Re: C++20 deprecation warning regarding enum arithmetic

["Pierre Chapuis" <lua@...>]

The value added is always TM_ADD, It's just a shift trick to find the metatable ID corresponding to a binary operation code.

If adding an integer works, it would be enough to do something like `#define TM_ADD_OFFSET ((int)TM_ADD)` and use this instead.

-- 
Pierre Chapuis

On Wed, Mar 22, 2023, at 16:12, Mouse wrote:
>> [...adding enums of different types...g++ warning...]
>> While you may of course cast one of the enum values to int before doing the $
>
> I would much prefer the cast.  The semantics are the same (or at least
> very close to) and IMO the cast makes it significantly clearer to the
> reader what's going on.
>
> However, I would also argue that using enums as "just funny ints",
> which semi-must be going on for adding enums from different
> enumerations to make any semantic sense at all, is a Very Bad Idea,
> even in C.  IMO manifest constants whose actual values are important
> should be #defines, not enum members.  (If there's a real need for
> value-important int-ish manifest constants with the scoping properties
> of enum constants rather than of #defines, I would say the right thing
> is to create a new kind of manifest constant, not to abuse enums.)
>
>> I wouldn't call it a bug, but [...]
>
> I would (would call it a bug, that is), but currently at just the
> semantic level.
>
> But what you or I think of it doesn't matter nearly as much as what the
> Lua people think of it.
>
> /~\ The ASCII				  Mouse
> \ / Ribbon Campaign
>  X  Against HTML		mouse@rodents-montreal.org
> / \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B