[clang] [llvm] Alisonz/tmp (PR #204233)
Alison Zhang via cfe-commits
cfe-commits at lists.llvm.org
Tue Jun 16 12:53:40 PDT 2026
https://github.com/azhan92 created https://github.com/llvm/llvm-project/pull/204233
None
>From 72b7bfae7524dbb6a2261de2703217d1c4e565c9 Mon Sep 17 00:00:00 2001
From: alisonzhang <alisonzhang at ibm.com>
Date: Tue, 16 Jun 2026 15:01:43 -0400
Subject: [PATCH 1/4] Changes from fexec-charset PR
---
clang/include/clang/Lex/Preprocessor.h | 4 ++++
clang/include/clang/Lex/TextEncodingConfig.h | 23 ++++++++++++++++++++
clang/lib/Frontend/CompilerInstance.cpp | 1 +
clang/lib/Lex/CMakeLists.txt | 1 +
clang/lib/Lex/TextEncodingConfig.cpp | 20 +++++++++++++++++
5 files changed, 49 insertions(+)
create mode 100644 clang/include/clang/Lex/TextEncodingConfig.h
create mode 100644 clang/lib/Lex/TextEncodingConfig.cpp
diff --git a/clang/include/clang/Lex/Preprocessor.h b/clang/include/clang/Lex/Preprocessor.h
index 8b684e85eb1c1..5f8d5caaafcac 100644
--- a/clang/include/clang/Lex/Preprocessor.h
+++ b/clang/include/clang/Lex/Preprocessor.h
@@ -30,6 +30,7 @@
#include "clang/Lex/ModuleMap.h"
#include "clang/Lex/PPCallbacks.h"
#include "clang/Lex/PPEmbedParameters.h"
+#include "clang/Lex/TextEncodingConfig.h"
#include "clang/Lex/Token.h"
#include "clang/Lex/TokenLexer.h"
#include "clang/Support/Compiler.h"
@@ -198,6 +199,7 @@ class Preprocessor {
std::unique_ptr<ScratchBuffer> ScratchBuf;
HeaderSearch &HeaderInfo;
ModuleLoader &TheModuleLoader;
+ TextEncodingConfig TEC;
/// External source of macros.
ExternalPreprocessorSource *ExternalSource;
@@ -1265,6 +1267,8 @@ class Preprocessor {
Builtin::Context &getBuiltinInfo() { return *BuiltinInfo; }
llvm::BumpPtrAllocator &getPreprocessorAllocator() { return BP; }
+ TextEncodingConfig &getTextEncodingConfig() { return TEC; }
+
void setExternalSource(ExternalPreprocessorSource *Source) {
ExternalSource = Source;
}
diff --git a/clang/include/clang/Lex/TextEncodingConfig.h b/clang/include/clang/Lex/TextEncodingConfig.h
new file mode 100644
index 0000000000000..a810b9ab61b1a
--- /dev/null
+++ b/clang/include/clang/Lex/TextEncodingConfig.h
@@ -0,0 +1,23 @@
+//===-- clang/Lex/TextEncodingConfig.h - Text Conversion Config -*- 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: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_CLANG_LEX_TEXTENCODINGCONFIG_H
+#define LLVM_CLANG_LEX_TEXTENCODINGCONFIG_H
+
+#include "clang/Basic/LangOptions.h"
+#include "llvm/ADT/StringRef.h"
+#include "llvm/Support/TextEncoding.h"
+
+enum ConversionAction { CA_NoConversion };
+
+class TextEncodingConfig {
+public:
+ llvm::TextEncodingConverter *getConverter(ConversionAction Action) const;
+};
+
+#endif
diff --git a/clang/lib/Frontend/CompilerInstance.cpp b/clang/lib/Frontend/CompilerInstance.cpp
index 9e88abbece7f2..09607e6de8ce7 100644
--- a/clang/lib/Frontend/CompilerInstance.cpp
+++ b/clang/lib/Frontend/CompilerInstance.cpp
@@ -34,6 +34,7 @@
#include "clang/Lex/HeaderSearch.h"
#include "clang/Lex/Preprocessor.h"
#include "clang/Lex/PreprocessorOptions.h"
+#include "clang/Lex/TextEncodingConfig.h"
#include "clang/Sema/CodeCompleteConsumer.h"
#include "clang/Sema/ParsedAttr.h"
#include "clang/Sema/Sema.h"
