[briandw]

True that Xcode needs yet another rebuild from scratch. If they forked it and abandoned the old project file and went with a swift first approach, could work. However adding support for Claude is still a huge win. Could lead the way to making the transition to a sane IDE possible / reasonable. This would require leadership that’s completely absent at the company.

[embedding-shape]

> If they forked it and abandoned the old project file and went with a swift first approach, could work.

Ever attempted this before at a large company and had success with it? I think I can count four times so far in ~15 years where people attempted to rewrite something medium/large-scale from scratch around me, was a success once, although scope was drastically cut at the end so almost a stretch to call it a success.

[briandw]

You are of course correct. It's not likely to succeed. "Could work" doesn't mean high chance of success. I was trying to imply the opposite. It's just that Xcode has so much baggage that the previous attempts have been very compromised.

[jamesfinlayson]

Seen it once - it got done because there was enough will to make it happen but it probably took a decade.

Whether it was successful or not is up for debate. It certainly was nowhere near as performant as the old system but probably 99% of features came across.

[neutronicus]

In this particular case they just need to release a tool that properly generates compile_commands.json and .clangd from a .xcodeproj.

Boom! emacs is the IDE now. Bonuses all around.

[vor_]

Why would being "swift first" affect anything? Why do you assume Xcode isn't already full of Swift code?