Update 1.105 now available

Blitz3D Forums/Blitz3D Programming/Update 1.105 now available

marksibly

(Posted 2010) [#1]

Hi,

A new update for blitz3d is now available from the product updates section of the accout page.

This is an attempt to fix the crashes people have been experiencing with 'ordered' sprites + nvidia cards + windows7.

The bug does appear to be in the nvidia driver, not in blitz3d so I'm not 100% sure this will fix it for everyone but it's worth a shot!

puki	(Posted 2010) [#2]

I assume that only Win7 people need to download this - and only those that have a problem in the first place.

This could get confusing in the future.

What happens if someone installs this on a XP/Vista system, or a XP/Vista/Win7 machine that has an ATI card - will it break anything?

Matty

(Posted 2010) [#3]

puki - I'd imagine anyone who wants their games to work( using sprites) on a windows 7 machine should apply the patch?

GIB3D

(Posted 2010) [#4]

Yay, it broke FastExt again... lol. No biggy though, I always have 1.104.

MikhailV

(Posted 2010) [#5]

@GIB3D Ohh, FastExt for B3D 1.105 in progress :)

MadJack

(Posted 2010) [#6]

Good stuff. I had to put out a patch about a month ago (for Tank Universal) related to W7 not liking Blitz3d sprites - particularly sprites with entityorder applied. Replacing them all with quads seemed to do the trick.

Also, now that I've left Blitz3d in favour of Blitzmax + Xors3d, I feel kind of nostalgic for good old B3d. Sniff.

Ross C

(Posted 2010) [#7]

OT - how is xors3d for you?

MadJack

(Posted 2010) [#8]

Ross
I'm getting good results. I was hoping for a Blitz3d DX9 version and this is it essentially (although I'm using Bmax now). Has access to shadows, shaders, post processing fx using the Blitz3d commandset. Base requirements are not as high as the Leadwerks engine.

There were a lot of bugs on release but the Xors team have been assiduous in knocking those on the head and at this point, it's stable.

_PJ_	(Posted 2010) [#9]

puki - I'd imagine anyone who wants their games to work( using sprites) on a windows 7 machine should apply the patch?

That would be the best option,
In theory, the patch should only address the issue for those users and have no other 'side-effects'. If something in the fix is incompatible or breaks when used on WinXP or whatever drivers or so, then that'd probably mean 1.06 :D
I intend to update B3D as a matter of course, so will note any new bugs on the bug forum, though I don't expect to see anything different whatsoever, being a Windows XP user with ATI card!

Ross C

(Posted 2010) [#10]

Thanks MadJack. Sounds good!

Dodo	(Posted 2010) [#11]

I use Xors too (with Blitz3d), its pretty stable and has come a long way. I will be converting the headers to PowerBasic might need to hire someone to code some shaders for me in the future ... any good shader programmers on this forum?

SabataRH

(Posted 2010) [#12]

This update fixed my AstroMiner game under Window 7, which has been broke since win7 was released despite several users on this forum swearing there wasn't anything wrong with b3d and win7... know it alls... you know ...
... thanks Mark.

ShadowTurtle

(Posted 2010) [#13]

ok blitz3d is realy old. So it has still great support. This is a great product feature.

one question i have to go: is it harder to make blitz3d directx9/10 compatible as making a "new" crossplattform library/compiler?
You can earn many money if you do a little "port" to dx10/dx11 ^^

Thanks mark for the support all this time.

SLotman

(Posted 2010) [#14]

Something isn't right... even on XP :(

Sprites that were fine, now sometimes disappear one frame to appear on the next, specially when moving (If player isn't moving the 'shadow sprite' remains correctly on screen). Reverting to 1.104 is fine on XP, couldn't test on 7 yet.

The code I use in my game to create the problematic sprite:

	p\ballpos=CreatePivot()

	p\Shadow=LoadSprite("Shadow.png",1+2)
    	ScaleSprite p\Shadow,30,30
	EntityAlpha p\Shadow,0.5
	
	SpriteViewMode p\Shadow,2 ;free moving sprite
	PositionEntity p\Shadow, EntityX(p\ballpos), EntityY(p\ballpos)+19, EntityZ(p\ballpos)
	RotateEntity p\Shadow, 90,0,0,1

Then, after updating 'ballpos' position, I do this:

	PositionEntity p\Shadow, EntityX(p\ballpos)-10,EntityY(p\ballpos)-19, EntityZ(p\ballpos)

.

