Maccy 1.x and towards 2.x

[p0deje]

TLDR
I plan to freeze current Maccy to 1.x version and start re-writing it to utilize newer technologies, which will unfortunately necessitate dropping support for older macOS versions. The new version will maintain the same look and feel as the current version, while being faster, more efficient and have better integration with modern Apple ecosystem.

Hello folks! 👋

Maccy was created 7 years ago as a hobby project where I attempted to learn Swift and macOS development to replace missing Parcellite clipboard manager when I moved from Linux to macOS. Since then, it has been download 600k+ times, got featured on the App Store and used by thousands of people every day. I've never thought it might become so popular. Thank you everyone for using it, it's an honor for me. 🙇

7 years is a lot of time and macOS has changed drastically since then. When I started, there were no SwiftUI, SwiftData and other modern approaches to developing AppKit applications. The way Maccy is built is based on the old NSMenu system that is used to power context menus in macOS. This allowed me to preserver plenty of backwards compatibility - Maccy still runs and works on older macOS 10.14 Mojave which was released in 2018. There are still some people downloading and using it on those versions, though it's just a fraction of usage compared to modern macOS.

The choice of NSMenu provided native UI so that Maccy is fast and feels like part of OS. However, it also caused multiple issues in the long run, some of which were worked around (e.g. support for Korean/Chinese language in search), while others are proven to be hard to work around (e.g. support for Japanese language). In summary, NSMenu were never designed to be used for such use cases and this caused plenty of problems with no straightforward way to work them around. I can elaborate more on these issues if necessary.

With all this technical debt and design limitations, I don't see a clear way forward rather than attempting to rewrite Maccy into a regular macOS application that uses regular windows and custom UI views for handling input events. In the very end, UI must remain the same so it's native and fast, but the underlying implementation must move away from NSMenu.

However, re-writing without using modern stack is quite painful so I'd like for Maccy to adopt all the recent SDKs by Apple, notably:

• SwiftUI for UI views instead of current AppKit;
• SwiftData for persistence instead of current Core Data;
• App Intents for Shortcuts.app integration instead of current Intents.

However, switching to these technologies comes at the price of dropping backwards compatibility. For example, SwiftData requires macOS 14 Sonoma or later. That's unavoidable even though I'd hate to do that. The installations per platform suggest that's acceptable - by the time rewrite is done the 90% of users would be using at least macOS 14. I don't have data for installations via Homebrew/GItHub, but on the App Store the downloads this year so far are the following:

macOS version	Downloads (%)
10.14	<1%
10.15	<1%
11	<1%
12	4%
13	9%
14	87%

With that said, my current plan is to:

1. Make a final release of current Maccy as version 1.0. Keep it supported for a while with bug fixes whenever it's possible.
2. Start working on Maccy 2.0 that uses SwiftUI, SwiftData, App Intents. Preserve current UI look and feel while solving number of issues (memory consumption, Japanese language support).
3. Explore opportunities that open with new technologies (iCloud synchronization, iOS version, integration with new Siri, etc).

I'd be happy to hear what Maccy users think about these plans!

[XInTheDark]

Sounds great! Looking forward to the next version!

[tommyming]

Sound great! As a "maccy-heavy" user, would say it helps me a lot on software development.
But one thing I am not sure about is SwiftData.

Although it works well with new tech stack e.g. SwiftUI, the issues on new Predicate system could be a concern.
Apart from that, I've heard that SwiftData has a huge change in this year's WWDC(I don't have the time to drill on the related sessions yet, sadly).
Not sure how much will it change the behavior of SwiftData 1.0.

[p0deje]

Although it works well with new tech stack e.g. SwiftUI, the issues on new Predicate system could be a concern.

What issues are you referring to? Maccy is not using any complicated data queries so I believe it should be okay.

[tommyming]

Thanks for the reply!

Disclaimer: I didn't dive deep into the code, so my comment is just opinon.

I faced some issues when getting my hands dirty on SwiftData when it was first released last year.
It is slightly different to NSPredicate.

My assumption right now is:
If Maccy is just performing a simple query when checking copy history, then it should be fine.

I can't give a very subtle opinion at this moment. But I agree that it is worth trying.

[edsonbittencourt]

It was love at first sight with Maccy. I would love to and am really excited about the intentions for the application. Congratulations.

[miodragnikac]

Looking forward to the iOS version 🙌

[weisJ]

That’s great news. Let me know if you need any help. I still have cursor positioning in my backlog :D

[CostGranda]

That's amazing news, go ahead! and thanks for the effort