[Bug c++/70170] [6 regression] bogus not a constant expression error comparing pointer to array to null

[msebor at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70170

Martin Sebor  changed:

   What|Removed |Added

 Status|ASSIGNED|RESOLVED
 Resolution|--- |FIXED

--- Comment #13 from Martin Sebor  ---
Fixed by r234698.

[Bug c++/70170] [6 regression] bogus not a constant expression error comparing pointer to array to null

[msebor at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70170

--- Comment #12 from Martin Sebor  ---
Author: msebor
Date: Sat Apr  2 17:14:48 2016
New Revision: 234698

URL: https://gcc.gnu.org/viewcvs?rev=234698=gcc=rev
Log:
PR c++/67376 - [5/6 regression] Comparison with pointer to past-the-end
of array fails inside constant expression
PR c++/70170 - [6 regression] bogus not a constant expression error comparing
pointer to array to null
PR c++/70172 - incorrect reinterpret_cast from integer to pointer error
on invalid constexpr initialization
PR c++/70228 - insufficient detail in diagnostics for a constexpr out of bounds
array subscript

gcc/testsuite/ChangeLog:
2016-04-02  Martin Sebor  

PR c++/67376
PR c++/70170
PR c++/70172
PR c++/70228
* g++.dg/cpp0x/constexpr-array-ptr10.C: New test.
* g++.dg/cpp0x/constexpr-array-ptr9.C: New test.
* g++.dg/cpp0x/constexpr-nullptr-1.C: New test.
* g++.dg/cpp0x/constexpr-array5.C: Adjust text of expected diagnostic.
* g++.dg/cpp0x/constexpr-string.C: Same.
* g++.dg/cpp0x/constexpr-wstring2.C: Same.
* g++.dg/cpp0x/pr65398.C: Same.
* g++.dg/ext/constexpr-vla1.C: Same.
* g++.dg/ext/constexpr-vla2.C: Same.
* g++.dg/ext/constexpr-vla3.C: Same.
* g++.dg/ubsan/pr63956.C: Same.

gcc/cp/ChangeLog:
2016-04-02  Martin Sebor  

PR c++/67376
PR c++/70170
PR c++/70172
PR c++/70228
* constexpr.c (diag_array_subscript): New function.
(cxx_eval_array_reference): Detect out of bounds array indices.

gcc/ChangeLog:
2016-04-02  Martin Sebor  

PR c++/67376
* fold-const.c (maybe_nonzero_address): New function.
(fold_comparison): Call it.  Fold equality and relational
expressions involving null pointers.
(tree_single_nonzero_warnv_p): Call maybe_nonzero_address.

Added:
trunk/gcc/testsuite/g++.dg/cpp0x/constexpr-array-ptr10.C
trunk/gcc/testsuite/g++.dg/cpp0x/constexpr-array-ptr9.C
trunk/gcc/testsuite/g++.dg/cpp0x/constexpr-nullptr-1.C
Modified:
trunk/gcc/ChangeLog
trunk/gcc/cp/ChangeLog
trunk/gcc/cp/constexpr.c
trunk/gcc/fold-const.c
trunk/gcc/testsuite/ChangeLog
trunk/gcc/testsuite/g++.dg/cpp0x/constexpr-array5.C
trunk/gcc/testsuite/g++.dg/cpp0x/constexpr-string.C
trunk/gcc/testsuite/g++.dg/cpp0x/constexpr-wstring2.C
trunk/gcc/testsuite/g++.dg/cpp0x/pr65398.C
trunk/gcc/testsuite/g++.dg/ext/constexpr-vla1.C
trunk/gcc/testsuite/g++.dg/ext/constexpr-vla2.C
trunk/gcc/testsuite/g++.dg/ext/constexpr-vla3.C
trunk/gcc/testsuite/g++.dg/ubsan/pr63956.C

[Bug c++/70170] [6 regression] bogus not a constant expression error comparing pointer to array to null

[msebor at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70170

Martin Sebor  changed:

   What|Removed |Added

   Assignee|mpolacek at gcc dot gnu.org|msebor at gcc dot 
gnu.org

--- Comment #11 from Martin Sebor  ---
Marek, since I posted my patch for review I've reassigned this bug to myself:
https://gcc.gnu.org/ml/gcc-patches/2016-03/msg00796.html

[Bug c++/70170] [6 regression] bogus not a constant expression error comparing pointer to array to null

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70170

--- Comment #10 from Jakub Jelinek  ---
(In reply to Jason Merrill from comment #9)
> (In reply to Jakub Jelinek from comment #7)
> > A small problem is that the second argument to POINTER_PLUS_EXPR is not
> > signed, but unsigned (sizetype).  Which is why I wrote "negative", negative
> > would mean having the most significant bit set or so.
> 
> This seems to cause a variety of problems.  Since it can be negative, surely
> it should have ssizetype?

