Hi! I've noticed the following typo in a comment, fixed thusly, committed as obvious to trunk.
2025-11-25 Jakub Jelinek <[email protected]> * alias.cc (get_alias_set): Fix comment typo, TYPE_CANOINCAL -> TYPE_CANONICAL. --- gcc/alias.cc.jj 2025-04-08 14:08:48.109325606 +0200 +++ gcc/alias.cc 2025-11-24 09:43:03.812342061 +0100 @@ -1099,7 +1099,7 @@ get_alias_set (tree t) p = build_pointer_type (p); gcc_checking_assert (p == TYPE_MAIN_VARIANT (p)); /* build_pointer_type should always return the canonical type. - For LTO TYPE_CANOINCAL may be NULL, because we do not compute + For LTO TYPE_CANONICAL may be NULL, because we do not compute them. Be sure that frontends do not glob canonical types of pointers in unexpected way and that p == TYPE_CANONICAL (p) in all other cases. */ Jakub
