Have the frontend and new swift-driver look in an external -sdk for non-Darwin platform runtime libraries and modules too
#79621
+167
−121
Conversation
This file contains hidden or 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
|
@swift-ci please test |
|
swiftlang/swift-driver#1822 |
1 similar comment
|
swiftlang/swift-driver#1822 |
finagolfin
commented
Feb 27, 2025
lib/Frontend/CompilerInvocation.cpp
Outdated
| FrontendOpts.UseSharedResourceFolder ? "swift" : "swift_static", | ||
| getPlatformNameForTriple(Triple)); | ||
| // Check for eg <sdkRoot>/usr/lib/swift/linux/ | ||
| if (llvm::sys::fs::exists(SDKResourcePath)) { |
There was a problem hiding this comment.
I initially tried simply checking if <sdkRoot>/usr/lib/swift/, ie a Swift resource directory, existed, but around 100 tests from the compiler validation suite failed with that, as they pass in a separate -sdk that simply has a Swift resource directory with a few different API notes or something, but does not contain the platform-specific files needed in <sdkRoot>/usr/lib/swift/<os>/. Actually checking for this OS-specific directory makes sure the Swift resource directory in -sdk contains the required platform-specific modules and libraries, and falls back to the Swift resource directory in the toolchain itself otherwise. However, that implies mixing and matching files from two different Swift resource directories, so maybe it will be better to enforce that only one Swift resource directory is used and modify all those tests instead?
For an example of such mixing already taking place, the Windows CI failed with this pull, right before trying to compile the trunk Foundation macros for the Windows host:
-- Check for working Swift compiler: T:/5/bin/swiftc.exe - broken
CMake Error at C:/Program Files/CMake/share/cmake-3.29/Modules/CMakeTestSwiftCompiler.cmake:40 (message):
The Swift compiler
"T:/5/bin/swiftc.exe"
is not able to compile a simple test program.
It fails with the following output:
Change Dir: 'T:/x86_64-unknown-windows-msvc/FoundationMacros/CMakeFiles/CMakeScratch/TryCompile-6u6j6e'
Run Build Command(s): C:/PROGRA~1/MICROS~2/2022/COMMUN~1/Common7/IDE/COMMON~1/MICROS~1/CMake/Ninja/ninja.exe -v cmTC_bc87d
[1/2][ 50%][0.084s] T:\5\bin\swiftc.exe -target x86_64-unknown-windows-msvc -j 36 -num-threads 36 -c -module-name cmTC_bc87d -sdk T:/toolchains/swift-6.0.3-RELEASE-windows10/LocalApp/Programs/Swift/Platforms/6.0.3/Windows.platform/Developer/SDKs/Windows.sdk -gnone -Xlinker /INCREMENTAL:NO -Xlinker /OPT:REF -Xlinker /OPT:ICF -incremental -output-file-map CMakeFiles\cmTC_bc87d.dir\\output-file-map.json T:\x86_64-unknown-windows-msvc\FoundationMacros\CMakeFiles\CMakeScratch\TryCompile-6u6j6e\main.swift
FAILED: CMakeFiles/cmTC_bc87d.dir/main.swift.obj
T:\5\bin\swiftc.exe -target x86_64-unknown-windows-msvc -j 36 -num-threads 36 -c -module-name cmTC_bc87d -sdk T:/toolchains/swift-6.0.3-RELEASE-windows10/LocalApp/Programs/Swift/Platforms/6.0.3/Windows.platform/Developer/SDKs/Windows.sdk -gnone -Xlinker /INCREMENTAL:NO -Xlinker /OPT:REF -Xlinker /OPT:ICF -incremental -output-file-map CMakeFiles\cmTC_bc87d.dir\\output-file-map.json T:\x86_64-unknown-windows-msvc\FoundationMacros\CMakeFiles\CMakeScratch\TryCompile-6u6j6e\main.swift
<unknown>:0: warning: using (deprecated) legacy driver, Swift installation does not contain swift-driver at: 'C:\Users\swift-ci\jenkins\workspace\swift-PR-windows\build\5\bin\swift-driver-new.exe'
<unknown>:0: warning: option '-incremental' is only supported in swift-driver
<unknown>:0: error: module compiled with Swift 6.0.3 cannot be imported by the Swift 6.2 compiler: T:/toolchains/swift-6.0.3-RELEASE-windows10/LocalApp/Programs/Swift/Platforms/6.0.3/Windows.platform/Developer/SDKs/Windows.sdk\usr\lib\swift\windows\Swift.swiftmodule\x86_64-unknown-windows-msvc.swiftmodule
ninja: build stopped: subcommand failed.