diff --git a/clang/lib/Lex/CMakeLists.txt b/clang/lib/Lex/CMakeLists.txt
index f61737cd68021..106a5d3b126be 100644
--- a/clang/lib/Lex/CMakeLists.txt
+++ b/clang/lib/Lex/CMakeLists.txt
@@ -29,6 +29,7 @@ add_clang_library(clangLex
Preprocessor.cpp
PreprocessorLexer.cpp
ScratchBuffer.cpp
+ TextEncodingConfig.cpp
TokenConcatenation.cpp
TokenLexer.cpp
diff --git a/clang/lib/Lex/TextEncodingConfig.cpp b/clang/lib/Lex/TextEncodingConfig.cpp
new file mode 100644
index 0000000000000..bb3f0d4b4abec
--- /dev/null
+++ b/clang/lib/Lex/TextEncodingConfig.cpp
@@ -0,0 +1,20 @@
+//===--- TextEncodingConfig.cpp -------------------------------------------===//
+//
+// 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: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include "clang/Lex/TextEncodingConfig.h"
+#include "clang/Basic/DiagnosticDriver.h"
+
+using namespace llvm;
+
+llvm::TextEncodingConverter *
+TextEncodingConfig::getConverter(ConversionAction Action) const {
+ switch (Action) {
+ default:
+ return nullptr;
+ }
+}
>From 660b75df3e808eb772982ad7cd14dd81855ac840 Mon Sep 17 00:00:00 2001
From: alisonzhang <alisonzhang at ibm.com>
Date: Tue, 16 Jun 2026 15:34:36 -0400
Subject: [PATCH 2/4] Changes from finput-charset PR
---
.../clang/Basic/DiagnosticCommonKinds.td | 3 +
clang/include/clang/Basic/SourceManager.h | 10 +-
.../include/clang/Frontend/CompilerInstance.h | 1 +
clang/include/clang/Lex/TextEncodingConfig.h | 4 +-
clang/lib/Basic/SourceManager.cpp | 108 ++++++++++++++----
clang/lib/Frontend/CompilerInstance.cpp | 19 +--
.../lib/Frontend/VerifyDiagnosticConsumer.cpp | 4 +-
clang/lib/Lex/ModuleMap.cpp | 5 +-
clang/lib/Lex/PPDirectives.cpp | 7 +-
clang/lib/Lex/Preprocessor.cpp | 5 +-
clang/lib/Lex/TextEncodingConfig.cpp | 2 +
clang/lib/Serialization/ASTReader.cpp | 7 +-
12 files changed, 136 insertions(+), 39 deletions(-)
diff --git a/clang/include/clang/Basic/DiagnosticCommonKinds.td b/clang/include/clang/Basic/DiagnosticCommonKinds.td
index f2ed2f4698b8d..8ebac3908b465 100644
--- a/clang/include/clang/Basic/DiagnosticCommonKinds.td
+++ b/clang/include/clang/Basic/DiagnosticCommonKinds.td
@@ -417,6 +417,9 @@ def note_file_sloc_usage : Note<
"%plural{0:|: plus %2B (%human2B) for macro expansions}2">;
def note_file_misc_sloc_usage : Note<
"%0 additional files entered using a total of %1B (%human1B) of space">;
+def warn_charset_conversion_failed : Warning<
+ "conversion from source encoding failed for '%0': %1; interpreting as IBM-1047">,
+ InGroup<DiagGroup<"charset-conversion-failed">>;
// Modules
def err_module_format_unhandled : Error<
diff --git a/clang/include/clang/Basic/SourceManager.h b/clang/include/clang/Basic/SourceManager.h
index 4217b8683da1e..f7d91d612e4ab 100644
--- a/clang/include/clang/Basic/SourceManager.h
+++ b/clang/include/clang/Basic/SourceManager.h
@@ -50,6 +50,7 @@
#include "llvm/Support/Allocator.h"
#include "llvm/Support/Compiler.h"
#include "llvm/Support/MemoryBuffer.h"
+#include "llvm/Support/TextEncoding.h"
#include <cassert>
#include <cstddef>
#include <map>
@@ -156,6 +157,11 @@ class alignas(8) ContentCache {
/// FIXME: Remove this once OrigEntry is a FileEntryRef with a stable name.
