[clang] [DependencyScanning] Add ability to scan TU with a buffer input (PR #125111)

https://github.com/benlangmuir approved this pull request. https://github.com/llvm/llvm-project/pull/125111 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [compiler-rt] [llvm] Revert "Reapply "[ORC] Enable JIT support for the compact-unwind frame..." with fixes." (PR #125098)

https://github.com/benlangmuir created https://github.com/llvm/llvm-project/pull/125098 This reverts commit d6524c8dfa37634257050ca71d16e117b802181c. This reverts commit b1bd73700a1fb6f450e0f6f9c405a9c8bde2cae7. This was causing bot failures on Darwin https://green.lab.llvm.org/job/llvm.org/j

[clang] [clang][modules] Separate parsing of modulemaps (PR #119740)

@@ -0,0 +1,141 @@ +//===- ModuleMapFile.h - Parsing and representation -*- C++ -*-===// +// +// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. +// See https://llvm.org/LICENSE.txt for license information. +// SPDX-License-Identifier: Ap

[clang] [clang][modules] Separate parsing of modulemaps (PR #119740)

https://github.com/benlangmuir edited https://github.com/llvm/llvm-project/pull/119740 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [clang][modules] Separate parsing of modulemaps (PR #119740)

@@ -3157,25 +2140,18 @@ bool ModuleMap::parseModuleMapFile(FileEntryRef File, bool IsSystem, assert((!Offset || *Offset <= Buffer->getBufferSize()) && "invalid buffer offset"); - // Parse this module map file. - Lexer L(SourceMgr.getLocForStartOfFile(ID), MMapLan

[clang] [clang][modules] Separate parsing of modulemaps (PR #119740)

@@ -0,0 +1,141 @@ +//===- ModuleMapFile.h - Parsing and representation -*- C++ -*-===// +// +// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. +// See https://llvm.org/LICENSE.txt for license information. +// SPDX-License-Identifier: Ap

[clang] [clang][modules] Separate parsing of modulemaps (PR #119740)

https://github.com/benlangmuir commented: I'm happy with the way this split of the code worked out! > Currently this has no effect other than slightly changing diagnostics. Can you say more about the ordering changes? I understand we can't always emit diagnostics in source order, but it's help

[clang] [clang][modules] Partially revert 48d0eb518 to fix -gmodules output (PR #124003)

https://github.com/benlangmuir created https://github.com/llvm/llvm-project/pull/124003 With the changes in 48d0eb518, the CodeGenOptions used to emit .pcm files with -fmodule-format=obj (-gmodules) were the ones from the original invocation, rather than the ones specifically crafted for outpu

[clang] [Modules] Delay deserialization of preferred_name attribute at r… (PR #122726)

benlangmuir wrote: There's probably a better way, but I just built my first clang normally then built a second one with `-DCMAKE_C_COMPILER` and `-DCMAKE_CXX_COMPILER` pointing to the first one, and `-DLLVM_ENABLE_MODULES=1` to enable modules. https://github.com/llvm/llvm-project/pull/122726 _

[clang] [Modules] Delay deserialization of preferred_name attribute at r… (PR #122726)

benlangmuir wrote: I see that this change has already been reverted by @ilya-biryukov , but FYI before this is re-applied: I am seeing crashes from this change if I attempt to bootstrap a build of clang with modules enabled on Darwin. I haven't yet managed to minimize a test case, but compili

[clang] [clang][modules] Fix local submodule visibility of macros from transitive import (PR #122955)

https://github.com/benlangmuir created https://github.com/llvm/llvm-project/pull/122955 When we mark a module visible, we normally mark all of its non-explicit submodules and other exports as visible. However, when we first enter a submodule we should not make them visible to the submodule its

[clang] [llvm] [Support] Return `LockFileManager` errors right away (PR #130627)

@@ -248,39 +236,14 @@ LockFileManager::LockFileManager(StringRef FileName) // There is a lock file that nobody owns; try to clean it up and get // ownership. -if ((EC = sys::fs::remove(LockFileName))) { - std::string S("failed to remove lockfile "); - S.a

[clang] [clang][modules] Correctly set module map systemness (PR #131940)

https://github.com/benlangmuir approved this pull request. https://github.com/llvm/llvm-project/pull/131940 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [clang][DependencyScanning] Track modules that resolve from sysroot. (PR #130634)

@@ -739,6 +748,12 @@ ModuleDepCollectorPP::handleTopLevelModule(const Module *M) { MDC.ScanInstance.getASTReader()->visitInputFileInfos( *MF, /*IncludeSystem=*/true, [&](const serialization::InputFileInfo &IFI, bool IsSystem) { +if (MD.IsInSysroot) { +

[clang] [clang-tools-extra] [clang][modules] Introduce new `ModuleCache` interface (PR #131193)

https://github.com/benlangmuir approved this pull request. https://github.com/llvm/llvm-project/pull/131193 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [clang][DependencyScanning] Track modules that resolve from sysroot. (PR #130634)

https://github.com/benlangmuir edited https://github.com/llvm/llvm-project/pull/130634 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [clang][DependencyScanning] Track modules that resolve from "shared" locations (PR #130634)

