Celestia Forums

I thought that the pre3 bug fix had cured things with basesplit1 BUT !
although it cured the strange tile assignations previously referred to, it didn't cure (or it enabled me to see previously invisible !) strange effects !!

There is now a problem in that the tiles for (say) level3 are being painted with (called up from) tiles of the same name in the lower (level2) folder/directory !

It is suggestive of further problems in indexing and loop counters of the form (m,n) or (i,j) ??

dunno, I'm still investigating the hows&whys but thought i'd post a prelim heads-up ! there is also a prob that some tiles are painted with an absent (white flat) tile unless an "enabling file/tile" exists in a lower structure/folder/directory ?? I'll post more when I get get my head round it all
No prob, curious is all, just that it stops climbing Mt Everest to lower
levels !!
Malcolm.

Oh! since posting the above,
I see people talking about 132pre4 ?? did I blink or did I fall asleep ?
I thougt pre3 woz the latest ?
Can anyone point me to the announce and does it change/fix my post about BaseSplit1 ??
Is there a forum for announces of pre/new releases that I am missing ?

Thanks,
Malcolm.

1.3.2pre4 has some annoying bugs . . . I'll have prerelease 5 available within a day, so just wait for that one.

--Chris

chris wrote:prerelease 5 available within a day, so just wait for that one.

Breathless anticipation ! hehee !!
Thanks, best wishes,
Malcolm

I wrote:the tiles for (say) level3 are being painted with (called up from) tiles of the same name in the lower (level2) folder/directory

Chris wrote:prerelease 5 available within a day, so just wait for that one.

Problem is still present in pre5 !
For example the tile position _x_y at level3 is being painted with the texture file level2_tx_x_y. (at least I think that is what is happening !_
Does that make sense ?

Another example, if I make a jpg file tx_7_1 and put it in level2
and then I make another different jpg file tx_7_1 and put it in level3 folder,
the level2 7_1 picture gets displayed twice, once correctly in level2_7_1 position on surface and again when zooming in to display a level3 surface the same (level2)picture, but smaller, gets displayed where the level3 tile ought to be

hth,
Malcolm

I don't know if this is the problem here, but I have seen the following: if not al VT tiles have the same colour depth, I see strange behaviour and sometimes tiles mixed in from different levels. So it seems worth while to check the colour depth of all of your tiles in the same VT.

Hi!
Last weekend I tried to play with VTs, just to understand how they work: I started from a "dds" file of Italy, splitting it and putting the 4 files (with correct numbering) in the next level.
At the end it was all OK, but in my trials I forgot to convert a "jpg" file to "dds" : what I saw in Celestia was really strange.
First of all, I saw (correctly) a white rectangle, but zooming onto it I saw couples of previous parts of the maps of previous levels (perhaps coming from buffers in my graphic card's memory).
So, as a rule of thumb, please verify all the files you handle.

Good bye
Pierluigi

Eeekk! Chris, It's a bit more complex than I first reported, I've been digging down ! and a pattern emerges :--

Levels 0 and 1 are painted correctly but
a tile position at level2 gets its texture from the same tile number but from folder level1
a tile position at level3 gets its texture from the same tile number but from folder level2
and so on :-
4 comes from 2
5 from 3
6 from 3
7 from 4
8 from 4
9 from 5
I couldnt go to level10, when I made a level10 folder Celestia crashed on deep zoom probably when it tried to call up the next resolution, it happened both at ctx resolutions 64 and at 512 1.3.2pre5, WinME.128MBram, Celestia reports :-Vendor: NVIDIA Corporation, Renderer: GeForce2 GTS/AGP/SSE, Version: 1.2.1, Max simultaneous textures: 2, Max texture size: 2048

BuzzThanks for the info but no, all my test tiles are 512x512 x24bit colour.

Pierluigi said "as a rule of thumb, please verify all the files you handle"
Yes very true, I agree, I have BaseSplit0 working OK, but for these BaseSplit1 tests I am using test tiles-files-jpgs of a simple 512x512 plain various colours with text inlaid of the form Z_x_y
where Z is the level number and x_y the tile number. So I hope all is well !

Anyhows, Chris has already cured one mis-sequencing-numbering problem, so I'm hopefull that he will suss this one out in the fullness of time when he gets a chance

Malcolm.

Chris any news on investigations into BaseSplit1 oddities ?
no worries, just curious,
Malcolm.