|
Re: Shaderjoy 1.16 |
Posted on: 2020/10/25 7:33
#321 |
---|---|---|
Just can't stay away
![]() ![]() Joined:
2007/7/14 21:30 From Lothric
Posts: 1205
|
@kas1e
I'm expecting the application to beep and exit normally regardless of the scrollbar position. For me, the test application is blocked when the scrollbar is up. Yes, I saw Trevor's presentation, pretty cool :) Shaderjoy was also featured on some Hans' Youtube video earlier. EDIT: added some video links in the first post. |
|
|
Re: Shaderjoy 1.16 |
Posted on: 2020/10/25 7:42
#322 |
---|---|---|
Home away from home
![]() ![]() Joined:
2007/9/11 12:31 From Russia
Posts: 6721
|
@Capehill
Quote:
Yeah, same for me, until it at the top it blocked. Then once I scroll it down its exit with a beep (with no corruption). Should we discuss it with TonyW maybe? In our case, it's anyway a bit different... Maybe we need to set there some interrupt handler like if console output close we then "interrupt/break" any further output or something, so control to ShaderJoy is back? |
|
|
Re: Shaderjoy 1.16 |
Posted on: 2020/10/25 7:49
#323 |
---|---|---|
Just can't stay away
![]() ![]() Joined:
2007/7/14 21:30 From Lothric
Posts: 1205
|
@kas1e
Yeah, I could ask Tony about this. So you don't have that last line corruption / overwite issue? Have you modified your console settings somehow, line count or such? I wonder about that output window closing part. Theoretically output window should re-open when new prints appear. |
|
|
Re: Shaderjoy 1.16 |
Posted on: 2020/10/28 14:54
#324 |
---|---|---|
Home away from home
![]() ![]() Joined:
2007/9/11 12:31 From Russia
Posts: 6721
|
@Capehill
Btw, shouldn't be uniform "iChannelTime" be declared in ShaderJoy as well? I find out those shaders which failed with an error about "iChannelTime" undeclared, and for such shaders, I had to add manually at top "uniform vec3 iChannelTime;" There is a list of those shaders for tests, in case you think such a uniform should be added to ShaderJoy: Quote:
|
|
|
Re: Shaderjoy 1.16 |
Posted on: 2020/10/28 16:34
#325 |
---|---|---|
Just can't stay away
![]() ![]() Joined:
2007/7/14 21:30 From Lothric
Posts: 1205
|
@kas1e
iChannelTime is waiting for audio support. Shaders that try to sample audio would work wrongly anyway. Anyway, that's a handy list for future testing, thanks! |
|
|
Re: Shaderjoy 1.16 |
Posted on: 2020/10/29 17:16
#326 |
---|---|---|
Just can't stay away
![]() ![]() Joined:
2007/7/14 21:30 From Lothric
Posts: 1205
|
@kas1e
Quote:
I'm now able to test this and it works for me. If you can figure out how to break I can check it again. |
|
|
Re: Shaderjoy 1.16 |
Posted on: 2020/10/29 21:10
#327 |
---|---|---|
Home away from home
![]() ![]() Joined:
2007/9/11 12:31 From Russia
Posts: 6721
|
@Capehill
You mean all 3 stages works ? For me only first 2 works as expected, and then when 3rd stage starts, it show only background, but no fuzz/blur of texture. I.e there should be no of "only background" at any pass |
|
|
Re: Shaderjoy 1.16 |
Posted on: 2020/10/30 16:42
#328 |
---|---|---|
Just can't stay away
![]() ![]() Joined:
2007/7/14 21:30 From Lothric
Posts: 1205
|
@kas1e
Yes. Can you switch to Detail debug level and then just quit the Shaderjoy and paste the final logs? |
|
|
Re: Shaderjoy 1.16 |
Posted on: 2020/10/30 17:21
#329 |
---|---|---|
Home away from home
![]() ![]() Joined:
2007/9/11 12:31 From Russia
Posts: 6721
|
@Capehil
Will check once will back home, but that about 7-10 days :( |
|
|
Re: Shaderjoy 1.17 |
Posted on: 2020/10/31 11:53
#330 |
---|---|---|
