p_ptr->command_cmd

[PaulBlay]

command_cmd of the player_type structure is s16b, but as far as I can tell it is never assigned values outside of the char range. Is more expected from command_cmd in the future?

As it is there are places, like

	/* Check for a "prevention" inscription */
	verify_inscrip[1] = p_ptr->command_cmd;

where I'm not sure if it is safe, or appropriate, to 'cast' p_ptr->command_cmd to (char) or not. I've been going through my (mostly your ;-) code tweaking it to remove compiler warnings and this is one of the few left. I don't want to cast stuff unless I'm pretty sure the RHS is really not going to exceed the cast specified.

[EDIT] Incidentally I think it would be nice if you could make a clean break between 'key pressed' and 'command action'. You could make macros and keymappings easier to read and more flexible for a start.

[PowerDiver]

command_cmd of the player_type structure is s16b, but as far as I can tell it is never assigned values outside of the char range. Is more expected from command_cmd in the future?

As it is there are places, like

	/* Check for a "prevention" inscription */
	verify_inscrip[1] = p_ptr->command_cmd;

where I'm not sure if it is safe, or appropriate, to 'cast' p_ptr->command_cmd to (char) or not. I've been going through my (mostly your ;-) code tweaking it to remove compiler warnings and this is one of the few left. I don't want to cast stuff unless I'm pretty sure the RHS is really not going to exceed the cast specified.

That's one of the things asserts are for.

[PaulBlay]

That's one of the things asserts are for.

Yeah, I've added a few of those, but most places it's pretty obvious.

(e.g. randint0(25) isn't going to exceed a char value ).

[takkaria]

command_cmd of the player_type structure is s16b, but as far as I can tell it is never assigned values outside of the char range. Is more expected from command_cmd in the future?

As it is there are places, like

	/* Check for a "prevention" inscription */
	verify_inscrip[1] = p_ptr->command_cmd;

where I'm not sure if it is safe, or appropriate, to 'cast' p_ptr->command_cmd to (char) or not. I've been going through my (mostly your ;-) code

I disclaim responsibility for the vast majority of Angband.

tweaking it to remove compiler warnings

What compiler are you using?

and this is one of the few left. I don't want to cast stuff unless I'm pretty sure the RHS is really not going to exceed the cast specified.

Looks like it's a fairly safe cast to make, yeah, but I have no exhaustive answer you for since I've not really looked at it and it involves inkey(), which I avoid like the plague.

[EDIT] Incidentally I think it would be nice if you could make a clean break between 'key pressed' and 'command action'. You could make macros and keymappings easier to read and more flexible for a start.

Yup, it would, and is somewhat slowly being worked on.

[PaulBlay]

What compiler are you using?

VC++ (which complains more) and Dev-C++ (which had been associated with an apparently compiler-related game crash but now seems to be behaving).

I now have only 4 warnings (down from 207). Wo-hoo!

Looks like it's a fairly safe cast to make, yeah, but I have no exhaustive answer you for since I've not really looked at it and it involves inkey(), which I avoid like the plague.

I'll stick an assert or two in there just in case then.

[Pete Mack]

@paul
Getting rid of all warnings is always a nice relaxing task that still feels satisfying to complete... And yes, Visual Studio, like gcc, is a big step up from off-brand compilers.