Re: Liberty Eiffel runtime C-code question

[Raphael Mack]

mh. When I look at the callers place I would say, that the (undefined
return value) is not used. So for me it looks like a bug, yes. Mainly
in sys/runtime/c/deep_twin.c, where the * should be removed.

What concerns me most with this fact is, that we don't seem to have a
test and not even a tutorial calling deep_twin.
Any volunteers to create this test?

As the additional * in this case would be easy to spot even for a C
compiler I wonder whether we should extend the C compiler warning flags
- at least for the tests on ET. - Which I (on the other side) also fear
a little.

Someone willing to compile with -Wall or maybe -Wreturn-type for the
start?

Regards,
Rapha

Am Montag, dem 31.01.2022 um 13:23 +0100 schrieb Hans Zwakenberg:
>  function se_deep_twin_trats() in deep_twin.c is defined as a
> function returning a void pointer: 
>  
>  void* se_deep_twin_trats(void) { 
> se_deep_twin_start_counter--; 
> if (se_deep_twin_start_counter == 0) { 
> se_hash_table_clear(se_deep_twin_memory); 
> }
>  
>  where does it return that pointer?   I don't quite get it.  Is this
> a bug? 
>  
>  cheers 
>  Hans 
>