The problem remported here is not demonstrated by this repository and may no longer exist. After upgrading crates and packages to the latest, except leaving yew pinned to 0.14 the bug was present on March 26th. However, by then yew 0.19 is current and once I upgraded to that, the bug doesn't appear.
This is the smallest demo I could come up with "quickly" that shows what appears to be a leak when using wee-alloc in a project configured per yew-wasm-pack-template. It is the result of me tracking down a memory leak in a client that I had written using std-web and converted to web-sys using yew-wasm-pack-template as a template.
FWIW, from my POV, this same program, but configured to run with wasm-pack did not show the leak, even with it using wee-alloc. I have not looked into what is different between the two environments, because my guess is there are others who would know off the top of their head. I don't mind spending more time trying to find out more about this bug, but I don't want to spend hours of my time doing something someone else can do in seconds.
The source code has wee-alloc set as the global allocator at the end
of src/lib.rs. If you use yarn start:dev and go to
http://localhost:8000 your browser will start using a lot of WASM
memory. If you use Chrome, you can use the Task Manager window to
see how much memory is used. It may take a few tens of seconds, but
you'll probably get an Uncaught (in promise) RuntimeError coming
from the rust_oom, and if you're using Task Manager, you'll see
about 2 GB of memory consumed.
You'll also get around 1,500 lines of "When in the course of human events" on your web page.
If you comment out the #[global_allocator] line, close the previous
run (if it's still open), open a new window or tab and go to
http://localhost:8000 you'll still get a big pause and perhaps your
fan will run (it's a debug build; this same problem manifests in a
release build; it's just that using them requires some sort of server),
but you'll get 1,500 lines of "When in the course...", there will be no
error on the JavaScript Console and the amount of memory you'll be using
once you get to steady state will be on the order of 50 MB.
I realize there's not a lot of info here. I also realize that wee-alloc never shrinks. However, I don't think this is simply a case of wee-alloc never shrinking, because although the log_data constantly grows upward, it only grows to 1,500 lines. I recognize that if every time you allocate a little more memory than you had before, you potentially leave a block that's not going to ever be usable again for that particular data structure (since it continues to grow) but as those left behind pieces get larger and larger, I would think they could be used for smaller objects. But beyond that, since this doesn't happen with wasm-pack, I think something more is going on.
Anyway, I hope the information I'm providing is useful. I'm happy to provide more information and I feel bad submitting PRs for things like this, because I really don't think it's a yew issue, per-se, and wee-alloc isn't enabled by default, but I also don't want to spam a bunch of different repositories with this info, so I figure making a sample program and reporting it once to someone is probably the right thing to do.