On Thu, May 07, 2026 at 11:04:35AM +0200, Claudio Jeker wrote:
> The accounting in chash.c has two issues:
> 
> - ch_table_resize() did not increase the global
>   type->t_counts->cc_num_extendible value and so a later _ch_destroy()
>   would underflow the accounting.
> 
> - ch_sub_free() can be called with NULL, NULL arguments and then the
>   accounting needs to be skipped.
> 
> Fixing these issues is trivial.

ok tb