Investigate a couple bogus warnings

Author: lloda

TODO

@@ -5,8 +5,8 @@ checkbox [C-ct] flip TODO
 
 #+STARTUP: logdone
 
-* TODO [1/7]
-  - [-] Bugs [5/16]
+* TODO [0/6]
+  - [-] Bugs [1/12]
     - [ ] We shouldn't need ra::iota(3, ra::dim_t(4)) to get iota to use Seq<dim_t> instead of Seq<int>.
     - [ ] Fix reduction-1.cc: "real_part(iter<1>(m)) = real_part(iter<1>(c))"
     - [ ] Invalid expressions that aren't caught by Match, either bc it makes assumptions, or bc
@@ -27,7 +27,9 @@ checkbox [C-ct] flip TODO
       way.
     - [ ] be namespace clean except for possibly the scalar block in ra.hh. Document exactly what is exported to ::.
     - [ ] bad uses of assert [ra17]
-  - [-] Features [3/17]
+    - [ ] poor shape mismatch report, e.g. [1 2] vs [2] produces 'len[0] is -192...' (magic number for MISS)
+      instead of e.g. 'len[0] MISS 1 vs 2'.
+  - [-] Features [0/15]
     - [ ] generalize cat
     - [-] support std::format
       - [X] basic support
@@ -65,7 +67,7 @@ checkbox [C-ct] flip TODO
         - [X] Scalar / unbeatable <2023-08-04 Fri 14:18>
         - [X] iota. But iota args need to integral constants for this to work.
       - [ ] Optimize iota(len...) + iota(len...)
-  - [-] Performance [2/4]
+  - [-] Performance [0/2]
     - [ ] bench/bench-stencil* is weird.
     - [ ] Bigd cases in bench-at.
   - [-] Building, tests [4/6]
@@ -101,44 +103,48 @@ Some of these aren't bugs in the sense that I expect to solve them, but more lik
 
