@HSVSphere @BrretHatinson Typically stacks grow slow-ish so the penalty of exceptions is bearable. If a thread's code has large stack allocations then its likely worse deal to be lazy.
In chrome we had to do a hunt for char name[MAX_PATH] because it started to matter when every stack frame had one.
@HSVSphere @BrretHatinson ... handles by assigning an actual page to the virtual address and resuming the execution.
The vmm might do some optimization on top, but regular usermode heaps don't do lazy paging, at least the ones most apps use.
@HSVSphere @BrretHatinson Probably not. Lets nerd this thing. The stack allocation is done by the vmm in kernel during thread creation. It will reserve 8M but the memory is not allocated. Once the thread starts running it will touch it each page causing a page fault exception which the vmm again ..
People with interesting lives do not get offended that they cannot be happy. Happy people are offended that they cannot have interesting lives.
-- Penelope Trunk
@aionescu partly is that the need for a general library is rare. For one-off research sysinternals is fine, for hooking you usually know exactly your target and the loader does the rest.
I got into computers because building stuff to solve my own problems is fun. I got lucky that it is also an in-demand and lucrative field. But to this day, I still like to somewhat irrationally reinvent the wheel with my main/only motivation being that its fun.
@samwhoo True. I mistyped something while setting Limine and it booted only to the panic shell.
And yes, "here is a table with 5 boot-loaders" please set up one of your choice"
The arch linux installer is amazing. It drops you into a console and makes you type 30% of what is needed to build a distro.
You want networking? here is a table comparing 6 networking packages. I'll wait here while you do your research.