Mischov

Various improvements will need to be made to `Meeseeks.Parser` before it can be made public.

I have at least partially confirmed high memory use for large HTML files. ### Summary The word size of the result of parsing HTML can be significantly larger than the...

Two aspects of OTP 21 have combined to make the situation better, though still not ideal. 1. As @michalmuskala alluded to, static maps are smaller in OTP 21, with the...

Thank you for the report. I believe I've confirmed what you said ``` iex(15)> IEx.pry eheap_alloc: Cannot allocate 47021480096 bytes of memory (of type "heap_frag"). ``` Do you believe that...

Further exploration suggests that if I call `IEx.pry` after I parse the document (step 4 above) it works fine, and takes the total memory usage from `107190608` to `156387424` (it's...

Yep, that's it. Running a `Task.async(fn -> Enum.count(r) end)` after step 7 above causes the same "cannot allocate" crash. Looks like when copying the document in each result they each...

Thank you! I was just in the process of writing a similar `List.duplicate` test to try and understand what was going on. That's actually a bit surprising- I incorrectly assumed...

I have always been resistant to returning the document from Rustler as a NIF resource instead of as Elixir data, since I appreciated the freedom that came from being able...

I definitely need to document the issue and provide warnings in the appropriate places. It's a pretty bad problem, but I'm not entirely surprised that it took so long to...

I'll leave the broader question of "can the library be fixed to handle this situation?" to Hans, but- > Can I get around this problem Yeah, to some definition of...