4027 replies to this topic

[fuzzel]

 

What about rev4300? Does it crash as many times as 4301 does?

 

Not so much,  4300 was more stable. It could crash too, but not while editing. 4300 crashed mostly when you run the table or on exit. But I have not reported those crash since I usually have open a VP9 table to copy something of simply to check something, and I know VPX and VP9 don't go along

 

hmm strange...4301 doesn't have any code changes it's only the updated FreeImage.dll that was changed. Your crash dump looks like it crashed while exiting the player. Unfortunately it doesn't point to something interesting.

 

 

got a crash dump finally
 

Crash report VPX rev4301

============

Process: VPinball beta 10.7.4301.exe

Reason: 0xC0000005 - EXCEPTION_ACCESS_VIOLATION at 0023:00539D82

Attempt to read from 0x00000000

Thread ID: 0xBE0 [3040]

 

Call stack

==========

00539D82 VPinball beta 10.7.4301.exe (0x0030045E 0x0019F36C 0x00000001 0x0019f36c)

004902FF VPinball beta 10.7.4301.exe (0x0019EBB0 0x0030045E 0x0030045E 0x0019d4a4)

00427AC2 VPinball beta 10.7.4301.exe (0x00755648 0x00A708C4 0x00000000 0x0073e170)

 

Environment

===========

Date/time: 26/9/2020, 22:28:30:606

Number of CPUs: 8

Processor type: 586

System: Windows 10 (10.0 19041)

 

Memory status

=============

Total Reserved: 81308K (79M) bytes

Total Commited: 501776K (490M) bytes

Total Free: 3611156K (3526M) bytes

Largest Free: 2095424K (2046M) bytes

 

Registers

=========

EAX=00000000 EBX=031A5FC0 ECX=00000000 EDX=0000000F

ESI=06539190 EDI=00000004 EBP=00000000 ESP=0019EC14 EIP=773B4A75

FLG=00010202 CS=0023 DS=002B SS=002B ES=002B FS=0053 GS=002B

 

 

Mini dump saved successfully.

 

 

the mini dump is 3.8gb
If you need that, it will take me a while to upload it.

What about rev4300? Does it crash as many times as 4301 does?

It sort of does seem so.

I just had it exit on me now.
It for some reason, does not consistently leave a crash txt and dump

It also seems (and it could be an illusion) to not crash so much, with auto save disabled.

 

Kind if ironic, i turned auto save on and set it for every 3 min, in case the editor crashed (and of course it did not exactly consistently auto save LOL)
And now seems that very thing might be making it crash.

I thought maybe it was doing that because i had my parts table open, and maybe it gave it trouble knowing which one to actually auto save
but seems to be doing it with only 1 table open, so i'm not sure?

The call stack of the crash report points to a null pointer in the sound manager while closing the sound manager dialog. I don't understand why there seems to be a null pointer at all but I changed the code to better handle resources maybe this is fixed in the next update.

[toxie]

 

It seems to love imploding on it.

Good.. Can repro with that one.. But no clue yet how it manages to fail in that spot..

[wiesshund]

new crash type

 

Crash report VPX rev4301

============

Process: VPinball beta 10.7.4301.exe

Reason: 0xC0000005 - EXCEPTION_ACCESS_VIOLATION at 0023:005E722F

Attempt to write to 0x073BDB67

Thread ID: 0x2250 [8784]

 

Call stack

==========

005E722F VPinball beta 10.7.4301.exe (0x073BDB67 0x19CDEBB0 0x00000001 0x19cdeb70)

004E5B40 VPinball beta 10.7.4301.exe (0x19CDEBB0 0x00000001 0x19CDEB70 0x34b7b680)

073BDB67  (0x18463B0C 0x1846420F 0xFFFFFFFF8B0C4689 0xffffffff85182444)

468B0CC4  (0x1846420F 0xFFFFFFFF8B0C4689 0xFFFFFFFF85182444 0xffffffff890274c0)

18463B0C  (0xFFFFFFFF8B0C4689 0xFFFFFFFF85182444 0xFFFFFFFF890274C0 0xffffffffc0335f38)

1846420F  (0xFFFFFFFF85182444 0xFFFFFFFF890274C0 0xFFFFFFFFC0335F38 0x0010c25e)

8B0C4689  (0xFFFFFFFF890274C0 0xFFFFFFFFC0335F38 0x0010C25E 0xffffffffcccccccc)