@@ -739,6 +795,12 @@ ModuleDepCollectorPP::handleTopLevelModule(const Module *M) { MDC.ScanInstance.getASTReader()->visitInputFileInfos( *MF, /*IncludeSystem=*/true, [&](const serialization::InputFileInfo &IFI, bool IsSystem) { +if (MD.IsShareable) { +

[clang] [clang][DependencyScanning] Track modules that resolve from "shared" locations (PR #130634)

@@ -157,6 +157,32 @@ static void optimizeCWD(CowCompilerInvocation &BuildInvocation, StringRef CWD) { } } +/// Check a subset of invocation options to determine whether the current +/// context can safely be considered as shareable. +static bool areOptionsInSharedDir(CowCom

[clang] [clang][DependencyScanning] Track modules that resolve from "shared" locations (PR #130634)

https://github.com/benlangmuir approved this pull request. https://github.com/llvm/llvm-project/pull/130634 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [clang] Remove deprecated `FileManager` APIs (PR #132063)

https://github.com/benlangmuir approved this pull request. https://github.com/llvm/llvm-project/pull/132063 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [clang] Do not share ownership of `HeaderSearchOptions` (PR #132984)

@@ -48,7 +48,7 @@ void ASTMergeAction::ExecuteAction() { /*ShouldOwnClient=*/true)); std::unique_ptr Unit = ASTUnit::LoadFromASTFile( ASTFiles[I], CI.getPCHContainerReader(), ASTUnit::LoadEverything, Diags, -CI.getFileSys

[clang] [clang] Do not share ownership of `HeaderSearchOptions` (PR #132984)

@@ -116,7 +116,7 @@ class ASTUnit { std::shared_ptr PP; IntrusiveRefCntPtr Ctx; std::shared_ptr TargetOpts; - std::shared_ptrHSOpts; + std::unique_ptr HSOpts; benlangmuir wrote: Is there still a reason to keep this a poi

[clang] [clang-tools-extra] [clang] Do not share ownership of `PreprocessorOptions` (PR #133467)

https://github.com/benlangmuir approved this pull request. Sounds great, thanks for explaining! https://github.com/llvm/llvm-project/pull/133467 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cf

[clang] [clang] Do not share ownership of `HeaderSearchOptions` (PR #132984)

@@ -569,7 +569,7 @@ CrossTranslationUnitContext::ASTLoader::loadFromDump(StringRef ASTDumpPath) { return ASTUnit::LoadFromASTFile( ASTDumpPath, CI.getPCHContainerOperations()->getRawReader(), ASTUnit::LoadEverything, Diags, CI.getFileSystemOpts(), - CI.getHe

[clang] [clang] Do not share ownership of `HeaderSearchOptions` (PR #132984)

https://github.com/benlangmuir edited https://github.com/llvm/llvm-project/pull/132984 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [clang] Do not share ownership of `HeaderSearchOptions` (PR #132984)

https://github.com/benlangmuir commented: Basically LGTM but there are two clients that I'm not familiar with. https://github.com/llvm/llvm-project/pull/132984 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailm

[clang] [clang-tools-extra] [clang][modules] Introduce new `ModuleCache` interface (PR #131193)

@@ -0,0 +1,44 @@ +//===--===// +// +// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. +// See https://llvm.org/LICENSE.txt for license information. +// SPDX-License-Identifier: Apac

[clang] [clang][DependencyScanning] Track modules that resolve from sysroot. (PR #130634)

https://github.com/benlangmuir commented: I have a similar concern to @jansvoboda11 that we probably need to check the invocation paths as well. > I would think that if ModuleDepCollector only found dependency inputs that > resolve to sysroot locations, the command line for building the spec

[clang] [clang][DependencyScanning] Track modules that resolve from sysroot. (PR #130634)

@@ -835,6 +850,13 @@ void ModuleDepCollectorPP::addAllSubmoduleDeps( }); } +void ModuleDepCollectorPP::addClangModule(const Module *M, const ModuleID ID, benlangmuir wrote: How about something like "addOneModuleDep"? If find it hard to guess from the name

[clang] [clang][DependencyScanning] Track modules that resolve from sysroot. (PR #130634)

https://github.com/benlangmuir edited https://github.com/llvm/llvm-project/pull/130634 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [Support] Return `LockFileManager` errors right away (PR #130627)

@@ -157,42 +157,35 @@ class RemoveUniqueLockFileOnSignal { } // end anonymous namespace -LockFileManager::LockFileManager(StringRef FileName) -{ - this->FileName = FileName; - if (std::error_code EC = sys::fs::make_absolute(this->FileName)) { -std::string S("failed to o

[clang] [clang][DependencyScanning] Track modules that resolve from sysroot. (PR #130634)