StringRef Filename;
+ /// Information on whether this is associated with a FileID for a file (as
+ /// opposed to a buffer) and, if so, what conversion (if any) was requested.
+ llvm::PointerIntPair<llvm::TextEncodingConverter *, 1u, bool>
+ FileIDConverterInfo;
+
/// A bump pointer allocated array of offsets for each source line.
///
/// This is lazily computed. The lines are owned by the SourceManager
@@ -918,6 +924,7 @@ class SourceManager : public RefCountedBase<SourceManager> {
/// being \#included from the specified IncludePosition.
FileID createFileID(FileEntryRef SourceFile, SourceLocation IncludePos,
SrcMgr::CharacteristicKind FileCharacter,
+ llvm::TextEncodingConverter *Converter = nullptr,
int LoadedID = 0,
SourceLocation::UIntTy LoadedOffset = 0);
@@ -942,7 +949,8 @@ class SourceManager : public RefCountedBase<SourceManager> {
/// Get the FileID for \p SourceFile if it exists. Otherwise, create a
/// new FileID for the \p SourceFile.
FileID getOrCreateFileID(FileEntryRef SourceFile,
- SrcMgr::CharacteristicKind FileCharacter);
+ SrcMgr::CharacteristicKind FileCharacter,
+ llvm::TextEncodingConverter *Converter = nullptr);
/// Creates an expansion SLocEntry for the substitution of an argument into a
/// function-like macro's body. Returns the start of the expansion.
diff --git a/clang/include/clang/Frontend/CompilerInstance.h b/clang/include/clang/Frontend/CompilerInstance.h
index bb0eddb918623..89a0d066afd6f 100644
--- a/clang/include/clang/Frontend/CompilerInstance.h
+++ b/clang/include/clang/Frontend/CompilerInstance.h
@@ -864,6 +864,7 @@ class CompilerInstance : public ModuleLoader {
///
/// \return True on success.
static bool InitializeSourceManager(const FrontendInputFile &Input,
+ llvm::TextEncodingConverter *,
DiagnosticsEngine &Diags,
FileManager &FileMgr,
SourceManager &SourceMgr);
diff --git a/clang/include/clang/Lex/TextEncodingConfig.h b/clang/include/clang/Lex/TextEncodingConfig.h
index a810b9ab61b1a..30e0fcf2ac919 100644
--- a/clang/include/clang/Lex/TextEncodingConfig.h
+++ b/clang/include/clang/Lex/TextEncodingConfig.h
@@ -13,9 +13,11 @@
#include "llvm/ADT/StringRef.h"
#include "llvm/Support/TextEncoding.h"
-enum ConversionAction { CA_NoConversion };
+enum ConversionAction { CA_NoConversion, CA_FromInputEncoding };
class TextEncodingConfig {
+std::unique_ptr<llvm::TextEncodingConverter> FromInputEncodingConverter;
+
public:
llvm::TextEncodingConverter *getConverter(ConversionAction Action) const;
};
diff --git a/clang/lib/Basic/SourceManager.cpp b/clang/lib/Basic/SourceManager.cpp
index b6cc6ec9365f5..c33ee69962864 100644
--- a/clang/lib/Basic/SourceManager.cpp
+++ b/clang/lib/Basic/SourceManager.cpp
@@ -16,6 +16,7 @@
#include "clang/Basic/LLVM.h"
#include "clang/Basic/SourceLocation.h"
#include "clang/Basic/SourceManagerInternals.h"
+#include "clang/Lex/TextEncodingConfig.h"
#include "llvm/ADT/DenseMap.h"
#include "llvm/ADT/MapVector.h"
#include "llvm/ADT/STLExtras.h"
@@ -31,6 +32,7 @@
#include "llvm/Support/ErrorHandling.h"
#include "llvm/Support/MemoryBuffer.h"
#include "llvm/Support/raw_ostream.h"
+#include "llvm/Support/SmallVectorMemoryBuffer.h"
#include <algorithm>
#include <cassert>
#include <cstddef>
@@ -136,7 +138,57 @@ ContentCache::getBufferOrNone(DiagnosticsEngine &Diag, FileManager &FM,
Buffer = std::move(*BufferOrError);
- // Check that the file's size fits in an 'unsigned' (with room for a
+ // Unless this is a named pipe (in which case we can handle a mismatch),
+ // check that the file's size is the same as in the file entry (which may
+ // have come from a stat cache).