85182444  (0xFFFFFFFFC0335F38 0x0010C25E 0xFFFFFFFFCCCCCCCC 0xffffffffcccccccc)

890274C0  (0x0010C25E 0xFFFFFFFFCCCCCCCC 0xFFFFFFFFCCCCCCCC 0xffffffffcccccccc)

C0335F38  (0xFFFFFFFFCCCCCCCC 0xFFFFFFFFCCCCCCCC 0xFFFFFFFFCCCCCCCC 0x1024448b)

0010C25E  (0xFFFFFFFFCCCCCCCC 0xFFFFFFFFCCCCCCCC 0x1024448B 0x04244c8b)

CCCCCCCC  (0xFFFFFFFFCCCCCCCC 0x1024448B 0x04244C8B 0x7400e883)

CCCCCCCC  (0x1024448B 0x04244C8B 0x7400E883 0x01e8830e)

CCCCCCCC  (0x04244C8B 0x7400E883 0x01E8830E 0x448b1075)

1024448B FreeImage.dll FreeImage_MakeThumbnail + 0x1D766B (0x7400E883 0x01E8830E 0x448B1075 0x41010824)

04244C8B  (0x01E8830E 0x448B1075 0x41010824 0xffffffff8b07eb18)

7400E883  (0x448B1075 0x41010824 0xFFFFFFFF8B07EB18 0xffffffff89082444)

01E8830E VPinball beta 10.7.4301.exe (0x41010824 0xFFFFFFFF8B07EB18 0xFFFFFFFF89082444 0x548b1841)

448B1075  (0xFFFFFFFF8B07EB18 0xFFFFFFFF89082444 0x548B1841 0xffffffffd2851424)

41010824  (0xFFFFFFFF89082444 0x548B1841 0xFFFFFFFFD2851424 0x418b0c74)

8B07EB18  (0x548B1841 0xFFFFFFFFD2851424 0x418B0C74 0xffffffffc7028918)

89082444  (0xFFFFFFFFD2851424 0x418B0C74 0xFFFFFFFFC7028918 0x00000442)

548B1841  (0x418B0C74 0xFFFFFFFFC7028918 0x00000442 0xffffffffc0330000)

D2851424  (0xFFFFFFFFC7028918 0x00000442 0xFFFFFFFFC0330000 0xffffffffcc0014c2)

418B0C74  (0x00000442 0xFFFFFFFFC0330000 0xFFFFFFFFCC0014C2 0xffffffffcccccccc)

C7028918  (0xFFFFFFFFC0330000 0xFFFFFFFFCC0014C2 0xFFFFFFFFCCCCCCCC 0x560cec83)

 

Environment

===========

Date/time: 26/9/2020, 16:32:05:790

Number of CPUs: 8

Processor type: 586

System: Windows 10 (10.0 19041)

 

Memory status

=============

Total Reserved: 122196K (119M) bytes

Total Commited: 2665448K (2602M) bytes

Total Free: 1406596K (1373M) bytes

Largest Free: 523364K (511M) bytes

 

Registers

=========

EAX=00000000 EBX=19CDEBC4 ECX=00000001 EDX=00000001

ESI=19CDEBB0 EDI=073BDB67 EBP=418B0C74 ESP=004E5BA0 EIP=C7028918

FLG=00210202 CS=0023 DS=002B SS=002B ES=002B FS=0053 GS=002B

 

 

Failed to save minidump.

[vogliadicane]

I'm working out some GI lights shape at the moment to cast some shadow, where a post or something else is in the way of the light.

 

As much as I liked this technique for existing at all I now think, it still looks too artificial, because the boders between light and no light are much too harsh.

So my feature request:

 

give the light object a "blur edges" property

 

Of course the icing on the cake would be, to let the blur scale up in dependence of distance to the light source.

"Distance scale blur (1-10)" with 1 being no effect on distance and 10 most possible blur


Just to make sure, I mean only the light's shape.
All in all that would draw us Non-Renderer Table authorth's abilities a bit more to those, who can do the render.

Edited by vogliadicane, 27 September 2020 - 05:27 PM.

[wiesshund]

I'm working out some GI lights shape at the moment to cast some shadow, where a post or something else is in the way of the light.

 

As much as I liked this technique for existing at all I now think, it still looks too artificial, because the boders between light and no light are much too harsh.