@@ -739,6 +748,12 @@ ModuleDepCollectorPP::handleTopLevelModule(const Module *M) { MDC.ScanInstance.getASTReader()->visitInputFileInfos( *MF, /*IncludeSystem=*/true, [&](const serialization::InputFileInfo &IFI, bool IsSystem) { +if (MD.IsInSysroot) { +

[clang] [clang][frontend] Expose `CompilerInstance::cloneForModuleCompile()` (PR #135405)

https://github.com/benlangmuir approved this pull request. https://github.com/llvm/llvm-project/pull/135405 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [Support] Return `LockFileManager` errors right away (PR #130627)

https://github.com/benlangmuir approved this pull request. https://github.com/llvm/llvm-project/pull/130627 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [llvm] [Support] Remove output file checks from `LockFileManager` (PR #130395)

https://github.com/benlangmuir approved this pull request. https://github.com/llvm/llvm-project/pull/130395 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [clang][deps] Remove dependency on llvm targets from lib DependencyScanning (PR #129774)

https://github.com/benlangmuir closed https://github.com/llvm/llvm-project/pull/129774 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [clang][deps] Ensure DiagnosticConsumer::finish is always called (PR #127110)

https://github.com/benlangmuir closed https://github.com/llvm/llvm-project/pull/127110 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [clang][deps] Ensure DiagnosticConsumer::finish is always called (PR #127110)

https://github.com/benlangmuir created https://github.com/llvm/llvm-project/pull/127110 When using the clang dependency scanner with an arbitrary DiagnosticConsumer, it is important that we always call finish(). Previously, if there was an error preventing us from reaching the scanning action,

[clang] [clang] Make serialized diagnostics more reliable (PR #100681)

https://github.com/benlangmuir commented: There's another call to finish() in DependencyScanningWorker that should be removed if we do this. https://github.com/llvm/llvm-project/pull/100681 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https

[clang] [clang][modules] Fix local submodule visibility of macros from transitive import (PR #122955)

https://github.com/benlangmuir closed https://github.com/llvm/llvm-project/pull/122955 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [clang][deps] Propagate the entire service (PR #128959)

@@ -130,11 +130,11 @@ class DependencyScanningWorker { DependencyActionController &Controller, StringRef ModuleName); - bool shouldEagerLoadModules() const { return EagerLoadModules; } - llvm::vfs::FileSys

[clang] [clang][deps] Propagate the entire service (PR #128959)

https://github.com/benlangmuir approved this pull request. https://github.com/llvm/llvm-project/pull/128959 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [clang][modules] Separate parsing of modulemaps (PR #119740)

https://github.com/benlangmuir approved this pull request. https://github.com/llvm/llvm-project/pull/119740 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [clang][deps] Respect `Lexer::cutOffLexing()` (PR #134404)

https://github.com/benlangmuir approved this pull request. https://github.com/llvm/llvm-project/pull/134404 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [clang-tools-extra] [clang] Do not share ownership of `PreprocessorOptions` (PR #133467)

benlangmuir wrote: Are you planning to do the same for LangOpts and HSOpts? What's the ultimate goal here? There's also this comment on `CompilerInvocationBase`: ``` /// ... It keeps individual option objects /// behind reference-counted pointers, which is useful for clients that want to /// ke

[clang] [clang][deps] Avoid unchecked error assertion (PR #134284)

https://github.com/benlangmuir approved this pull request. https://github.com/llvm/llvm-project/pull/134284 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [clang][modules] Invalidate module cache when SDKSettings.json changes (PR #139751)

https://github.com/benlangmuir edited https://github.com/llvm/llvm-project/pull/139751 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[clang] [clang][modules] Invalidate module cache when SDKSettings.json changes (PR #139751)

https://github.com/benlangmuir approved this pull request. Nice find, thanks for fixing. https://github.com/llvm/llvm-project/pull/139751 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commit

[clang] [clang][modules] Invalidate module cache when SDKSettings.json changes (PR #139751)

@@ -0,0 +1,49 @@ +// This test checks that the module cache gets invalidated when the SDKSettings.json file changes. + +// RUN: rm -rf %t +// RUN: split-file %s %t + +//--- AppleTVOS15.0.sdk/SDKSettings-old.json +{ + "DisplayName": "tvOS 15.0", + "Version": "15.0", + "Canonica

[clang] [Modules] Don't fail when an unused textual header is missing. (PR #138227)

benlangmuir wrote: > I don't think we can ever get rid of that pattern and we will keep relying on > Clang supporting this. > We would need to example I shared above to keep working. > I think the change we want is to always prioritize modular over textual, > before we look at non-private vs p

[clang] [Modules] Don't fail when an unused textual header is missing. (PR #138227)

benlangmuir wrote: > I don't think we can ever get rid of that pattern and we will keep relying on > Clang supporting this. > We would need to example I shared above to keep working. I don't have a strong opinion on the best path forward here, but I'm not clear if you're actually willing to

<    1   2   3   4   5