Hi!

The following testcase ICEs in the -Wmisleading-indentation warning.
For C++11 and later expansion statement can appear in the set of keywords
the warning sees and it is the first keyword of it (for expansion stmt
of a pair of keywords), so RID_TEMPLATE.

Fixed thusly, bootstrapped/regtested on x86_64-linux and i686-linux,
ok for trunk?

2026-01-20  Jakub Jelinek  <[email protected]>

        PR c++/123694
        * c-indentation.cc (guard_tinfo_to_string): Handle RID_TEMPLATE
        for C++11 or later.

        * g++.dg/cpp26/expansion-stmt28.C: New test.

--- gcc/c-family/c-indentation.cc.jj    2026-01-02 09:56:09.910340858 +0100
+++ gcc/c-family/c-indentation.cc       2026-01-19 18:32:05.483488251 +0100
@@ -606,6 +606,9 @@ guard_tinfo_to_string (enum rid keyword)
       return "do";
     case RID_SWITCH:
       return "switch";
+    case RID_TEMPLATE:
+      gcc_assert (cxx_dialect >= cxx11);
+      return "template for";
     default:
       gcc_unreachable ();
     }
--- gcc/testsuite/g++.dg/cpp26/expansion-stmt28.C.jj    2026-01-19 
18:36:16.350153705 +0100
+++ gcc/testsuite/g++.dg/cpp26/expansion-stmt28.C       2026-01-19 
18:35:29.797956009 +0100
@@ -0,0 +1,13 @@
+// PR c++/123694
+// { dg-do compile { target c++11 } }
+// { dg-options "-Wmisleading-indentation" }
+
+int
+foo ()
+{
+  int b = 0;
+  template for (constexpr auto a : { 1, 2L, 3.0 })     // { dg-warning 
"'template for' only available with" "" { target c++23_down } }
+    b += a;                                            // { dg-warning "this 
'template for' clause does not guard\\.\\.\\." "" { target *-*-* } .-1 }
+    b++;                                               // { dg-message 
"\\.\\.\\.this statement, but the latter is misleadingly indented as if it were 
guarded by the 'template for'" }
+  return b;
+}

        Jakub

Reply via email to