So my feature request:

 

give the light object a "blur edges" property

 

Of course the icing on the cake would be, to let the blur scale up in dependence of distance to the light source.

"Distance scale blur (1-10)" with 1 being no effect on distance and 10 most possible blur


Just to make sure, I mean only the light's shape.
All in all that would draw us Non-Renderer Table authorth's abilities a bit more to those, who can do the render.

 

Use a flasher, and a shadow mask?

[fuzzel]

 

new crash type

 

Crash report VPX rev4301

============

Process: VPinball beta 10.7.4301.exe

Reason: 0xC0000005 - EXCEPTION_ACCESS_VIOLATION at 0023:005E722F

Attempt to write to 0x073BDB67

Thread ID: 0x2250 [8784]

 

Call stack

==========

005E722F VPinball beta 10.7.4301.exe (0x073BDB67 0x19CDEBB0 0x00000001 0x19cdeb70)

004E5B40 VPinball beta 10.7.4301.exe (0x19CDEBB0 0x00000001 0x19CDEB70 0x34b7b680)

073BDB67  (0x18463B0C 0x1846420F 0xFFFFFFFF8B0C4689 0xffffffff85182444)

468B0CC4  (0x1846420F 0xFFFFFFFF8B0C4689 0xFFFFFFFF85182444 0xffffffff890274c0)

18463B0C  (0xFFFFFFFF8B0C4689 0xFFFFFFFF85182444 0xFFFFFFFF890274C0 0xffffffffc0335f38)

1846420F  (0xFFFFFFFF85182444 0xFFFFFFFF890274C0 0xFFFFFFFFC0335F38 0x0010c25e)

8B0C4689  (0xFFFFFFFF890274C0 0xFFFFFFFFC0335F38 0x0010C25E 0xffffffffcccccccc)

85182444  (0xFFFFFFFFC0335F38 0x0010C25E 0xFFFFFFFFCCCCCCCC 0xffffffffcccccccc)

890274C0  (0x0010C25E 0xFFFFFFFFCCCCCCCC 0xFFFFFFFFCCCCCCCC 0xffffffffcccccccc)

C0335F38  (0xFFFFFFFFCCCCCCCC 0xFFFFFFFFCCCCCCCC 0xFFFFFFFFCCCCCCCC 0x1024448b)

0010C25E  (0xFFFFFFFFCCCCCCCC 0xFFFFFFFFCCCCCCCC 0x1024448B 0x04244c8b)

CCCCCCCC  (0xFFFFFFFFCCCCCCCC 0x1024448B 0x04244C8B 0x7400e883)

CCCCCCCC  (0x1024448B 0x04244C8B 0x7400E883 0x01e8830e)

CCCCCCCC  (0x04244C8B 0x7400E883 0x01E8830E 0x448b1075)

1024448B FreeImage.dll FreeImage_MakeThumbnail + 0x1D766B (0x7400E883 0x01E8830E 0x448B1075 0x41010824)

04244C8B  (0x01E8830E 0x448B1075 0x41010824 0xffffffff8b07eb18)

7400E883  (0x448B1075 0x41010824 0xFFFFFFFF8B07EB18 0xffffffff89082444)

01E8830E VPinball beta 10.7.4301.exe (0x41010824 0xFFFFFFFF8B07EB18 0xFFFFFFFF89082444 0x548b1841)

448B1075  (0xFFFFFFFF8B07EB18 0xFFFFFFFF89082444 0x548B1841 0xffffffffd2851424)

41010824  (0xFFFFFFFF89082444 0x548B1841 0xFFFFFFFFD2851424 0x418b0c74)

8B07EB18  (0x548B1841 0xFFFFFFFFD2851424 0x418B0C74 0xffffffffc7028918)

89082444  (0xFFFFFFFFD2851424 0x418B0C74 0xFFFFFFFFC7028918 0x00000442)

548B1841  (0x418B0C74 0xFFFFFFFFC7028918 0x00000442 0xffffffffc0330000)

D2851424  (0xFFFFFFFFC7028918 0x00000442 0xFFFFFFFFC0330000 0xffffffffcc0014c2)

418B0C74  (0x00000442 0xFFFFFFFFC0330000 0xFFFFFFFFCC0014C2 0xffffffffcccccccc)

C7028918  (0xFFFFFFFFC0330000 0xFFFFFFFFCC0014C2 0xFFFFFFFFCCCCCCCC 0x560cec83)