 * Numbered issues through the source [ra..]
   1. [ ] test/bug10.cc
-  2. [ ]
-  3. [ ]
+  2. [ ] test/maxwell.cc, fromb(terminal case) -Waggressive-loop-optimizations with gcc14/15 -DRA_CHECK=0 -O3.
+    assert(ak>0) clears the warning, but [​[assume(ak>0)]] doesn't.
+  3. [ ] test/fromu.cc, test/genfrom.cc,fromb(terminal case) warnings -Warray-bounds -Wstringop-overflow with
+     gcc14/15 -DRA_CHECK=0 -O2 or -O3. assert(bk<ra::size(bv))) clears both warnings. [​[assume(...)]] clears
+     -Wstringop-overflow but not -Warray-bounds.
   4. [ ] test/optimize.cc: Expression objects only keep pointers to data objects. This means that it
-     is unsafe to define expression objects unless none of the data objects are
-     temporaries. E.g. `auto e = temp + temp;` is unsafe. Either `obj a; obj b; auto e = a + b;` or
-     `obj o = temp + temp;` is safe. Generally there's no reason to ever define expression objects
-     explicitly.
+    is unsafe to define expression objects unless none of the data objects are
+    temporaries. E.g. `auto e = temp + temp;` is unsafe. Either `obj a; obj b; auto e = a + b;` or
+    `obj o = temp + temp;` is safe. Generally there's no reason to ever define expression objects
+    explicitly.
+     1. [ ]
   5. [ ]
   6. [ ]
-  7. [ ]
-  8. [X] test/operators.cc: Some simple expressions with scalars fail in VAL(). The patch in ra.hh
+  7. [X] test/operators.cc: Some simple expressions with scalars fail in VAL(). The patch in ra.hh
      triggers the address sanitizer in test/ra-9.cc.
-  9. [ ] test/ra-6.cc: ra::Ptr doesn't hold copies. This enables restarting (see [ra39]), so
+  8. [ ] test/ra-6.cc: ra::Ptr doesn't hold copies. This enables restarting (see [ra39]), so
      ra::ptr(temp) must only be used as temp. Really forbidding auto ll = ra::ptr(lvalue) would also
      be ok.
-  10. [ ] test/ra-0.cc: size(SmallArray) requires ra:: to avoid collision with std::size, but not
-      size(Big) (??).
+  9. [ ] test/ra-0.cc: size(SmallArray) requires ra:: to avoid collision with std::size, but not
+     size(Big) (??).
+  10. [ ]
   11. [ ]
-  12. [ ]
-  13. [ ] ra/ply.hh (i/o), ra/expr.hh, ra/base.hh, test/compatibility.cc: std::string used to
+  12. [ ] ra/ply.hh (i/o), ra/expr.hh, ra/base.hh, test/compatibility.cc: std::string used to
       be registered as scalar, but that clashes with how std::ranges sees it. OTOH we don't want
       format(std::string_view) to print it as a foreign vector, so we have an exception for it
       there. As things stand, you can register it as scalar or not.
-  14. [ ] CellBig needs to copy its Dimv in some cases, which also complicates ViewBig::iter<>. Problem
+  13. [ ] CellBig needs to copy its Dimv in some cases, which also complicates ViewBig::iter<>. Problem
       is demonstrated in ra-5.cc.
-  15. [X] Conversion-to-scalar operators for dynamic-rank ViewBig(). I thought this could be
+  14. [X] Conversion-to-scalar operators for dynamic-rank ViewBig(). I thought this could be
       https://wg21.link/cwg976 but gcc 14 doesn't fix it. There are two issues here, 1) why isn't
       const conversion enough and 2) ambiguity with Small's 'any' constructor (cf 'c++ converting
       constructor vs conversion operator').
       - <2025-06-17 Tue> This was fixed mostly by replacing the (auto && x) constructors in the
         array classes by (const & x). This may cause iterators to be copied, so maybe we can redo
         (auto && x) in the future if we find out it matters.
-  16. [ ] Ambiguity in ravel vs nested constructors bc bc 1 converts to int2. Should be explicit
+  15. [ ] Ambiguity in ravel vs nested constructors bc bc 1 converts to int2. Should be explicit
       (and ravel constructors also).
-  17. [ ] assert() is used in some cases for runtime conditions that aren't ra::'s fault. This is
+  16. [ ] assert() is used in some cases for runtime conditions that aren't ra::'s fault. This is
       bad because if RA_ASSERT is defined to throw, the caller probably expects to be able to handle
       the error. On the other hand if RA_CHECK is 0, we shouldn't continue. But ofc we do in
       index checks etc. So is this different?
+  17. [ ]
   18. [ ]
   19. [ ]
   20. [ ]
@@ -153,28 +159,27 @@ Some of these aren't bugs in the sense that I expect to solve them, but more lik
   29. [ ]
   30. [ ]
   31. [ ]
-  32. [ ]
-  33. [X] test/frame-old.cc, test/fromb.cc: When mixing beaten & unbeaten subscripts and the
+  32. [X] test/frame-old.cc, test/fromb.cc: When mixing beaten & unbeaten subscripts and the
       unbeaten subscripts go first, the result is a nested expression. This has to be fixed in the
       view operator()s. Fixed after rewrite of from() e9ed62aa3954dd2e3dc528c51da7269971f9c553
       <2025-08-22 Fri>.
+  33. [ ]
   34. [ ]
   35. [ ]
   36. [ ]
   37. [ ]
-  38. [ ]
-  39. [ ] test/ra-9.cc: There's no reason to restart, since the Iterator methods are all static. So
+  38. [ ] test/ra-9.cc: There's no reason to restart, since the Iterator methods are all static. So
       start(ra::Scalar) just fwds. But that means that Scalar::c maintain constness, so a const
       overload is required for ScalarFlat::operator*.
+  39. [ ]
   40. [ ]
-  41. [ ]
-  42. [ ] Ways to test that compile time assertions or constraints are triggered appropriately,
+  41. [ ] Ways to test that compile time assertions or constraints are triggered appropriately,
       esp. for things like ct array lookup.  Some are doable with concepts, see e.g. ra-14.cc or
       big-0.cc.
+  42. [ ]
   43. [ ]
   44. [ ]
   45. [ ]
-  46. [ ]
 
 * Old bugs or issues
   - [X] Before merging saveload [1/1]

examples/maxwell.cc

@@ -67,9 +67,9 @@ int main()
         constexpr int k = k_;
         const int o = DA.len(k);
         if (o>=2) {
-            DA(HH<k>, iota(o-2, 1), HH<4-k>, k) = (A(HH<k>, iota(o-2, 2)) - A(HH<k>, iota(o-2, 0)));
-            DA(HH<k>, 0, HH<4-k>, k) = (A(HH<k>, 1) - A(HH<k>, o-1));
-            DA(HH<k>, o-1, HH<4-k>, k) = (A(HH<k>, 0) - A(HH<k>, o-2));
+            DA(HH<k>, iota(o-2, 1), HH<4-k>, k) = A(HH<k>, iota(o-2, 2)) - A(HH<k>, iota(o-2, 0));
+            DA(HH<k>, 0, HH<4-k>, k) = A(HH<k>, 1) - A(HH<k>, o-1);
+            DA(HH<k>, o-1, HH<4-k>, k) = A(HH<k>, 0) - A(HH<k>, o-2);
             DA(HH<5>, k) *= factor;
         }
     };

