[Freeciv-Dev] Re: [Patch] Stricter input checking (PR#1764)

[Raimar Falke <rf13@xxxxxxxxxxxxxxxxx>]

On Thu, Jul 18, 2002 at 10:21:51PM -0400, Ross W. Wetmore wrote:
> 
> This is a foolish mistake.
> 
> In general, requiring realness in virtually all cases is not bad. 
> Requiring normalized in any but carefully orchestrated ones is very 
> bad. 
> 
> Since normalize_map_pos() not only does a fast check for real, 
> but also insures the output coordinates are normalized, there is
> absolutely no need for an is_normal_map_pos test anywhere in the 
> code - ony realness tests using normalized_map_pos() are needed. 

Maybe you haven't notices but we are now dealing only with normal map
positions in the code.

> The replacements here are in fact errors waiting to happen and 
> will bring back the chaos of the void tile years where realness 
> was generally ignored.

Now this is FUD.

> Is_normal_map_pos was only added for debug purposes and was recognized 
> at the time as having no valid use outside of this. It is an 
> incomplete test for validity. It's use here is thus still invalid.

> Realness of coordinates is a hard condition that can only be dealt 
> with at source and is an error elsewhere.

Ack.

> Normalized is a soft and easily remedied one that can be handled
> anywhere in code

Ack.

> - for example
> as local efficiency mods. 

> The first *is* required of game coordinates.

Ack.

> The second is meaningless by itself as an incomplete
> test. Normalized is also not a rigorously defined concept while real
> is. Thus "passing normalized coordinates" is only useful wrt a local
> definition of normalized.

No ack. We have a definition of normal map positions.

> The long arguments are in the email history - go read them. Start 
> with the initial round with Gaute who was repudiated unanimously
> by the list.
> 
> Since Tony seems to be moving towards a new policy of considering pros 
> and cons, and I presume thus moving away from the rather autocratic and
> or chaotic historical Freeciv flavour, it surely behooves maintainers to
> perhaps be the first to respect this.

        Raimar

-- 
 email: rf13@xxxxxxxxxxxxxxxxx
 Windows: Where do you want to go today?
 Linux: Where do you want to go tomorrow?
 BSD: Are you guys coming or what?