The issue appears to be that they're trying to build the trunk Foundation macros with the freshly-built trunk 6.2 Swift compiler, but passing in a Swift 6.0.3 SDK with -sdk T:/toolchains/swift-6.0.3-RELEASE-windows10/LocalApp/Programs/Swift/Platforms/6.0.3/Windows.platform/Developer/SDKs/Windows.sdk to build it. The trunk Swift 6.2 compiler then quietly ignores that 6.0.3 Swift resource directory and likely uses the 6.2 Swift resource directory next to the toolchain instead. This pull, by enforcing that the Swift resource directory in the 6.0.3 -sdk that is passed in is used because <6.0.3-sdkRoot>/usr/lib/swift/windows/ exists, then fails because the 6.0.3 modules cannot be used with the trunk Swift 6.2 compiler.
@weliveindetail, do you know if that's merely a configuration mistake when building the Foundation and Testing macros on the Windows CI, which should be easily fixed, or something more complicated?
finagolfin
commented
Feb 27, 2025
| @@ -2258,6 +2259,18 @@ static bool ParseSearchPathArgs(SearchPathOptions &Opts, ArgList &Args, | |||
|
|
|||
| if (const Arg *A = Args.getLastArg(OPT_resource_dir)) | |||
| Opts.RuntimeResourcePath = A->getValue(); | |||
| else if (!Triple.isOSDarwin() && Args.hasArg(OPT_sdk)) { | |||
There was a problem hiding this comment.
This code explicitly follows what the original C++ Driver has long done when looking for the Swift runtime libraries, swiftrt.o, and a few other files found in the Swift resource directory:
if (const Arg *A = args.getLastArg(options::OPT_resource_dir)) {
StringRef value = A->getValue();
resourceDirPath.append(value.begin(), value.end());
} else if (!getTriple().isOSDarwin() && args.hasArg(options::OPT_sdk)) {
StringRef value = args.getLastArg(options::OPT_sdk)->getValue();
resourceDirPath.append(value.begin(), value.end());
llvm::sys::path::append(resourceDirPath, "usr");
CompilerInvocation::appendSwiftLibDir(resourceDirPath, shared);
} else {
auto programPath = getDriver().getSwiftProgramPath();
CompilerInvocation::computeRuntimeResourcePathFromExecutablePath(
programPath, shared, resourceDirPath);
}
Note how the SDK is only looked in if a non-Darwin -sdk is explicitly specified: Saleem later tried to expand that to Darwin also in #26361, but he may have never got it to work.
That C++ Driver setup now matches this Frontend setup, because the default in both is now to look relative to the compiler, which is done first in this Frontend here, ie usr/bin/../lib/swift/. If a -resource-dir is set, that is given first priority, then a non-Darwin -sdk is given second priority, ie the C++ Driver and the Frontend now match in where they look.
This is important for two reasons:
- The new swift-driver simply queries the Frontend and uses whatever Swift resource directory it uses, so now the new swift-driver finally matches the original C++ Driver's behavior, and piecemeal workarounds like that in [Unix] Go back to only checking the runtime resource path for swiftrt.o swift-driver#1822 can now be eliminated.
- The Frontend will now look in the same Swift resource directory for stdlib/corelibs swiftmodules as the swift-driver is looking for runtime libraries and
swiftrt.o, eliminating any confusion between the two by centralizing the Swift resource directory lookup here. That already found one bug in the Windows CI, see my other code comment.
However, unlike the C++ Driver, my -sdk code below actually checks if the -sdk path contains a Swift resource directory for the platform triple and does not use the -sdk for this if not, falling back to the aforementioned default next to the compiler in that case. This is because an -sdk is not guaranteed to have a Swift resource directory and may have only a C/C++ sysroot.