ra/ply.hh

@@ -535,13 +535,13 @@ constexpr auto
 fromb(auto pl, auto ds, auto && bv, int bk, auto && a, int ak)
 {
     if constexpr (dobv) {
-#pragma GCC diagnostic push
-#pragma GCC diagnostic warning "-Waggressive-loop-optimizations" // gcc14/15.2 -DRA_CHECK=0 -O3
+#pragma GCC diagnostic push // gcc14/15 -DRA_CHECK=0 --no-sanitize -O3 [ra02]
+#pragma GCC diagnostic warning "-Waggressive-loop-optimizations"
         for (; ak<ra::rank(a); ++ak, ++bk) {
 #pragma GCC diagnostic pop
-#pragma GCC diagnostic push
-#pragma GCC diagnostic warning "-Warray-bounds" // gcc14/15.2 -DRA_CHECK=0 --no-sanitize -O2 -O3
-#pragma GCC diagnostic warning "-Wstringop-overflow" // gcc14/15.2 -DRA_CHECK=0 --no-sanitize -O2 -O3
+#pragma GCC diagnostic push // gcc14/15 -DRA_CHECK=0 --no-sanitize -O2 -O3 [ra03]
+#pragma GCC diagnostic warning "-Warray-bounds"
+#pragma GCC diagnostic warning "-Wstringop-overflow"
             bv[bk] = a.dimv[ak];
 #pragma GCC diagnostic pop
         }

test/bug10.cc

@@ -14,7 +14,7 @@
 using std::cout;
 
 // FIXME [ra01] bug only shows up with all of: --no-sanitize -O>2, no assert (1)
-// see the patch in ViewSmall(nested braces)
+// patched with a barrier in ViewSmall(nested braces)
 
 int main()
 {

test/fromu.cc

@@ -98,13 +98,13 @@ int main()
             tr.info("a(rank0, rank1)").test_eq(ra::Small<real, 2>{9, 7}, from(a, 1, Vint{1, 0}));
             tr.info("a(rank1, rank0)").test_eq(ra::Small<real, 2>{9, 1}, from(a, Vint{1, 0}, ra::Small<int>(1)));
             tr.info("a(rank0, rank1)").test_eq(ra::Small<real, 2>{9, 7}, from(a, ra::Small<int>(1), Vint{1, 0}));
-// using .iter() wass necessary when mixed u/b weren't accepted. Now rank 0 are always beaten, so it doesn't matter.
+// needed .iter() when mixed u/b weren't accepted. Now rank 0 are always beaten, so it doesn't matter.
             tr.info("a(rank1, rank0)").test_eq(ra::Small<real, 2>{9, 1}, from(a, Vint{1, 0}, ra::Small<int>(1).iter()));
             tr.info("a(rank0, rank1)").test_eq(ra::Small<real, 2>{9, 7}, from(a, ra::Small<int>(1).iter(), Vint{1, 0}));
         };
         check_selection_unbeatable_2(Ureal<2>({2, 2}, {1, 2, 3, 4}));
         check_selection_unbeatable_2(ra::Small<real, 2, 2>({1, 2, 3, 4}));
-        check_selection_unbeatable_2(Ureal<>({2, 2}, {1, 2, 3, 4}));
+        check_selection_unbeatable_2(Ureal<>({2, 2}, {1, 2, 3, 4})); // [ra03]
     }
     tr.section("mixed scalar/unbeatable, 2D -> 1D");
     {

test/genfrom.cc

@@ -350,7 +350,7 @@ int main()
     tr.section("unbeaten ViewBig<... ANY> IIb");
     {
         auto i = ra::Small<int, 2, 3> {{3, 2, 1}, {4, 5, 6}};
-        auto b = from(ra::ViewBig<ra::Seq<ra::dim_t>, ra::ANY>(ra::ii({10, 2})), i, 0);
+        auto b = from(ra::ViewBig<ra::Seq<ra::dim_t>, ra::ANY>(ra::ii({10, 2})), i, 0); // [ra03]
         ra::Small<int, 2, 3> c = b;
         println(cout, "c {:c}", c);
         tr.strict().test_eq(ra::Small<int, 2, 3> {{6, 4, 2}, {8, 10, 12}}, c);

test/io.cc

@@ -199,7 +199,7 @@ int main()
         o << fmt(ra::nstyle, A);
         tr.test_eq(std::string("1 2\n3 4"), o.str());
     }
-    tr.section("[ra02a] printing Map");
+    tr.section("printing Map");
     {
         iocheck(tr.info("output of map (1)"),
                 ra::map_([](double i) { return -i; }, start(ra::Small<double, 3>{0, 1, 2})),
@@ -214,7 +214,7 @@ int main()
                 ra::map_([](int i) { return -i; }, a.iter()),
                 ra::Unique<int, 2>({2, 3}, { -1, -2, -3, -4, -5, -6 }));
     }
-    tr.section("[ra02b] printing array iterators");
+    tr.section("printing array iterators");
     {
         ra::Unique<int, 2> a({3, 2}, { 1, 2, 3, 4, 5, 6 });
         iocheck(tr.info("output of array through its iterator"), a.iter(), a);
@@ -223,7 +223,7 @@ int main()
                 transpose(a, {1, 0}).iter(),
                 ra::Unique<int>({2, 3}, { 1, 3, 5, 2, 4, 6 }));
     }
-    tr.section("[ra02c] printing array iterators");
+    tr.section("printing array iterators");
     {
         ra::Small<int, 3, 2> a { 1, 2, 3, 4, 5, 6 };
         iocheck(tr.info("output of array through its iterator"), a.iter(), a);