 

Environment

===========

Date/time: 26/9/2020, 16:32:05:790

Number of CPUs: 8

Processor type: 586

System: Windows 10 (10.0 19041)

 

Memory status

=============

Total Reserved: 122196K (119M) bytes

Total Commited: 2665448K (2602M) bytes

Total Free: 1406596K (1373M) bytes

Largest Free: 523364K (511M) bytes

 

Registers

=========

EAX=00000000 EBX=19CDEBC4 ECX=00000001 EDX=00000001

ESI=19CDEBB0 EDI=073BDB67 EBP=418B0C74 ESP=004E5BA0 EIP=C7028918

FLG=00210202 CS=0023 DS=002B SS=002B ES=002B FS=0053 GS=002B

 

 

Failed to save minidump.

 

From the call stack it looks like something that was triggered by the new FreeImage.dll. Do you still have so much crashes If you replace the dll with the one that came with rev4300?

[Mitchell]

Disable Lighting only works in the option? I had to use walls drop so it can change off or on. Looks cool.

Edited by Mitchell, 27 September 2020 - 06:44 PM.

[vogliadicane]

 

I'm working out some GI lights shape at the moment to cast some shadow, where a post or something else is in the way of the light.

 

As much as I liked this technique for existing at all I now think, it still looks too artificial, because the boders between light and no light are much too harsh.

So my feature request:

 

give the light object a "blur edges" property

 

Of course the icing on the cake would be, to let the blur scale up in dependence of distance to the light source.

"Distance scale blur (1-10)" with 1 being no effect on distance and 10 most possible blur


Just to make sure, I mean only the light's shape.
All in all that would draw us Non-Renderer Table authorth's abilities a bit more to those, who can do the render.

 

Use a flasher, and a shadow mask?

 

a valid proposal. But why not expand something that's anyways here? Instead of pointing out "hey there's a workaround for your personal problem", when there is a simple fix for many?

[toxie]

 

 

It seems to love imploding on it.

Good.. Can repro with that one.. But no clue yet how it manages to fail in that spot..

 

Stone old bug actually, since the initial open source release of VP. 

[toxie]

I also reverted FreeImage back again to 3.17.0.   :/

I hoped that my custom build of the dll would resolve the problems we saw earlier when going to 3.18.0, but apparently no.  

[wiesshund]

a valid proposal. But why not expand something that's anyways here? Instead of pointing out "hey there's a workaround for your personal problem", when there is a simple fix for many?

 

 

 

 

Well, mostly because this thread is for the current beta, and issues seen in it, so i try not to load down the thread with baggage of stuff from my wishlist
that does not pertain to what they are presently working on.

I'd make a separate thread for a feature proposal.

[vogliadicane]

feature proposals are absolutely common in our beta threads.

[fuzzel]

rev4307 is up:

 

- FreeImage.dll reverted back to previous version

- Fix crash while loading/saving of table resources 

- Fix memory consumption while saving a table

[fourbanks]

rev4307 is up:

 

- FreeImage.dll reverted back to previous version

- Fix crash while loading/saving of table resources 

- Fix memory consumption while saving a table

is it possible for me to keep build 4301 as its stable my end ?

[Thalamus]

@fourbanks : Don't you worry. At some time, freeimage will probably be updated down the line. And, of course, if it works for you - use it. No one is holding you back on that.

[fourbanks]

@fourbanks : Don't you worry. At some time, freeimage will probably be updated down the line. And, of course, if it works for you - use it. No one is holding you back on that.

thank you as always my friend

[toxie]

Yup, i also keep the new dll for my own setup, no problem.

[wiesshund]

Yup, i also keep the new dll for my own setup, no problem.

 

With autosave OFF, i have not had the sound manager or freeimage DLL crash

I actually have not had any crash, and trust me, i tried to crash it
even to the point of copying all table objects, and pasting them about 100 times into the same table.

Lagged like all hell (expected) but would not crash, and even manually saved the abomination successfully

[toxie]

..which is absolutely weird and i have no clue why this works when saving manually

 

But nevertheless, the new build with the huge memory usage removed should work anyways for you, right?

[wiesshund]

..which is absolutely weird and i have no clue why this works when saving manually

 

But nevertheless, the new build with the huge memory usage removed should work anyways for you, right?

 

I have not gotten to try to kill the 4307 build yet.

 

We shall see if i can break it