[Ep. 53] SDL VSync (SDL_RENDERER_PRESENTVSYNC) | Introduction to SDL2 and SDL3

[@stevphiericardo2790]

Present_vsync is replaced in sdl3, its entirely different how vsync handled now

[@MikeShah]

Indeed, SDL3 still changing. Waiting for official announcement of first release -- hopefully soon!

[@VoidloniXaarii]

Yeayyy! More SDL ! 🎉🎉🎉

[@MikeShah]

😁

[@VoidloniXaarii]

@@MikeShah 🥰

[@jaoschmidt3786]

this was difficult to watch all your playlist is perfect but I think this specific bitmask explanation was weird I think you wanted to see if a specific flag is true automatically: Just make an AND bitwise operation with the flag you want and info.flag. If it doesn't return 0 then the flag is active so you can write: cout

[@MikeShah]

I could have simplified this -- agreed, one of the joys of live coding 😅

[@shaileshdubey7197]

sir what if we use the smart pointer in a class and create a base pointer which hold the reference of derived type..in that case do we need virtual destructor.. smart pointer data type are class type in which they being written.

[@VoidloniXaarii]

I can't wait until the series finally gets to drawing some spheres, cubes or maybe even a loaded .obj model in sdl. Would be so kickass to see 3d objects rendered in both Linux and windows using the same modern c++ code

[@MikeShah]

Some of those topics will be in the OpenGL series :) When SDL3 comes with the 3D API, I will also likely have those topics covered here.

[@VoidloniXaarii]

@@MikeShah so hard to wait! Been wishing for this from the beginning. Any ETA no matter how rough! Having a Linux/Windows independent way of rendering a model using GPU is THE DREAM using modern c++. I would have thought this would be trivial in modern 2023 world yet after many tens of hours of research it seems to me Godot is the closest to easy loading and rendering of a 3d model... But that is not be l naturally modern c++ centric 😭

[@MikeShah]

@@VoidloniXaarii As soon as SDL 3 comes out, I'll be looking at the GPU API :) So this summer around June/July I'd guess we see more. I think we are starting to see more APIs that just keep things in C++ which is indeed very. nice! Keep an eye on the OpenGL series, that's going to start heating up in the next week or so with some new videos.

[@VoidloniXaarii]

Thank you so much! So looking forward to that!

[@只是約翰紐約市]

You SDL2 videos help me a lot! But maybe you could make a video about SDL2 gfx library?

[@MikeShah]

Cheers! I haven't used SDL gfx library unfortunately, but if I do I'll write a tutorial

[@JamesListener]

Oh, no! 9:43 - when you put && there, it is totally wrong. If you wanted a logic result, you should've cast the result into bool. But with logic 'or ' instead of bitwise 'or' you totally change the math here. It is gonna give you 'true' whenever left and right operands are not zeros. So logic 'or' doesnt give you any idea about "does my result contains that flag", it gives you the answer "is my result not zero?" (flag by deafult is not zero). Meanwhile bitwise 'or' (single &) gives you a result which contains '1' in bits which are both '1' in left and right operands. And if one of operands is flag (flag always contains only one '1' bit), then the result is zero (if the other operand doesn't contain flag) or equal to flag (if it contains flag). So you can just check if the result zero, or cast the result to bool. Mostly it is done this way: if (flags & SOME_FLAG) /* do smth when flag is on */; else /*do smth when flag is off*/;

[@MikeShah]

Correct, && and & are different -- I think I was just playing around with it a bit to show where the result comes from.

[@HolographicKode]

That flag explanation is kinda messy :) you should have converted the info.flags into binary and look at the bits that way.