1z/array-prvalue2.C

diff --git a/gcc/cp/typeck.cc b/gcc/cp/typeck.cc
index a6e2f4ee7da..1fe30fc7c48 100644
--- a/gcc/cp/typeck.cc
+++ b/gcc/cp/typeck.cc
@@ -3932,13 +3932,14 @@ cp_build_indirect_ref (location_t loc, tree ptr, ref_operator errorstring,
    If INDEX is of some user-defined type, it must be converted to
    integer type.  Otherwise, to make a compatible PLUS_EXPR, it
    will inherit the type of the array, which will be some pointer type.
-   
+
    LOC is the location to use in building the array reference.  */
 
 tree
 cp_build_array_ref (location_t loc, tree array, tree idx,
 		    tsubst_flags_t complain)
 {
+  tree first = NULL_TREE;
   tree ret;
 
   if (idx == 0)
@@ -3983,6 +3984,14 @@ cp_build_array_ref (location_t loc, tree array, tree idx,
 
   bool non_lvalue = convert_vector_to_array_for_subscript (loc, &array, idx);
 
+  /* 0[array] */
+  if (TREE_CODE (TREE_TYPE (idx)) == ARRAY_TYPE)
+    {
+      std::swap (array, idx);
+      if (flag_strong_eval_order == 2 && TREE_SIDE_EFFECTS (array))
+	idx = first = save_expr (idx);
+    }
+
   if (TREE_CODE (TREE_TYPE (array)) == ARRAY_TYPE)
     {
       tree rval, type;
@@ -4058,15 +4067,16 @@ cp_build_array_ref (location_t loc, tree array, tree idx,
       protected_set_expr_location (ret, loc);
       if (non_lvalue)
 	ret = non_lvalue_loc (loc, ret);
+      if (first)
+	ret = build2_loc (loc, COMPOUND_EXPR, TREE_TYPE (ret), first, ret);
       return ret;
     }
 
   {
     tree ar = cp_default_conversion (array, complain);
     tree ind = cp_default_conversion (idx, complain);
-    tree first = NULL_TREE;
 
-    if (flag_strong_eval_order == 2 && TREE_SIDE_EFFECTS (ind))
+    if (!first && flag_strong_eval_order == 2 && TREE_SIDE_EFFECTS (ind))
       ar = first = save_expr (ar);
 
     /* Put the integer in IND to simplify error checking.  */
diff --git a/gcc/testsuite/g++.dg/cpp1z/array-prvalue2.C b/gcc/testsuite/g++.dg/cpp1z/array-prvalue2.C
new file mode 100644
index 00000000000..60a038cfd16
--- /dev/null
+++ b/gcc/testsuite/g++.dg/cpp1z/array-prvalue2.C
@@ -0,0 +1,5 @@
+// PR c++/103185
+// { dg-do compile { target c++11 } }
+
+using intarr = int[];
+static_assert(__is_same(decltype(0[intarr{0}]), int&&), "");
diff --git a/gcc/testsuite/g++.dg/cpp1z/eval-order3.C b/gcc/testsuite/g++.dg/cpp1z/eval-order3.C
index 5773591d49f..897cdef7483 100644
--- a/gcc/testsuite/g++.dg/cpp1z/eval-order3.C
+++ b/gcc/testsuite/g++.dg/cpp1z/eval-order3.C
@@ -166,6 +166,11 @@ void g()
   ip(24)[f(25)-25] = 0;
   last=0;
 
+  // even with swapped operands
+  (f(20)-20)[afn(21)].memfn(f(22),23);
+  (f(24)-24)[ip(25)] = 0;
+  last=0;
+
   // a << b
   aref(24) << f(25);
   iref(26) << f(27);

base-commit: a3ff76278efe006dc0b50249c8e5baf565bff56b
-- 
2.39.3


^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2023-12-20 17:31 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-12-20 17:31 [pushed] c++: xvalue array subscript [PR103185] Jason Merrill

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).