Report an uniterable for range once (#7661)

Building a `for` loop looks `Core.Iterate` up twice: once for
`NewCursor` to make the cursor, and again for `Next` to advance it. A
range that implements neither failed both lookups and reported both, so
a loop over something that isn't iterable produced two errors saying the
same thing about the same expression.

The second lookup is skipped when the first already failed, which is
what `BuildBinaryOperator`'s `diagnose` parameter is for. The
`ErrorInst` it returns instead does not reach the produced SemIR: the
loop is abandoned on the error either way.

Assisted-by: Claude Code

---------

Co-authored-by: josh11b <15258583+josh11b@users.noreply.github.com>
Co-authored-by: Richard Smith <richard@metafoo.co.uk>

toolchain/check/handle_loop_statement.cpp

@@ -224,11 +224,15 @@ auto HandleParseNode(Context& context, Parse::ForHeaderId node_id) -> bool {
       context, node_id,
       {.type_id = GetPointerType(context, cursor_type_inst_id),
        .lvalue_id = cursor_var_id});
+  // A range that implements neither fails both lookups; reporting the second
+  // would say the same thing about the same expression a second time.
+  // TODO: We should only perform the impl lookup once.
   auto element_id =
       BuildBinaryOperator(context, node_id,
                           {.interface_name = CoreIdentifier::Iterate,
                            .op_name = CoreIdentifier::Next},
-                          range_id, cursor_addr_id);
+                          range_id, cursor_addr_id,
+                          /*diagnose=*/cursor_id != SemIR::ErrorInst::InstId);
   // We need to convert away from an initializing expression in order to call
   // `HasValue` and then separately pattern-match against the element.
   // TODO: Instead, form a `.Some(pattern_id)` pattern and pattern-match against