Tested aarch64-linux. Pushed to trunk.

-- >8 --

There's no need to check for self-assignment here, it would just add
extra code for an unlikely case. Add a comment saying so.

libstdc++-v3/ChangeLog:

        PR libstdc++/100147
        * include/bits/gslice.h (operator=): Add comment about lack of
        self-assignment check.
---
 libstdc++-v3/include/bits/gslice.h | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/libstdc++-v3/include/bits/gslice.h 
b/libstdc++-v3/include/bits/gslice.h
index 3d78da2bd75..d9a8b929a57 100644
--- a/libstdc++-v3/include/bits/gslice.h
+++ b/libstdc++-v3/include/bits/gslice.h
@@ -169,6 +169,8 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION
   inline gslice&
   gslice::operator=(const gslice& __g)
   {
+    // Safe for self-assignment. Checking for it would add overhead just to
+    // optimize a case that should never happen anyway.
     if (__g._M_index)
       __g._M_index->_M_increment_use();
     if (_M_index && _M_index->_M_decrement_use() == 0)
-- 
2.43.0

Reply via email to