+ // The buffer will always be larger than the file size on z/OS in the presence
+ // of characters outside the base character set.
+ assert(Buffer->getBufferSize() >= (size_t)ContentsEntry->getSize());
+ if (!ContentsEntry->isNamedPipe() &&
+ Buffer->getBufferSize() < (size_t)ContentsEntry->getSize()) {
+ Diag.Report(Loc, diag::err_file_modified) << ContentsEntry->getName();
+
+ return std::nullopt;
+ }
+
+ // Convert source from the input charset to UTF-8 if necessary.
+ if (Converter) {
+ StringRef OriginalBuf = Buffer->getBuffer();
+
+ llvm::SmallString<0> UTF8Buf;
+ UTF8Buf.reserve(OriginalBuf.size() + 1);
+
+ std::error_code EC = Converter->convert(OriginalBuf, UTF8Buf);
+ if (EC) {
+ Diag.Report(Loc, diag::warn_charset_conversion_failed)
+ << ContentsEntry->getName() << EC.message();
+#ifdef __MVS__
+ // On z/OS, if conversion fails, try converting from IBM-1047 to UTF-8
+ std::unique_ptr<llvm::TextEncodingConverter> FallbackConverter =
+ TextEncodingConfig::createInputConverterFromFiletag(1047, Diag);
+
+ if (FallbackConverter) {
+ // Try converting with IBM-1047 converter
+ UTF8Buf.clear();
+ UTF8Buf.reserve(OriginalBuf.size() + 1);
+ EC = FallbackConverter->convert(OriginalBuf, UTF8Buf);
+
+ if (!EC) {
+ auto NewBuf = std::make_unique<llvm::SmallVectorMemoryBuffer>(
+ std::move(UTF8Buf), Buffer->getBufferIdentifier());
+ Buffer = std::move(NewBuf);
+ } else {
+ // TODO: Reclaim memory if the buffer size exceeds the content.
+ auto NewBuf = std::make_unique<llvm::SmallVectorMemoryBuffer>(
+ std::move(UTF8Buf), Buffer->getBufferIdentifier());
+ Buffer = std::move(NewBuf);
+ }
+ }
+#endif
+ }
+ }
+
+ // Check that the buffer's size fits in an 'unsigned' (with room for a
// past-the-end value). This is deeply regrettable, but various parts of
// Clang (including elsewhere in this file!) use 'unsigned' to represent file
// offsets, line numbers, string literal lengths, and so on, and fail
@@ -151,22 +203,15 @@ ContentCache::getBufferOrNone(DiagnosticsEngine &Diag, FileManager &FM,
return std::nullopt;
}
- // Unless this is a named pipe (in which case we can handle a mismatch),
- // check that the file's size is the same as in the file entry (which may
- // have come from a stat cache).
- // The buffer will always be larger than the file size on z/OS in the presence
- // of characters outside the base character set.
- assert(Buffer->getBufferSize() >= (size_t)ContentsEntry->getSize());
- if (!ContentsEntry->isNamedPipe() &&
- Buffer->getBufferSize() < (size_t)ContentsEntry->getSize()) {
- Diag.Report(Loc, diag::err_file_modified) << ContentsEntry->getName();
-
- return std::nullopt;
- }
-
- // If the buffer is valid, check to see if it has a UTF Byte Order Mark
- // (BOM). We only support UTF-8 with and without a BOM right now. See
- // http://en.wikipedia.org/wiki/Byte_order_mark for more information.
+ // If the buffer is valid, check to see if it has a UTF Byte Order Mark (BOM)
+ // Note that any conversion requested using `-finput-charset` (if successful)
+ // has already occurred, so we are expecting UTF-8 with or without a BOM.
+ //
+ // In theory, if we see a non-UTF-8 BOM, we can assume that an appropriate
+ // conversion was not supplied via `-finput-charset` and we could try to
+ // convert based on the BOM.
+ //
+ // See http://en.wikipedia.org/wiki/Byte_order_mark for more information.
StringRef BufStr = Buffer->getBuffer();
const char *InvalidBOM = getInvalidBOM(BufStr);
@@ -537,15 +582,30 @@ FileID SourceManager::getNextFileID(FileID FID) const {
/// being \#included from the specified IncludePosition.
