m4b
I think this can easily be fixed by simply calling default and requiring anyone to also derive Copy and Default on their types, which kind of sucks? E.g., in your...
I think the best way to fix this is to use `MaybeUninit`; if anyone feels like making a PR that would be appreciated :)
Hi @roblabla thanks for opening the issue! So: 1. I don't like the function name per say, **but** I like the idea! 2. It would also be nice to cleanup...
Ok, but just fyi, I don't really want external deps in scroll; I know that's super annoying, but I'd like to keep it dependency free in general. That being said...
please don't assert if size is 0. just return an allocation error if possible, or at least do a checked sub and unwrap so it doesn't do weird stuff in...
ok, agreed assert better than debug overflow unwrap; even better it's caught even farther upstream, but yea. Is creating a buffer of size 0 in vulkan really a validation error?...
If it’s invalid usage for size 0 simple panic with error message is best imho, and then documenting this as a panics condition :) Asserts are for pre and post...
Out of bounds panics with bounds len violation message, last I checked. I’m sure there are asserts in vec but I can’t remember last time an assert violation bubbled up...
In any event I mostly care about fixing the debug overflow which wasn’t super helpful wrt tracking down cause, and I don’t really care if it’s a panic or assert,...
Wouldn't I have to do this in the generated code directly, otherwise this violates coherency, no?