* [PATCH] c++: Fix up constexpr handling of char/signed char/short pre/post inc/decrement [PR105774]
@ 2022-10-24 7:27 Jakub Jelinek
2022-10-24 14:07 ` Jason Merrill
0 siblings, 1 reply; 2+ messages in thread
From: Jakub Jelinek @ 2022-10-24 7:27 UTC (permalink / raw)
To: Jason Merrill; +Cc: gcc-patches
Hi!
signed char, char or short int pre/post inc/decrement are represented by
normal {PRE,POST}_{INC,DEC}REMENT_EXPRs in the FE and only gimplification
ensures that the {PLUS,MINUS}_EXPR is done in unsigned version of those
types:
case PREINCREMENT_EXPR:
case PREDECREMENT_EXPR:
case POSTINCREMENT_EXPR:
case POSTDECREMENT_EXPR:
{
tree type = TREE_TYPE (TREE_OPERAND (*expr_p, 0));
if (INTEGRAL_TYPE_P (type) && c_promoting_integer_type_p (type))
{
if (!TYPE_OVERFLOW_WRAPS (type))
type = unsigned_type_for (type);
return gimplify_self_mod_expr (expr_p, pre_p, post_p, 1, type);
}
break;
}
This means during constant evaluation we need to do it similarly (either
using unsigned_type_for or using widening to integer_type_node).
The following patch does the latter.
Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
2022-10-24 Jakub Jelinek <jakub@redhat.com>
PR c++/105774
* constexpr.cc (cxx_eval_increment_expression): For signed types
that promote to int, evaluate PLUS_EXPR or MINUS_EXPR in int type.
* g++.dg/cpp1y/constexpr-105774.C: New test.
--- gcc/cp/constexpr.cc.jj 2022-10-21 18:04:47.869797312 +0200
+++ gcc/cp/constexpr.cc 2022-10-23 18:43:27.003390282 +0200
@@ -6234,6 +6234,18 @@ cxx_eval_increment_expression (const con
offset = fold_build1 (NEGATE_EXPR, TREE_TYPE (offset), offset);
mod = fold_build2 (POINTER_PLUS_EXPR, type, val, offset);
}
+ else if (c_promoting_integer_type_p (type)
+ && !TYPE_UNSIGNED (type)
+ && TYPE_PRECISION (type) < TYPE_PRECISION (integer_type_node))
+ {
+ offset = fold_convert (integer_type_node, offset);
+ mod = fold_convert (integer_type_node, val);
+ tree t = fold_build2 (inc ? PLUS_EXPR : MINUS_EXPR, integer_type_node,
+ mod, offset);
+ mod = fold_convert (type, t);
+ if (TREE_OVERFLOW_P (mod) && !TREE_OVERFLOW_P (t))
+ TREE_OVERFLOW (mod) = false;
+ }
else
mod = fold_build2 (inc ? PLUS_EXPR : MINUS_EXPR, type, val, offset);
if (!ptr)
--- gcc/testsuite/g++.dg/cpp1y/constexpr-105774.C.jj 2022-10-23 18:44:15.587729613 +0200
+++ gcc/testsuite/g++.dg/cpp1y/constexpr-105774.C 2022-10-23 18:33:54.754170726 +0200
@@ -0,0 +1,15 @@
+// PR c++/105774
+// { dg-do compile { target c++14 } }
+
+constexpr signed char
+foo ()
+{
+#if __SCHAR_MAX__ < __INT_MAX__
+ signed char x = __SCHAR_MAX__;
+#else
+ signed char x = 0;
+#endif
+ return ++x;
+}
+
+constexpr auto a = foo ();
Jakub
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [PATCH] c++: Fix up constexpr handling of char/signed char/short pre/post inc/decrement [PR105774]
2022-10-24 7:27 [PATCH] c++: Fix up constexpr handling of char/signed char/short pre/post inc/decrement [PR105774] Jakub Jelinek
@ 2022-10-24 14:07 ` Jason Merrill
0 siblings, 0 replies; 2+ messages in thread
From: Jason Merrill @ 2022-10-24 14:07 UTC (permalink / raw)
To: Jakub Jelinek; +Cc: gcc-patches
On 10/24/22 03:27, Jakub Jelinek wrote:
> Hi!