We should probably tighten this up to require an explicit -sdk to have a Swift resource directory, with the only exception when an explicit -resource-dir is also specified, but I'm open to debate here. The C++ Driver doesn't even check if the non-Darwin -sdk has a Swift resource directory and simply assumes one is there, we can do a bit better than that.
|
Alright, this tiny pull and the resulting swiftlang/swift-driver#1822 cleanup now pass the linux CI, have no effect on Darwin, and found a seeming bug in the Windows CI. I'm going to test this more in comparison to the C++ Driver and examine more tests to see if I can use it to flush out more cross-compilation bugs with an explicit non-Darwin In the meantime, the basic functionality works and is ready for review. I will add tests once the few remaining smaller design choices mentioned above are hammered out. @artemcm, @etcwilde, and @compnerd, please take a look: I'm looking for feedback on both this current patch and the remaining issues and questions in my detailed comments above. |
|
swiftlang/swift-driver#1827 |
|
swiftlang/swift-driver#1827 |
finagolfin
requested review from
zoecarver,
hyp,
egorzhdan,
beccadax,
ian-twilightcoder,
Xazax-hun,
j-hui and
fahadnayyar
as code owners
finagolfin
removed request for
hyp,
Xazax-hun,
egorzhdan,
zoecarver,
ian-twilightcoder and
fahadnayyar
finagolfin
requested review from
compnerd,
xymus and
susmonteiro
as code owners
|
Alright, this is getting close to the finish line, ready for some review and help on the last sticking points. One of the tests consistently fails on macOS, which appears to be a pre-existing Darwin-specific bug that my test changes surfaced, unrelated to the two compiler changes in this pull because they change nothing when targeting Darwin. Specifically, I modify @beccadax, I notice you added some extra shims checking in the ClangImporter before in #23287, which doesn't appear to have anything to do with this bug, but I wonder if you have any suggestions as a result. The Windows CI got pretty far, but fails with a cryptic error when trying to run the XCTest tests: Previously, it failed when building, so I noticed that it was the one build that was special-cased not to use the just-built Windows Swift SDK and I just switched it over to use it and remove the VFS flags it alone was using, only for it to fail when running the tests now. @compnerd, I notice you special-cased this |
Hmm, I don't remember the full details of how the problem surfaced. I am hesitant to say that is the case though as |
|
@compnerd, those are two completely different CI errors, one for mac and another for Windows. There is no connection between them, I just raised them both in the same comment. I'll try isolating the XCTest testing issue alone in a separate pull and see what the error is when switching it to the SDK build without the compiler changes here, ie why you special-cased it previously. |
|
Check if Windows install error is spurious. swiftlang/swift-corelibs-libdispatch#892 @swift-ci smoke test windows |
compnerd
requested changes
Sep 11, 2025
Comment on lines
+27
to
+33
| WARNING(warning_no_resource_dir_in_sdk, none, | ||
| "You passed in an external -sdk without a Swift runtime.\n" | ||
| " Either specify a directory containing the runtime libraries with\n" | ||
| " the -resource-dir flag, or use -sysroot instead to point at a C/C++\n" | ||
| " sysroot alone. Falling back to this path for the Swift runtime modules\n" | ||
| " and libraries:\n" | ||
| " %0", (StringRef)) |
There was a problem hiding this comment.
This feels wrong. I don't think that we should be trying to format with indentation here. Do any other warnings try to do this?
There was a problem hiding this comment.
I thought it read better, but had not finalized it yet. I will look at the other warnings next and match them.
| // Prefer `-sdk` paths. | ||
| if (!searchPathOpts.getSDKPath().empty()) { | ||
| // Prefer `-sdk` paths for Darwin. | ||
| if (triple.isOSDarwin() && !searchPathOpts.getSDKPath().empty()) { |
There was a problem hiding this comment.
Wait, why are we not preferring -sdk paths?
There was a problem hiding this comment.
Once the Frontend change is made for non-Darwin, the RuntimeResourcePath in the very next block has been set to <-sdk>/usr/lib/swift, so this explicit getSDKPath() check is not needed when targeting non-Darwin. 😺
| @@ -1592,7 +1592,7 @@ function Build-CMakeProject { | |||
| # Disable EnC as that introduces padding in the conformance tables | |||
| $SwiftFlags += @("-Xlinker", "/INCREMENTAL:NO") | |||
| # Swift requires COMDAT folding and de-duplication | |||
| $SwiftFlags += @("-Xlinker", "/OPT:REF", "-Xlinker", "/OPT:ICF") | |||
| $SwiftFlags += @("-Xlinker", "/OPT:REF", "-Xlinker", "/OPT:ICF", "-v") | |||
| @@ -2892,7 +2892,8 @@ function Build-Dispatch([Hashtable] $Platform) { | |||
| -Defines @{ | |||
| ENABLE_SWIFT = "YES"; | |||
| dispatch_INSTALL_ARCH_SUBDIR = "YES"; | |||
| } | |||
| BUILT_FIRST_SDK = if ($Platform -eq $KnownPlatforms["WindowsX64"]) { "NO" } else { "YES" }; | |||
There was a problem hiding this comment.
