I'm happy to share a new macOS app I've been working on - XCCacheAgent, which targets the classic "project build time" problem. Works best for projects that have many Swift packages. Give it a try, folks! Feedback is all welcome! https://t.co/HilZJbmGNT

Quick intro to XCCacheAgent:

Maintaining a mobile device farm for E2E testing isn’t easy - and making it stable, even harder. My new blog post shares my journey in maintaining such farms. Read it here: https://t.co/vzinQWN5xD

any strong reason to switch to Apple container from Docker? have just tried Apple container and it kinda looks like a Docker clone, but in native

An amazing read from Tripadvisor https://t.co/3ekGkYrSlq

Though SPM supports declaring binary targets, producing an xcframework from a Swift package isn't easy. Traditional approaches rely on an Xcode project, which doesn't apply to SPM. Check out my latest blog post for an approach without Xcode projects. https://t.co/q8CZWyn2Ia

I'm excited to announce the release of xccache v1 🎉 This includes a redesigned caching mechanism that now supports partial cache hits, increasing the cache-hit rate in the project. Give it a try, friends! https://t.co/s8uEuBpfqI

Happy to share a side project I'm working on called XCCache - a lightweight build caching tool for Swift packages. Give it a try, folks! I'd love to hear your feedback if any https://t.co/s8uEuBoHBa

Crowdstrike Analysis: It was a NULL pointer from the memory unsafe C++ language. Since I am a professional C++ programmer, let me decode this stack trace dump for you.

@softwarejameson This is a great discussion and show cases of SDUI from various companies with prod and cons analysis

I wrote a new blog post about remotes and authentication in Git. Do u really know your tool? https://t.co/Yzd53jdNgE

Debugging for quite some time, and u realize it's a trailing newline that caused the issue :/

im just gonna say it: these six round tech interviews need to stop

Hats off to the author for such an amazing article. It feels like time travel reading this https://t.co/57UIF1w5gE

Still don’t know how to bypass the “verify it’s you” step from Google Login when running E2E tests on CI. Is there any practice to overcome this security check?

Let say 15% (p=0.85) of the tasks are non-parallel (job preparation, test warm-up), the speed-up is 1/((1-p)+p/N) = 1.7 (when N=2), 2.3 (when N=3). Now we have 5 jobs, it should be around 3.1x gain :)

2 years ago, when we first split our UI test job (taking 40m) into N jobs, we observed around 1.7x & 2.4x speed-up with 2 & 3 jobs respectively. I thought we might need 4-5 jobs in the next 6 months. What's the expected speed-up then? This law suggests a very good guess.

TIL: Amdahl’s law

This kind of analysis has become part of my quarterly work besides iOS-related work. I feel like CI/CD infra for iOS projects is kinda unique in the sense that some problems only exist in the iOS world, but not in other platforms

Yesterday, a team member asked this question in the weekly sync-up: “what (recent) books made you a better engineer” One of mine is “The Mayfield Handbook of Technical and Scientific Writing” https://t.co/wGjVYW8jy0