As part of the work on changing side-effects removal in SPARK, a special
case was introduced to generate an Itype_Reference for Itypes in slices.
This was based on a misunderstanding of existing checks for bounds when
analyzing slices. These Itype_Reference are actually not needed to get
the corresponding run-time checks in GNATprove, and are actually harmful
in some cases (inside quantified expressions) as the insertion point for
the Itype_Reference ends up being outside of the quantifier scope,
leading to unprovable checks.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* sem_res.adb (Set_Slice_Subtype): Revert special-case
introduced previously, which is not needed as Itypes created for
slices are precisely always used.
diff --git a/gcc/ada/sem_res.adb b/gcc/ada/sem_res.adb
--- a/gcc/ada/sem_res.adb
+++ b/gcc/ada/sem_res.adb
@@ -12607,10 +12607,9 @@ package body Sem_Res is
-- the point where actions for the slice are analyzed). Note that this
-- is different from freezing the itype immediately, which might be
-- premature (e.g. if the slice is within a transient scope). This needs
- -- to be done only if expansion is enabled, or in GNATprove mode to
- -- capture the associated run-time exceptions if any.
+ -- to be done only if expansion is enabled.
- elsif Expander_Active or GNATprove_Mode then
+ elsif Expander_Active then
Ensure_Defined (Typ => Slice_Subtype, N => N);
end if;
end Set_Slice_Subtype;