Hi, i have worked on the new project for some time, i think it can work, but it's not finished yet. here is a quick demo of what it would look like.

Thanks Egon,
i found that setUniform(float array) can be used to set boolean uniform too, if the uniform in glsl program is written as 'bool'.

to AeroShark333:
i tried gles3.0 and i am quite sure the app was in es3, but i still can't use bitwise operation (shader doesn't compile if '&' is used).
i just used 3 as the argument when calling the setglesclientversion() method.

i was thinking about more flexibilities of data format.
for example, since 'short' is not possible, i may use 1 int to store 2 short values. i could save some ram in this way.

but now i found a bigger problem:
i can't use bitwise operation in shader(like int&2 ). so i can't use 1 int to store 32 boolean values. and jpct-ae doesn't provide setUniform(string,bool val). i guess i could use 1 int to store 4 booleans by some calculation, but it's not convenient.

can i call GL functions directly so jpct-ae doesn't need to provide all possible signatures of setUniform?

Thanks.

hello Egon,
it seems that there is no setUniform(string,int[] val). but in java doc of glsl shader there is: uniform int blendingmode[4].
is that right?

another question , can there be uniform short array?
another question , can a uniform array be longer than 1024 elements?

thanks.

Great!
i also wished for a game that has auto-generated roads, preferably a rally game, so i would have unlimited stages to drive on. but that project would be too difficult for me.

i wish blitting can be 90 degree rotated and can be subtractive in color.
rotated blitting can be used to mimic the rotation of screen without setting it from Android , and without using different textures.
subtractive color is easier to achieve dark text on light color background, without using different textures. ( i currently use textures with white text on black background and no alpha, additive makes the text look good on dark background)
are these possible?

i think it uses the video encode/decode processor to do the encoding in vram then transfer the data stream to system ram/network/storage, just like recording video from physical camera.

i have an A53@2.3G + mali T830 device and used the screen recorder built in the Android7 system. i get almost 30fps video recording my 3D app. (res 1920x1080)
that means Android can read the screen into system RAM at high rate. but the image was produced in GPU. does that mean there is a faster way than fb.getPixels ?

Thanks.
about the controls, you need to spend some time in the QUICK START and USER CONTROL chapters in the User Manual, just pay attention to the phase TOUCH UI.
or, if you have some teenager kids, let them read the User Manual and learn to fly, they can show it to you soon.

hello everyone,
i released a new demo . one main difference is that , this version has on-screen touch controls, no need for keyboard (but keyboard still works). just find the new file named 2023 08 in the link posted above, if you want to try it.

i might have a method to avoid uploading new mesh every time a vertex is moved or added or deleted etc. - using shader and dynamic uniforms.
would you be interested in making an app like this? you can sell it on google app store. i think it takes only a few months to finish the basic program. i think many people would like to use such an app , for casual modeling. and it's output file can be re-edited by other modeler softwares (3dsmax etc).
i might be able to make an app myself but it would be very limited and non-professional, since it's not my main goal, and i am not a professional programmer.

thanks for your confirmation.
i guess using jpct-ae, one can make a modeler app which is usable but limited in functionalities and lower performance and non-professional looking. tricks must be used to solve some problems. that way, the app's worth is a doubt.
at the meantime, how about windows version of jpct? a modeler made with jpct, designed for tap control (tablet PC), is this easier? is this kind of app worth the efforts?

i wish i could make 3D models on android devices, using tap control only. so that i dont need windows PC and mouse.
i wonder whether it's possible to make an app with basic modeling functionalities, like some of those in 3DSMAX.
one problem i anticipate is that, models are highly dynamic in a modeler app, i've never tried that with JPCT.
another problem is that, there are symbols, lines, translucent objects, etc.. i dont know whether GLES is good enough to handle all those elements.
anyone interested in such app or with technical insight?

i mean, faster than using only gl renderer to render all display ports.

what if gl renderer and software renderer are used to render 2 different display port? not combining the images, but still in 1 process. is this do-able? would it be faster?