>
> signed char, char or short int pre/post inc/decrement are represented by
> normal {PRE,POST}_{INC,DEC}REMENT_EXPRs in the FE and only gimplification
> ensures that the {PLUS,MINUS}_EXPR is done in unsigned version of those
> types:
> case PREINCREMENT_EXPR:
> case PREDECREMENT_EXPR:
> case POSTINCREMENT_EXPR:
> case POSTDECREMENT_EXPR:
> {
> tree type = TREE_TYPE (TREE_OPERAND (*expr_p, 0));
> if (INTEGRAL_TYPE_P (type) && c_promoting_integer_type_p (type))
> {
> if (!TYPE_OVERFLOW_WRAPS (type))
> type = unsigned_type_for (type);
> return gimplify_self_mod_expr (expr_p, pre_p, post_p, 1, type);
> }
> break;
> }
> This means during constant evaluation we need to do it similarly (either
> using unsigned_type_for or using widening to integer_type_node).
> The following patch does the latter.
>
> Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
OK.
> 2022-10-24 Jakub Jelinek <jakub@redhat.com>
>
> PR c++/105774
> * constexpr.cc (cxx_eval_increment_expression): For signed types
> that promote to int, evaluate PLUS_EXPR or MINUS_EXPR in int type.
>
> * g++.dg/cpp1y/constexpr-105774.C: New test.
>
> --- gcc/cp/constexpr.cc.jj 2022-10-21 18:04:47.869797312 +0200
> +++ gcc/cp/constexpr.cc 2022-10-23 18:43:27.003390282 +0200
> @@ -6234,6 +6234,18 @@ cxx_eval_increment_expression (const con
> offset = fold_build1 (NEGATE_EXPR, TREE_TYPE (offset), offset);
> mod = fold_build2 (POINTER_PLUS_EXPR, type, val, offset);
> }
> + else if (c_promoting_integer_type_p (type)
> + && !TYPE_UNSIGNED (type)
> + && TYPE_PRECISION (type) < TYPE_PRECISION (integer_type_node))
> + {
> + offset = fold_convert (integer_type_node, offset);
> + mod = fold_convert (integer_type_node, val);
> + tree t = fold_build2 (inc ? PLUS_EXPR : MINUS_EXPR, integer_type_node,
> + mod, offset);
> + mod = fold_convert (type, t);
> + if (TREE_OVERFLOW_P (mod) && !TREE_OVERFLOW_P (t))
> + TREE_OVERFLOW (mod) = false;
> + }
> else
> mod = fold_build2 (inc ? PLUS_EXPR : MINUS_EXPR, type, val, offset);
> if (!ptr)
> --- gcc/testsuite/g++.dg/cpp1y/constexpr-105774.C.jj 2022-10-23 18:44:15.587729613 +0200
> +++ gcc/testsuite/g++.dg/cpp1y/constexpr-105774.C 2022-10-23 18:33:54.754170726 +0200
> @@ -0,0 +1,15 @@
> +// PR c++/105774
> +// { dg-do compile { target c++14 } }
> +
> +constexpr signed char
> +foo ()
> +{
> +#if __SCHAR_MAX__ < __INT_MAX__
> + signed char x = __SCHAR_MAX__;
> +#else
> + signed char x = 0;
> +#endif
> + return ++x;
> +}
> +
> +constexpr auto a = foo ();
>
> Jakub
>
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2022-10-24 14:07 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-10-24 7:27 [PATCH] c++: Fix up constexpr handling of char/signed char/short pre/post inc/decrement [PR105774] Jakub Jelinek
2022-10-24 14:07 ` 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).