Haven-res.jar hit-box fix. Fixed after multiple attempts.

[Xcom]

Wait. What an idiot.

I forgot I am using custom images for walls / cubs / cheese racks. Ofc it would mess up cause the custom res for these items haven't had a neg fix.

Sorry for false alarm.

Small update:

I used the script to edit the costum res files. Oddly they should have ended up

Code: Select all

Size bs   (10, 10)
Offset bc (-5, -5)

But instead they ended up with

Code: Select all

Size bs   (11, 11)
Offset bc (-6, -6)

[Xcom]

I had a suspicion that rounding were made differently.

Bumping around "gfx/terobjs/pow" and "gfx/terobjs/furniture/table" are also rounded up instead of down.

Code: Select all

Name      gfx/terobjs/pow
Size bs   (10, 10)
Offset bc (-5, -5)

should be

Code: Select all

Size bs   (9, 9)
Offset bc (-4, -4)

Currently

Code: Select all

Name      gfx/terobjs/furniture/table
Size bs   (11, 11)
Offset bc (-6, -6)

should be

Code: Select all

Size bs   (10, 10)
Offset bc (-5, -5)

Does the s2m conversion use a int rounding? That must be the reason why they come out differently as it should round down all numbers.

[boshaw]

Does the s2m conversion use a int rounding? That must be the reason why they come out differently as it should round down all numbers.

It floors all numbers, so they should be rounded down

[Xcom]

Does the s2m conversion use a int rounding? That must be the reason why they come out differently as it should round down all numbers.

It floors all numbers, so they should be rounded down

Very strange. The roundings must be different as about 25% of the converted s2m objects are over sized.

[Xcom]

I had a thorrow look through the numbers.

The s2m conversion should go through this calculation.

x = rounddown(c.x / 4) + rounddown(c.y / 2)

y = rounddown(c.y / 2) - rounddown(c.x / 4)

But the script must be doing this and throws off the sizes of some objects

x = rounddown( (c.x / 4) + (c.y / 2) )

y = rounddown( (c.y / 2) - (c.x / 4) )

Another thing to note. Most errors happen in BC and not in BS.

[boshaw]

I had a thorrow look through the numbers.

The s2m conversion should go through this calculation.

x = rounddown(c.x / 4) + rounddown(c.y / 2)

y = rounddown(c.y / 2) - rounddown(c.x / 4)

But the script must be doing this and throws off the sizes of some objects

x = rounddown( (c.x / 4) + (c.y / 2) )

y = rounddown( (c.y / 2) - (c.x / 4) )

Another thing to note. Most errors happen in BC and not in BS.

You're right, i overlooked that. I update the source to do the truncation properly for that: https://github.com/boshaw/salem-layer-u ... master.zip

[Xcom]

Just did all the updates and tested it. Looks perfect now.

Thanks again for all the help. Filter should edit all hit-boxes correctly except for the few listed objects:

These should have a smaller hit-box then currently
Chicken
Hen
Cock
Rabbits

These shouldn't have a hit-box when they do.
Moth
Dragonfly
Log on ground - fixable, but left in for there red outline when hiding trees.

[Oddity]

I would like a list of all the objects and their hitbox dimensions. <3

[Xcom]

Alot of new objects with hitboxes have been added. But I could make a haven-res.jar out of the new ones and the old ones and somehow upload it and link. Dono how to upload though.

[mvgulik]

I have always been kinda mystified about the general conclusion that "bc(Y)" is a(and only a) "offset" and "bs(Y)" is a(and only a) size.
Is that conclusion still uphold ?

Looking at the new Yew tree res data ... I think that conclusion is wrong (at least for tree's that is).

Code: Select all

01.res\neg   #bc(Y)    0   #bs(Y)   0
02.res\neg   #bc(Y)   -4   #bs(Y)   4
03.res\neg   #bc(Y)   -4   #bs(Y)   5
04.res\neg   #bc(Y)   -5   #bs(Y)   5
05.res\neg   #bc(Y)   -9   #bs(Y)   9
06.res\neg   #bc(Y)   -9   #bs(Y)   9


The data change from res "03" to res "04" would mean that only the "offset"/position of the tree hitbox is changed. Which seems to have no good reason for doing that.

I think ... that bc(Y) is the size(or point-offset if you like) of the Top (triangle)part of the hitbox diamond. And "bs(Y)" is used for the size of the bottom part. (Size being here the distance of the Top/Bottom corner of the hitbox diamond from the objects center point.)
Looking at the (auto generated it seems) changes between res "03", "04" and "05" this system makes more sense to me. CQ: The Tree hitbox is growing one unit in size, but as the hitbox is a mirror like shape, its done by just increasing one side(triangle) of the hitbox when needed.