[feature request] basic logic: for loops, conditional statements, event triggers

Why is it not possible to generate images sequencially in a for loop, or trigger next "batch" on latent image input that would work like infinite image generation?

Currently the only way to do something similar is to set a latent batch the max size your GPU VRAM can handle.

Why can't I have a workflow that runs a set amount of times sequentially or generates images until it's killed?

This prevents running the generation unattended over-night and in my opinion would be a really useful feature.

Another case for conditions is - img2img vs prompt input.

Why cannot I do branching logic in a single workflow that would behave differently based on whether I put a text prompt in or a source image?