It looks like this 'shadow' sprite sometimes is rotating around it's position (?) and moving away from the pivot... very strange.

This should show better the problem:

SLotman

(Posted 2010) [#15]

Can anyone confirm the problem I found? Or are sprites working normally for you guys?

bye	(Posted 2010) [#16]

What exactly is the crash/problem with B3D/Win7? I've since moved on from B3D but might need to update some older projects..

From what I can gather, it crashes with sprites but not quads, and only when they use the z-buffer, and only on Win7 with Nvidia.. but the fix now crashes WinXP?

wtf... can someone clarify this spaghetti? Does B3D work on Win7 or not?.. does the fix work on XP or not?

MadJack

(Posted 2010) [#17]

Cicero

Can't clarify regards the new patch, but my game (Tank Universal) was experiencing bad glitching where B3d native sprites were being used. If those sprites had an entityorder attached, this often led to a BSOD. This appeared to be under Nvidia only.

Replacing B3d sprites with quads appears to have solved the issue. I've not had reports of other glitches so far (fingers crossed).

bye	(Posted 2010) [#18]

@MadJack,

Thanks for the info. Seems like a pretty serious issue to me. I wonder if Mark can shed some more light on it.

xlsior

(Posted 2010) [#19]

Does B3D work on Win7 or not?

Most B3D programs work on most computers -- but some don't.

I realize that that's not really helpful, but that's the way it is. :-?

skidracer

(Posted 2010) [#20]

Could people please stop posting noise in this thread (and for that matter kicking the B in other threads).

For several months nVidia 64 bit drivers have suffered a significant vertex memory lock issue with Blitz3D/SDK sprites (while their 32 bit drivers continue to perform flawlessly).

BRL does need the support of the community at this time, until now 64 bit Windows has "just worked", but with the arrival of this nVidia driver issue the testing and support of Blitz3D in future promises to be a slightly more arduous task.

bye	(Posted 2010) [#21]

Could people please stop posting noise in this thread...

If B3D users are concerned about a bug, it's hardly noise. The info about this issue is all over the place, I'm sure people just want a clear understanding of what the bug is, exactly when it happens etc.. so they can implement workarounds (such as using quads)

... (and for that matter kicking the B in other threads)

I've been a long-time supporter and user of B3D for years. It's only recently I've started to look at other languages because it seems BRL's focus is now firmly on BMax users. From this side of the fence, that's how it looks, so it's not so much 'kicking' the language, but honest exasperation at the lack of info on what seems like a very serious bug.

For several months nVidia 64 bit drivers have suffered a significant vertex memory lock issue with Blitz3D/SDK sprites (while their 32 bit drivers continue to perform flawlessly).

That's the kind of information I was asking for. Thanks.

No offense intended, I'm just a naturally blunt person with genuine concerns.

SLotman

(Posted 2010) [#22]

Regarding the bug I am experiencing, when I replace "CreateSprite" with the function below everything works as expected:

Function CreateSpriteFixed(x#,y#,z#, width#, height#)
Local quadmesh

    quadmesh=CreateMesh(parent)
    v=CreateSurface(quadmesh)
    AddVertex v, x-width#, y+height#, z#, 0,0 
    AddVertex v, x+width#, y+height#, z#, 1,0
    AddVertex v, x-width#, y-height#, z#, 0,1 
    AddVertex v, x+width#, y-height#, z#, 1,1

    AddTriangle(v,0,1,2)
    AddTriangle(v,2,1,3)

    Return quadmesh
End Function

So it's definitively a problem on 1.105 :(
Unfortunately, I can't seem to reproduce it outside of my game... so I really don't know what's happening.

jfk EO-11110

(Posted 2010) [#23]

I think it's the task of nvidia programmers, to implement a proper DX7 handler. So send them mails and wait for an update. If they don't fix it, consider nvidia a product that simply doesn't work.

Neochrome

(Posted 2010) [#24]

RESOLVED! i am running win7 Ultimate 64 and the bug killed my development!

i am using a psuedo sprite engine, and this works lovely!
i have nvidia, but this problem has seriously made me look towards the ATi for the furture :(

anyone had any issuse with Win7 Dx7 support with ATi cards?