Just can't stay away
![]() ![]() Joined:
2007/7/14 21:30 From Lothric
Posts: 1205
|
1.17 is available:
- Add FLOATTEXTURE tooltype. Note: there is an open driver issue #610. - Fix mouse Y position (info bar height wasn't taken into account). - Change iFrame uniform type from float to int. - Load dropped shader file into Image slot when there is no editor view (tabs). - Add sobel.frag. - Refactor logging. |
|
|
Re: Shaderjoy 1.17 |
Posted on: 2020/11/1 11:46
#331 |
---|---|---|
Home away from home
![]() ![]() Joined:
2007/9/11 12:31 From Russia
Posts: 6721
|
@Capehill
Btw, while away from amiga HW at the moment, still download the 1.17 archive and find out in directory ShaderToy a file with name "new_common.txt", with such content: Quote:
That just some test-leftovers?:) Or used for something now? Also if doesn't mind, can clean a bit about logging refactoring? Maybe it's something about an issue which we discuss lately related to that closing-console/can't close ShaderJoy? |
|
|
Re: Shaderjoy 1.17 |
Posted on: 2020/11/1 11:59
#332 |
---|---|---|
Just can't stay away
![]() ![]() Joined:
2007/7/14 21:30 From Lothric
Posts: 1205
|
@kas1e
new_common.txt should be used when one wants to add new Common tab. It could be internal string too but this way it's easier to customize. Oh, and log refactoring was only about going through all prints, and start using new logging levels (earlier it was only 2 levels). I will need to get back to that console issue some day. |
|
|
Re: Shaderjoy 1.17 |
Posted on: 2020/11/1 12:05
#333 |
---|---|---|
Home away from home
![]() ![]() Joined:
2007/9/11 12:31 From Russia
Posts: 6721
|
@Capehill
Quote:
But common part always different in all shaders (at least with shadertoy), some of them just had a bunch of defines, without functions, etc. Or you mean that just for an easer way to made own shader than using shadertoy ones ?:) But then it placed in "ShaderToy" directory, meaning it for "ShaderToy"? |
|
|
Re: Shaderjoy 1.17 |
Posted on: 12/4 16:20
#334 |
---|---|---|
Home away from home
![]() ![]() Joined:
2007/9/11 12:31 From Russia
Posts: 6721
|
@Capehill
What version do you set inside of shaderjoy for all the shaders? 120 or 130? Or none and let ogles2 use the default one? Why I ask, because there is bunch of shaders which fail because of those errors: 'min' no matching overloaded function found: 'max' no matching overloaded function found: 'rot' no matching overloaded function found: 'round' no matching overloaded function found: That all happens because of lower versioning set in shaderjoy internally (or not set ?), and those ones are from "#version 150". So the question is what kind of version you have set now, and can you raise it to #version 150 as default (because why it should be lower anyway?) There is a list of shaders you may want to try to test with the current version in shaderjoy, and if you will set internally #verison 150 : 'rot' no matching overloaded function found: https://www.shadertoy.com/view/3dVyD3 'max' no matching overloaded function found: https://www.shadertoy.com/view/llj3Wy https://www.shadertoy.com/view/4tdSDX 'min' no matching overloaded function found: https://www.shadertoy.com/view/tsdfW8 https://www.shadertoy.com/view/WdKcWd https://www.shadertoy.com/view/WsVcRd https://www.shadertoy.com/view/wlj3zV https://www.shadertoy.com/view/tlSSDV https://www.shadertoy.com/view/WlXyzS https://www.shadertoy.com/view/wtsyzs Thanks for worry :) |
|
|
Re: Shaderjoy 1.17 |
Posted on: 12/4 16:51
#335 |
---|---|---|
Just can't stay away
