[swift-evolution] A case for postponing ABI stability

Tino Heth 2th at gmx.de
Fri Jan 27 06:26:27 CST 2017


> I don’t disagree with your overall point, but I do want to emphasize that forcing apps to bundle the stdlib and runtime is more than just suboptimal.

Wouldn't it be possible to have several versions of the runtime bundled with the OS? Frameworks on macOS still have a filesystem layout that is build around the idea of having several versions in one bundle.

Also, breaking ABI isn't that uncommon: C++ did so in the past (and imho that language did things that are worse ;-)

I guess the problem is at least partly a phycological one:
ABI stability was supposed to be one major feature of Swift 3, and it's somewhat depressing when big goals are delayed over and over.

Maybe you could "stabilise" the ABI temporarily, and plan for a breaking change in two or three years, when the dust has settled?
Doing so would reduce the pressure, and long-term Mac developers are used to huge breaking changes anyway ;-)
It could also free resources to work on features and ideas that help to realise shortcomings worth to be addressed in a "really stable" ABI.

[have to remember to write something for the "switch to Discourse"-thread — seems this message got lost]
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.swift.org/pipermail/swift-evolution/attachments/20170127/cb7bb32b/attachment.html>


More information about the swift-evolution mailing list