I'll let Richi or Andrew Pinski comment on that design decision.  I fear
changing that in stage4 is not really possible.

[Bug c++/70170] [6 regression] bogus not a constant expression error comparing pointer to array to null

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70170

--- Comment #9 from Jason Merrill  ---
(In reply to Jakub Jelinek from comment #7)
> A small problem is that the second argument to POINTER_PLUS_EXPR is not
> signed, but unsigned (sizetype).  Which is why I wrote "negative", negative
> would mean having the most significant bit set or so.

This seems to cause a variety of problems.  Since it can be negative, surely it
should have ssizetype?

[Bug c++/70170] [6 regression] bogus not a constant expression error comparing pointer to array to null

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70170

--- Comment #8 from Marek Polacek  ---
Ah, true... well there's wi::sign_mask so maybe that.  But given Comment 5 I'll
hold off for now.

[Bug c++/70170] [6 regression] bogus not a constant expression error comparing pointer to array to null

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70170

--- Comment #7 from Jakub Jelinek  ---
A small problem is that the second argument to POINTER_PLUS_EXPR is not signed,
but unsigned (sizetype).  Which is why I wrote "negative", negative would mean
having the most significant bit set or so.

[Bug c++/70170] [6 regression] bogus not a constant expression error comparing pointer to array to null

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70170

--- Comment #6 from Marek Polacek  ---
-fno-delete-null-pointer-checks seems to change nothing.  But the case you
mention is something I wanted to think about more before posting a real patch.  

I suppose I could add a check that the offset part of POINTER_PLUS_EXPR is >=
0, then we should be safe (pointer arithmetic is undefined if it overflows of
course for both C/C++).

[Bug c++/70170] [6 regression] bogus not a constant expression error comparing pointer to array to null

[msebor at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70170

--- Comment #5 from Martin Sebor  ---
It turns out that my patch for bug 67376 (still being tested) fixes this bug.

[Bug c++/70170] [6 regression] bogus not a constant expression error comparing pointer to array to null

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70170

--- Comment #4 from Jakub Jelinek  ---
Shouldn't -fno-delete-null-pointer-checks disable that?  Or maybe only if the
constant is "negative", because with -fno-delete-null-pointer-checks in theory
some object could live at address NULL and you could have:
ptr = _at_0[0];
ptr += 4;
...
if (ptr - 4 == NULL)

[Bug c++/70170] [6 regression] bogus not a constant expression error comparing pointer to array to null

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70170

--- Comment #3 from Marek Polacek  ---
I think that to fix this, we need to add a new pattern to match.pd that deals
with "(ptr +p off) !=/== 0B".  Thus something like:

--- a/gcc/match.pd
+++ b/gcc/match.pd
@@ -2263,6 +2263,11 @@ DEFINE_INT_AND_FLOAT_ROUND_FN (RINT)
  (simplify
   (cmp (convert? addr@0) integer_zerop)
   (if (tree_single_nonzero_warnv_p (@0, NULL))
+   { constant_boolean_node (cmp == NE_EXPR, type); }))
+
+ (simplify
+  (cmp (pointer_plus (convert? @0) INTEGER_CST@1) integer_zerop)
+  (if (tree_single_nonzero_warnv_p (@0, NULL))
{ constant_boolean_node (cmp == NE_EXPR, type); })))

 /* If we have (A & C) == C where C is a power of 2, convert this into

[Bug c++/70170] [6 regression] bogus not a constant expression error comparing pointer to array to null

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70170

Marek Polacek  changed:

   What|Removed |Added

 Status|NEW |ASSIGNED
 CC||mpolacek at gcc dot gnu.org
   Assignee|unassigned at gcc dot gnu.org  |mpolacek at gcc dot 
gnu.org

--- Comment #2 from Marek Polacek  ---
I'll look but I think there's a dup somewhere, or at least something very
similar.

[Bug c++/70170] [6 regression] bogus not a constant expression error comparing pointer to array to null

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70170

Jakub Jelinek  changed:

   What|Removed |Added

 Status|UNCONFIRMED |NEW
   Last reconfirmed||2016-03-11
 CC||jakub at gcc dot gnu.org,
   ||jason at gcc dot gnu.org
 Ever confirmed|0   |1

--- Comment #1 from Jakub Jelinek  ---
Indeed, started with r230365.

[Bug c++/70170] [6 regression] bogus not a constant expression error comparing pointer to array to null

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70170

Richard Biener  changed:

   What|Removed |Added

   Priority|P3  |P1
   Target Milestone|--- |6.0