Emi

With [Mach engine](https://machengine.org) we have two dependencies that would be very, very painful to remove (would set us back years): 1. DirectXShaderCompiler (DXC) - a Microsoft fork of LLVM 3.7,...

A positive long-term effect of this change is that it would push us as a community away from wrapping C++ code and towards more pure-Zig solutions. Many of the comments...

Helped by https://github.com/hexops/mach/issues/903 Being tracked historically at https://wrench.machengine.org/stats/

I've been able to reproduce this with a similarly large pre-built static library, on macOS M3, using a slightly older Zig version: ``` 0.12.0-dev.3180+83e578a18 ``` Steps to reproduce with the...

* We aim to support most desktop/laptop hardware that was released in the last 5 years (see also #989) * We will support Android and iOS in the near future,...

FYI in Mach we don't use a sysroot; instead we treat frameworks as just frameworks and just add them to the paths: https://github.com/hexops/xcode-frameworks/blob/fb8ce986a98a72cdcbb844bbcfbafeb3133564a2/build.zig#L18-L22

@kubkon I realize now we're facing this issue as well: * https://github.com/hexops/mach/issues/1009 * https://github.com/hexops/mach/issues/1010 * https://github.com/hexops/mach/issues/1011 Can you shed more light on why this behavior changed? I know this worked...

From my POV options 1/2 make sense, and it sounds like 3 has a runtime penalty which would be nice to avoid. As I understand it, that means the linker...

Thanks for explaining that @menduz - I think the problem with that approach is that using `b.sysroot` prevents you (as expected) from linking other frameworks from other places on a...

Wanted to mention I've added this to Mach's `Most important Zig issues` tracking issue - since currently all our cross-compilation to macOS targets is broken as a result of this...