FileID SourceManager::createFileID(FileEntryRef SourceFile,
SourceLocation IncludePos,
+ llvm::TextEncodingConverter *Converter,
SrcMgr::CharacteristicKind FileCharacter,
int LoadedID,
SourceLocation::UIntTy LoadedOffset) {
SrcMgr::ContentCache &IR = getOrCreateContentCache(SourceFile,
isSystem(FileCharacter));
+ #ifndef NDEBUG
+ // Either the content cache has never been used for a FileID (and, if we are
+ // being asked to use a converter, there should be no valid buffer set up for
+ // it) or the conversion (or lack thereof) should be the same as that used
+ // previously.
+ auto [CacheConverter, CacheUsedByFileID] = IR.FileIDConverterInfo;
+ if (CacheUsedByFileID)
+ assert(CacheConverter == Converter);
+ else
+ assert(!Converter || IR.IsBufferInvalid || !IR.getBufferIfLoaded());
+#endif
+ IR.FileIDConverterInfo.setPointerAndInt(Converter, true);
+
// If this is a named pipe, immediately load the buffer to ensure subsequent
// calls to ContentCache::getSize() are accurate.
- if (IR.ContentsEntry->isNamedPipe())
+ // Do the same if character-encoding conversion was requested.
+ if (IR.ContentsEntry->isNamedPipe() || Converter)
(void)IR.getBufferOrNone(Diag, getFileManager(), SourceLocation());
return createFileIDImpl(IR, SourceFile.getName(), IncludePos, FileCharacter,
@@ -583,10 +643,12 @@ FileID SourceManager::createFileID(const llvm::MemoryBufferRef &Buffer,
/// new FileID for the \p SourceFile.
FileID
SourceManager::getOrCreateFileID(FileEntryRef SourceFile,
- SrcMgr::CharacteristicKind FileCharacter) {
+ SrcMgr::CharacteristicKind FileCharacter,
+ llvm::TextEncodingConverter *Converter) {
FileID ID = translateFile(SourceFile);
- return ID.isValid() ? ID : createFileID(SourceFile, SourceLocation(),
- FileCharacter);
+ return ID.isValid() ? ID
+ : createFileID(SourceFile, SourceLocation(),
+ FileCharacter, Converter);
}
/// createFileID - Create a new FileID for the specified ContentCache and
@@ -2340,8 +2402,8 @@ SourceManagerForFile::SourceManagerForFile(StringRef FileName,
std::make_unique<DiagnosticsEngine>(DiagnosticIDs::create(), *DiagOpts);
SourceMgr = std::make_unique<SourceManager>(*Diagnostics, *FileMgr);
FileEntryRef FE = llvm::cantFail(FileMgr->getFileRef(FileName));
- FileID ID =
- SourceMgr->createFileID(FE, SourceLocation(), clang::SrcMgr::C_User);
+ FileID ID = SourceMgr->createFileID(
+ FE, SourceLocation(), clang::SrcMgr::C_User, /*Converter=*/nullptr);
assert(ID.isValid());
SourceMgr->setMainFileID(ID);
}
diff --git a/clang/lib/Frontend/CompilerInstance.cpp b/clang/lib/Frontend/CompilerInstance.cpp
index 09607e6de8ce7..e94d05ba33fbb 100644
--- a/clang/lib/Frontend/CompilerInstance.cpp
+++ b/clang/lib/Frontend/CompilerInstance.cpp
@@ -907,15 +907,20 @@ CompilerInstance::createOutputFileImpl(StringRef OutputPath, bool Binary,
// Initialization Utilities
bool CompilerInstance::InitializeSourceManager(const FrontendInputFile &Input){
- return InitializeSourceManager(Input, getDiagnostics(), getFileManager(),
- getSourceManager());
+ // Retrieve the converter to the internal charset if it exists.