I was planning to change this to USE_PREBUILT_SDK or something like that, name suggestion welcome.
| @@ -2939,6 +2940,7 @@ function Build-Foundation([Hashtable] $Platform) { | |||
| _SwiftFoundationICU_SourceDIR = "$SourceCache\swift-foundation-icu"; | |||
| _SwiftCollections_SourceDIR = "$SourceCache\swift-collections"; | |||
| SwiftFoundation_MACRO = "$(Get-ProjectBinaryCache $BuildPlatform BootstrapFoundationMacros)\bin" | |||
| BUILT_FIRST_SDK = if ($Platform -eq $KnownPlatforms["WindowsX64"]) { "NO" } else { "YES" }; | |||
| @@ -3049,6 +3047,9 @@ function Build-Testing([Hashtable] $Platform) { | |||
| Foundation_DIR = (Get-ProjectCMakeModules $Platform DynamicFoundation); | |||
| SwiftTesting_MACRO = "$(Get-ProjectBinaryCache $BuildPlatform BootstrapTestingMacros)\TestingMacros.dll"; | |||
| SwiftTesting_INSTALL_NESTED_SUBDIR = "YES"; | |||
| CMAKE_Swift_FLAGS = @( | |||
| "-L$(Get-SwiftSDK $Platform.OS)\usr\lib\swift\$($Platform.OS.ToString())" | |||
| ); | |||
There was a problem hiding this comment.
Please retain the lexicographical ordering
| @@ -3668,7 +3669,7 @@ function Test-SourceKitLSP { | |||
| "-Xlinker", "$(Get-ProjectBinaryCache $BuildPlatform SourceKitLSP)\lib\CSourcekitd.lib", | |||
| "-Xswiftc", "-I$SourceCache\sourcekit-lsp\Sources\CCompletionScoring\include", | |||
| "-Xswiftc", "-I$(Get-ProjectBinaryCache $BuildPlatform SourceKitLSP)\swift", | |||
| "-Xlinker", "-L$(Get-ProjectBinaryCache $BuildPlatform SourceKitLSP)\lib" | |||
| "-Xlinker", "-L$(Get-ProjectBinaryCache $BuildPlatform SourceKitLSP)\lib", "-v" | |||
utils/build.ps1
Outdated
| @@ -4122,7 +4123,7 @@ if (-not $IsCrossCompiling) { | |||
| if ($Test -contains "llbuild") { Invoke-BuildStep Test-LLBuild $BuildPlatform } | |||
| if ($Test -contains "swiftpm") { Invoke-BuildStep Test-PackageManager $BuildPlatform } | |||
| if ($Test -contains "swift-format") { Invoke-BuildStep Test-Format $BuildPlatform } | |||
| if ($Test -contains "sourcekit-lsp") { Invoke-BuildStep Test-SourceKitLSP $BuildPlatform} | |||
| # if ($Test -contains "sourcekit-lsp") { Invoke-BuildStep Test-SourceKitLSP $BuildPlatform} | |||
|
@compnerd, thanks for the review, but I am mainly looking for feedback on the two actual compiler changes, one of which you commented on, and the tests right now. This is still a WIP for the Since I've never built Swift for Windows and don't use Windows at all, I think I will need help from you or @Steelskin to debug and fix those last two issues on the Windows CI. Specifically,
Do you or Fabrice have any idea what's going on there? If one of you can try this out locally and debug those two failures, it would be very helpful, as I simply could not get XCTest to dump any useful info on the Windows CI with all the debug methods I tried in swiftlang/swift-corelibs-xctest#512. |
Strip out `Build-XCTest` from `Build-SDK`. This is done against the default SDK now for each architecture after the full SDK is built.
… non-Darwin platform runtime libraries and modules too as done originally in swiftlang#25990 with the legacy C++ Driver, but since lost in the new swift-driver. Only difference is this checks if a Swift resource directory exists in `-sdk` and falls back to the default if not.
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.
as done originally in #25990 with the legacy C++ Driver, but since lost in the new swift-driver
This is another attempt to fix swiftlang/swift-driver#1562 and avoid the various workarounds that have been put in elsewhere to spackle over this root cause.
This worked for me with the compiler validation suite running natively on an Android device, with the exception of one SourceKit test. Let's run it through all the platforms on the CI before explaining fully what it's doing and adding a test to make sure it doesn't regress.