Nathan Moinvaziri
Nathan Moinvaziri
@Dead2 I previously did this but it was not merged properly into the `develop` branch in #1640. https://github.com/zlib-ng/zlib-ng/pull/1640/commits/eb41c8ce9b4d855c0abbdb56e6dd4f5f515009fd
@mtl1979 your PR is missing some parameters I had in my commit.
AFAIK, we don't use tokenless uploads and we always specify the token. So I don't think anything changes wrt forks.
This appears to be complete as your PR is now merged into `develop`.
@eschnett, a regression has been fixed in `develop` branch if you want to try it out.
We could do what we did with `inflate_fast` (in that it uses a template and there is a separate version of it for each instruction set) but do it for...
> Replacing that function pointer directly with quick_insert_string_sse42 makes the overhead go away. What does the overhead equate to in terms of performance?
Your branch changes look interesting.
The following changes have been made in `develop`: * `functable` no longer uses `Z_TLS` * Implemented `-D WITH_RUNTIME_CPU_DETECTION=OFF` which should further reduce any overhead * `inflate_fast` is also now a...