swift: 5.8 -> 5.9.2 #256956
Draft
swift: 5.8 -> 5.9.2 #256956
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
stephank
force-pushed
the
feat/swift-5.9
branch
2 times, most recently
from
85c7c9a
to
8898e5d
Compare
|
I have not had time to look at this at all yet, but figured I would jot down a note that I hit the following error building on |
stephank
force-pushed
the
feat/swift-5.9
branch
from
8898e5d
to
ec0a472
Compare
|
Updated this to Swift 5.9.2. Whatever issue I was having that was causing the compiler to segfault seems to have been fixed between 3.9 and 3.9.2.
This cuts off the actual error, but I got a similar end-of-log on Linux which was caused by the wrapper changes in this PR. That should be fixed now. Haven't yet tested Darwin again, though. Now fighting a weird issue in swift-format: I wonder if this is specific to swift-argument-parser 1.3.0. It's used by other components in the toolchain, but they're still on older versions. |
|
The only error I hit now when building the Swift 5.9.2 derivation on |
stephank
force-pushed
the
feat/swift-5.9
branch
from
ec0a472
to
1c55f04
Compare
|
Fixed swift-format by downgrading the swift-argument-parser dependency to 1.2.3 for now. I'm not sure why Darwin fails like that. |
13 tasks
|
#282851 should fix the Next error: ( |
stephank
force-pushed
the
feat/swift-5.9
branch
from
1c55f04
to
178944d
Compare
|
I think we have to remove Concurrency backdeploy, but that's unfortunate. It doesn't build with Swift 5.9 anymore, and sources were removed in the Swift 5.10 branch: swiftlang/swift@e54b183 Because of this, I'm not sure if Swift code that uses Concurrency can be built to target macOS <12 using our toolchain. I don't see a way to remedy this. If I understand correctly, Xcode relies on old binaries to accomplish this. Without Concurrency backdeploy, the build now completes (aarch64-darwin), but installation fails: I've been trying to debug this a number of times already, but the CMake is dense here. I don't understand why this fails. |
ofborg
bot
added
10.rebuild-darwin-stdenv
10.rebuild-darwin: 501+
10.rebuild-darwin: 5001+
and removed
10.rebuild-darwin: 11-100
labels
We no longer need the swift-driver patch to handle bash fds.
This doesn't build with Swift 5.9 anymore, and sources were removed in the Swift 5.10 branch: swiftlang/swift@e54b183 Because of this, I'm not sure if Swift code that uses Concurrency can be built to target macOS <12 using our toolchain. That's very unfortunate, but I don't see a way to remedy this. If I understand correctly, Xcode relies on old binaries to accomplish this.
stephank
force-pushed
the
feat/swift-5.9
branch
from
178944d
to
e18cfdb
Compare
ofborg
bot
added
10.rebuild-darwin: 11-100
and removed
10.rebuild-darwin: 501+
10.rebuild-darwin: 5001+
labels
1 task
|
I was not able to figure out why the build path for diff --git a/pkgs/development/compilers/swift/compiler/default.nix b/pkgs/development/compilers/swift/compiler/default.nix
index 38c40106388b..c56edcbbeaef 100644
--- a/pkgs/development/compilers/swift/compiler/default.nix
+++ b/pkgs/development/compilers/swift/compiler/default.nix
@@ -519,6 +519,9 @@ in stdenv.mkDerivation {
mv llvm/bin/clang-16{-unwrapped,}
mv swift/bin/swift-frontend{-unwrapped,}
+ substituteInPlace swift/stdlib/public/libexec/swift-backtrace/cmake_install.cmake \
+ --replace "./libexec/swift/macosx/swift-backtrace" "stdlib/public/libexec/swift-backtrace/OSX/x86_64"
+
mkdir $out $lib
# Install clang binaries only. We hide these with the wrapper, so they are
This patch is clearly specific to my macos x86_64 system, but it could easily be generalized if it was even the right path to take. I more just wanted to post this finding in case it helped unblock others. I am going to keep trying to get farther in the meantime (I did hit another failure later on in the process, but nothing that looks too mysterious yet). [EDIT] the next issue was missing the |
Open
|
@mattpolzin you fix seems to work on arm64 as well. I am now trying to tackle the missing cryptokit issue. Also is there way to reuse the cache when changing the nix file. Waiting for rebuilds everytime on my m1 is annoying [EDIT] looking further it feels like we also need to package CryptoKit. just to be sure i will try a linux build as well. if this issue doesnt happen on linux it might be about the apple sdk |
13 tasks
|
If it makes things easier re: CryptoKit, the 12.3 SDK was added in #326406. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description of changes
https://www.swift.org/blog/swift-5.9-released/
Incomplete. Opening a draft PR to indicate I'm working on it, and share progress.
Current state:
installPhasebroken on Darwin.Things done
sandbox = trueset innix.conf? (See Nix manual)nix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD". Note: all changes have to be committed, also see nixpkgs-review usage./result/bin/)