(sorry for poor spelling, just back from a MAJOR long shift)

Leon Drake

(Posted 2010) [#25]

My ATI card works fine for sprites in b3d on win 7, i honestly have yet to buy another nvidia card after one i purchased a while back caught my system on fire due to a faulty fan.

Neochrome

(Posted 2010) [#26]

Ah, bad luck man,
I am using at the moment NVidia 9600GT (prettly old) but without the update, it worked for a bit and just suddenly went very very slow.

Patched up with 1.105.. works like a dream... tho MAJORLY glitchy on BSP maps and ZOrdering :( sort of put my iTanks game in the pot again...

bye	(Posted 2010) [#27]

Another serious win7 bug; just got a bug report about win7+ATI card issue on all versions (1.05 and lower) that doesn't display anything on sprites OR quads when you texture them with an animtexture (e.g:- a bitmapfont texture).

Tried several fixes to no avail, very very frustrating when projects have to be abandoned.

I wouldn't mind, as my new projects are not written in B3D, but years of work on other stuff is now worthless garbage.

I won't hold my breath for a fix.

LineOf7s

(Posted 2010) [#28]

Not if you piggyback your complaint on the back of a 3-week-old update announcement thread.

I guess it makes it easier to ensure there's something to complain about, though.

bye	(Posted 2010) [#29]

Not if you piggyback your complaint on the back of a 3-week-old update announcement thread.

If you say so. I'm sure nobody wants me to start a new thread for every win7 bug, especially since this was already discussed earlier in the thread if you care to read it.

I guess it makes it easier to ensure there's something to complain about, though.

The idea that I'd take the time out to complain about a bug with no intention of if being seen is just silly. I won't even dignify it beyond that.

LineOf7s

(Posted 2010) [#30]

Look, I have nothing against you, but in the same post in this "Update 1.105 now available" thread you mentioned "another serious win7 bug" and "I won't hold my breath for a fix". Well duh. You're only a half-step away from posting it in the (picking a random thread) "My Games are on Steam!" thread as far as obfuscation of your "bug report" goes.

So...
1) If you indeed do not have "no intention of it being seen" then perhaps posting it into the Blitz3D Bug Reports forum might be an idea, following these guidelines when you do.
2) Yes, actually, if you've indeed found "another serious win7 bug" then I'm sure people would be very grateful for you to start a new thread - in the correct forum - in order to give it the best chance of being fixed. If you care enough to do so, though.

But to tack your 'report' on as a passing reference in a barely-relevant thread only to finish up with a snide remark about not holding your breath for a fix is little more than the act of a curmudgeon, and it's to that I reacted.

bye	(Posted 2010) [#31]

I appreciate the tutorial, but perhaps your time would be better spent on something else. I don't want to give you the impression I'm interested in your take on forum etiquette, but it was nice of you to spend so much time and effort trying to derail and obfuscate my post.

LineOf7s

(Posted 2010) [#32]

I appreciate the depths of your indifference.

RyanH

(Posted 2010) [#33]

I have to agree with LineOf7s. No offense to anyone here, but I really don't understand why bug reports are getting posted here, where they may or may not even be read, instead of in the bug reports forum.

bye	(Posted 2010) [#34]

It's more of a complaint than a bug report, since I don't expect any action will ever be taken. Can we please refrain from telling each-other where, when and how to post, it's distracting and annoying.

If anyone else feels like debating this for hours, come to IRC and do it, instead of flooding threads with repetitive "I agree with x, you should post like this" (irc.blitzed.org #blitzbasic)

Thanks!

Neochrome

(Posted 2010) [#35]

i just want the fix :(

i dont know enough of the internals to Blitz to offer a fix my self, but i am just hopefull that this thread is noticed by mark.

Doggie

(Posted 2010) [#36]

Ha, there was a time, long ago, when I thought I'd take up computer programming to relieve the stress of modern life. Sweet dreams are made of these...

DOG

Blitzplotter

(Posted 2010) [#37]

Thanks for the continued updates Mark. I love BLitz3D as a development language, I recently bought a radio control plane emulator where you connect a transmitter to control various planes. This is entriely possible with Blitz3D & the physics libs that are available.

And, despite the small amount of negativity now and then, the posts help identify more fixes (;-)