[VERY HIGH] Billboard position incorrect again

[celebaglar]

The billboard at (x,y-2) in the "north" view is being displayed with the wrong vertical offset again (too high up the screen).

e.g. From any of the liths in the domain of the Moon, look at the corresponding one directly north or south. See?

[celebaglar]

With hindisght, not having noticed this effect anywhere else, I think it must be a localised problem.

Screenshot to follow shortly.

[Natmus]

Celebaglar said:

With hindisght, not having noticed this effect anywhere else, I think it must be a localised problem.

May not localised, but the line of liths reveal it like nowhere else.

Screenshot to follow shortly.

I have one. Standing in the middle of the southern line looking at the northern one (ignore the wolfie in the centre, please). It's clear the one directly to the north is to far up in the picture compared to the ones next to it. The two furthest away seem a tad high up as well, considering we should be looking at a line of liths running east-west.

[celebaglar]

Sorry. I meant that the problem only seems to affect liths as far as I have seen. For instance, if it affected forests this would be very obvious.

I should add that I can confirm this does affect all liths in the "polar" displays (N, E, S, W) when the lith is 2 leagues directly ahead of the observer.

Aside from checking code, my suggestion would be to check the lith graphic for that particular scale.

[sparrowhawk]

Celebaglar said:

The billboard at (x,y-2) in the "north" view is being displayed with the wrong vertical offset again (too high up the screen).

e.g. From any of the liths in the domain of the Moon, look at the corresponding one directly north or south. See?

Offsets have not changed since you sent the new ones. Could the problem have always been there?

[celebaglar]

Jean-Yves said:

Offsets have not changed since you sent the new ones. Could the problem have always been there?

Not prior to 0.2, certainly. I can't say whether it appeared with 0.2 or with the UKLinux.net power cut.

However, if the problem was the offsets, you'd notice it every time you looked north, west, east or south at a big forest because you'd have trees sticking out way above the rest of the forest.

I have only noticed the problem with liths, but it does affect liths everywhere. It's most obvious up in Moon, but I've also looked at liths in Corelay and seen armies poking out from underneath them.

I can't say how the problem may have appeared, but looking from the outside I'd make a wild guess that the lith graphic at that particular scale somehow has a transparent section under the lith itself, making it appear as if the offset is wrong. If that particular scale only appears at that location, it would explain why it is not noticeable anywhere else.

I will take a look at the offsets I gave you to make sure they are correct, but I cannot check the graphics.

[Natmus]

Looks to me like the three centre liths actually are the same size and use the same graphic, and if this is so it must have something to do with the offsets as well, not just the graphics.

[celebaglar]

Alright, my initial investigations point at the problem with the lith graphic more than ever.

The displacement value is fine.

//Size 3
array(3, -1, -2, 88, -41),
array(3, -1, 2, 88, 241),
array(3, -2, -1, 88, 41),
array(3, -2, 1, 88, 159),

//Size 2
array(2, -2, 0, 86, 96),

//Size 1
array(1, -1, -1, 78, -17),
array(1, -1, 1, 78, 183),

//Size 0
array(0, -1, 0, 66, 64),

As you can see, the displacement value for Size 2 is fine.

However, I can now confirm that the lith graphic of Size 2 always shows further up the screen than it should, even on the NE view. Here are two screenshots from the same location. The first is looking north and the second northeast.



This strongly suggests that the problem is with the Size 2 graphic for liths. If you prefer, you could package all the graphics you use in this version and send them to me. I will check all of them to make sure they all have the same baseline and return them with any corrections if they need to be made.

[sparrowhawk]

OK will do so on Tuesday or maybe tonight if I have some free time! Busy bee at the moment...

[sparrowhawk]

OK, couldn't wait - looked into it further:

Seems like lith2.png and lith3.png were the wrong way around. Have renamed them (swapped their names around). Hopefully this fixes things...

[sparrowhawk]

Any joy?

[Natmus]

Jean-Yves said:

Any joy?

I checked yesterday (monday) evening, and the liths were still out of wack, same thing as before. I even took a screenshot of it, but then I saw the changelog hadn't been updated, so I thougt you hadn't changed the code yet.

So, if you have changed the code yesterday or earlier, it didn't work.

[celebaglar]

Jean-Yves said:

Any joy?

Now the Size 2 seems identical to the Size 3. Same physical size, same displacement.

[sparrowhawk]

Seems as though there was a problem with the FTPing of files yesterday. Have uploaded again, could you check again please?

Bill, I have sent you the gfx as requested. Could you take a look and just check that I am not going mad? (Madder?)

[celebaglar]

Jean-Yves said:

Seems as though there was a problem with the FTPing of files yesterday. Have uploaded again, could you check again please?

Bill, I have sent you the gfx as requested. Could you take a look and just check that I am not going mad? (Madder?)

OK, I've had a very quick look.

The display now looks OK-ish. The "-ish" bit is because of the apparent discrepancy you mention. Looking at the lith graphics as a series seems to show that size 3 is the one which looks bigger than you'd expect, but I need to do the calculations before I can confirm that.

Could you please send me another terrain type (mountains or citadels for instance) so I can compare the relative proportions? Cheers.

In fact, it may help if you send me all the terrain graphics. I can use a Konqueror function to produce a gallery at the touch of a button, complete with the actual sizes. This will allow for a very quick and easy scan through all the graphics to see if I can spot any discrepancies. A matter of minutes.