![]() ![]() Joined:
2007/7/14 21:30 From Lothric
Posts: 1205
|
@kas1e
Shaderjoy is current using "#version 310 es". Are you running version 1.17 by the way? https://www.shadertoy.com/view/llj3Wy has: "max(0,-iFrame)" and iFrame was changed from float to int in 1.17, so it _should_ work now...I will try these shaders on my side. |
|
|
Re: Shaderjoy 1.17 |
Posted on: 12/4 17:41
#336 |
---|---|---|
Home away from home
![]() ![]() Joined:
2007/9/11 12:31 From Russia
Posts: 6721
|
@Capehill
Omg! They all seem to start to works now! I didn't test all failed I found (have in my list about 50 of them with such errors), but so far 10 random ones works! (taking aside those ones which hog CPU at 100%:) ) Some auto-fix for non reported bug, thanks a bunch !:) |
|
|
Re: Shaderjoy 1.17 |
Posted on: 12/5 9:04
#337 |
---|---|---|
Home away from home
![]() ![]() Joined:
2007/9/11 12:31 From Russia
Posts: 6721
|
@Capehill
Rechecked all the shaders which give me errors about "no matching overloaded function found", and for the "min", "max" and "round" they are gone. Through found one shader, which still brings error about "rot": https://www.shadertoy.com/view/3dVyD3 But that one IMHO not worth worry, as it just one single shader from thousands tested. Another also not that big of worry issue, is I found that for some heavy shaders I have such a warning: Warning : 40:1: 'preprocessor evaluation': undefined macro in expression not allowed in es profile HW_PERFOMANCE: That one harmless, things works after it show ups, but still maybe worth to look at. There is test ones with this warning: https://www.shadertoy.com/view/tt2XzG https://www.shadertoy.com/view/4sBGD1 |
|
|
Re: Shaderjoy 1.17 |
Posted on: 12/5 9:13
#338 |
---|---|---|
Just can't stay away
![]() ![]() Joined:
2007/7/14 21:30 From Lothric
Posts: 1205
|
@kas1e
Pumpkin Rave works, audio excluded. You need to load Common part first because rot() is defined there and used by both Buffer A and Image. Regarding preprocessor warnings, I suppose it might be possible to let user define #version. EDIT: "Analyze" button of Shadertoy indicates that the web version is using "#version 300 es". |
|
|
Re: Shaderjoy 1.17 |
Posted on: 12/5 9:54
#339 |
---|---|---|
Home away from home
![]() ![]() Joined:
2007/9/11 12:31 From Russia
Posts: 6721
|
@Capehill
About rot: right, I just tried to include the "common" part to the image, and fail to see that bufferA use it as well, damn. Btw, check this shader: https://www.shadertoy.com/view/lldXRB On our part, we have too many "star dots", while on win32 realization there just very-very few of them. I think it's again related to "smoothstep" maybe ? That part:
// stars
|
|
|
Re: Shaderjoy 1.17 |
Posted on: 12/6 12:09
#340 |
---|---|---|
Not too shy to talk
![]() ![]() Joined:
2015/6/11 9:51 From Cologne
Posts: 435
|
@kas1e
Quote: Warning : 40:1: 'preprocessor evaluation': undefined macro in expression not allowed in es profile HW_PERFOMANCE: This is no bug on our side, neither in ogles2.lib nor in ShaderJoy. The picky GLSL preprocessor simply warns you that the #if condition cannot really be evaluated as the author probably intended because the preprocessor value to be tested doesn't exist in the first place. Note that those two shaders in question were made by the same guy ![]() #ifndef HW_PERFORMANCE #define HW_PERFORMANCE 0 #endif on top of his shaders so that his toggle is actually defined if you as the user didn't add an explicit #define HW_PERFORMANCE X on top by yourself. So, no action to be taken on our side. |
|