Kogia-sima

Disabling userland proxy may alleviate this overhead. https://franckpachot.medium.com/high-cpu-usage-in-docker-proxy-with-chatty-database-application-disable-userland-proxy-415ffa064955

I'm experiencing exactly same issue.

I’m sorry for the late reply (I was just doing some experiments at lab). The former error is a duplicate of https://github.com/Kogia-sima/sailfish/issues/33. The progress of this issue is blocked due...

For now `sailfish-compiler` just transforms the input template into pure Rust code, write them into the temporary artifact (which is placed in target/debug/build/sailfish-compiler-\*\*\*/out/templates/ directory), and then call `std::include!` for the...

@d4h0 Thank you for your suggestion. At first, I'd like to recommend using [`include!` macro](https://sailfish.netlify.app/en/syntax/includes/), which allows embedding another template. Also it prevents users from escaping rendering results of child...

@botika I appreciate your tips, though it has nothing to do with this issue. Your experiments examines how compiler can optimize fixed buffer. However this issue is about the indirect...

I found that [enhanced for-loop optimization](https://github.com/Kogia-sima/sailfish/commit/279121b858e1388236c133c6fe7c197d97239f60) could cause significant performance drop (~40%) in big-table benchmark. That makes me feel even more confused. ### Previous compiler output ```rust { __sf_rt::render_text!(__sf_buf, "\n");...

@botika This isn't that simple. Here are the objdump results for `master` and `perf/join-after-optimize` branch master branch ```nasm 00000000000167f0 : 167f0: 55 push rbp 167f1: 41 57 push r15 167f3:...

Note: no significant difference in Teams benchmark

There is some concern about a licence in big table benchmark. We need another example https://choosealicense.com/no-permission/