Eric Biggers

It is very annoying writing intrinsics code with MSVC. See all the limitations of MSVC that have been discussed above. Additionally, MSVC is closed source and only available on one,...

Visual Studio's clang-cl is better than MSVC, but when clang is operating in MSVC compatibility mode it still inherits some of the limitations of MSVC. Therefore the code that uses...

> Are you using my changes for msvc (you need to port my msbuild stuff to whatever you’re using) Note that the master branch is using CMake now, so it...

I'd like to keep this issue open so that people don't open duplicate issues. I also need to investigate @pps83's observation "In general, I haven't seen issues with msvc building...

This is basically done now, with a few exceptions. The biggest one, at least in theory, is that BMI2 code generation is not yet enabled in the decompressor. With gcc...

At level 9 the bottleneck is usually the inner loop in `hc_matchfinder_longest_match()`. This is one very specific thing, so it is plausible that a CPU that is faster in general...

It doesn't really make sense to talk about just the `load_u32_unaligned` part of `hc_matchfinder_longest_match`, since `load_u32_unaligned` is always inlined and compiles down to a single load. The fact that you...

There are several edge cases where for performance reasons, libdeflate is intentionally more accepting than zlib, **in a safe way**. The specific case that your example triggers is the case...

> Hi, @ebiggers , thanks for your explantion. Generally, any corruption of the compressed data should be timely notified to the users because it may lead to severe effects and...

https://github.com/ebiggers/libdeflate/commit/65da37643ab5f5f3e4e8a9b9e91bd018ebd7f5ba handled `poc1` and `poc4`. https://github.com/ebiggers/libdeflate/commit/2574818ea1741fcb83c2ac918b1567e185ec163a would handle `poc2` and `poc3`, but it would be a more complex change. I'm not sure it's worth merging, given that the existing behavior...