+ llvm::TextEncodingConverter *Converter =
+ hasPreprocessor() ? getPreprocessor().getTextEncodingConfig().getConverter(
+ CA_FromInputEncoding)
+ : nullptr;
+
+ return InitializeSourceManager(Input, Converter, getDiagnostics(),
+ getFileManager(), getSourceManager());
}
// static
-bool CompilerInstance::InitializeSourceManager(const FrontendInputFile &Input,
- DiagnosticsEngine &Diags,
- FileManager &FileMgr,
- SourceManager &SourceMgr) {
+bool CompilerInstance::InitializeSourceManager(
+ const FrontendInputFile &Input, llvm::TextEncodingConverter *Converter,
+ DiagnosticsEngine &Diags, FileManager &FileMgr, SourceManager &SourceMgr) {
SrcMgr::CharacteristicKind Kind =
Input.getKind().getFormat() == InputKind::ModuleMap
? Input.isSystem() ? SrcMgr::C_System_ModuleMap
@@ -923,7 +928,7 @@ bool CompilerInstance::InitializeSourceManager(const FrontendInputFile &Input,
: Input.isSystem() ? SrcMgr::C_System : SrcMgr::C_User;
if (Input.isBuffer()) {
- SourceMgr.setMainFileID(SourceMgr.createFileID(Input.getBuffer(), Kind));
+ SourceMgr.setMainFileID(SourceMgr.createFileID(Input.getBuffer(), Kind, Converter));
assert(SourceMgr.getMainFileID().isValid() &&
"Couldn't establish MainFileID!");
return true;
diff --git a/clang/lib/Frontend/VerifyDiagnosticConsumer.cpp b/clang/lib/Frontend/VerifyDiagnosticConsumer.cpp
index 1bfe644b2525a..01e3b20e0c7cb 100644
--- a/clang/lib/Frontend/VerifyDiagnosticConsumer.cpp
+++ b/clang/lib/Frontend/VerifyDiagnosticConsumer.cpp
@@ -610,8 +610,10 @@ static bool ParseDirective(StringRef S, ExpectedData *ED, SourceManager &SM,
}
FileID FID = SM.translateFile(*File);
+ // FIXME: Figure out character-encoding converter treatment.
if (FID.isInvalid())
- FID = SM.createFileID(*File, Pos, SrcMgr::C_User);
+ FID = SM.createFileID(*File, Pos, SrcMgr::C_User,
+ /*Converter=*/nullptr);
if (PH.Next(Line) && Line > 0)
ExpectedLoc = SM.translateLineCol(FID, Line, 1);
diff --git a/clang/lib/Lex/ModuleMap.cpp b/clang/lib/Lex/ModuleMap.cpp
index 6c07386f89010..c7c1e04b76ea1 100644
--- a/clang/lib/Lex/ModuleMap.cpp
+++ b/clang/lib/Lex/ModuleMap.cpp
@@ -1473,7 +1473,10 @@ bool ModuleMap::parseModuleMapFile(FileEntryRef File, bool IsSystem,
if (LocalFID.isInvalid()) {
auto FileCharacter =
IsSystem ? SrcMgr::C_System_ModuleMap : SrcMgr::C_User_ModuleMap;
- LocalFID = SourceMgr.createFileID(File, ExternModuleLoc, FileCharacter);
+ // FIXME: Module map files are also textual "source files". For consistency,
+ // conversion should occur.
+ LocalFID = SourceMgr.createFileID(File, ExternModuleLoc, FileCharacter,
+ /*Converter=*/nullptr);
}
ID = LocalFID;
}
diff --git a/clang/lib/Lex/PPDirectives.cpp b/clang/lib/Lex/PPDirectives.cpp
index eb21a510dcf83..2e095fce02c0f 100644
--- a/clang/lib/Lex/PPDirectives.cpp
+++ b/clang/lib/Lex/PPDirectives.cpp
@@ -2796,7 +2796,12 @@ Preprocessor::ImportAction Preprocessor::HandleHeaderIncludeOrImport(
// position on the file where it will be included and after the expansions.
if (IncludePos.isMacroID())
IncludePos = SourceMgr.getExpansionRange(IncludePos).getEnd();
- FileID FID = SourceMgr.createFileID(*File, IncludePos, FileCharacter);
+ // Retrieve the converter to the internal charset if it exists.
+ llvm::TextEncodingConverter *Converter =
+ getTextEncodingConfig().getConverter(CA_FromInputEncoding);
+
+ FileID FID =
+ SourceMgr.createFileID(*File, IncludePos, FileCharacter, Converter);
if (!FID.isValid()) {
TheModuleLoader.HadFatalFailure = true;
return ImportAction::Failure;
diff --git a/clang/lib/Lex/Preprocessor.cpp b/clang/lib/Lex/Preprocessor.cpp
index 1e21b4a94cea3..c5e32468bcc7e 100644
--- a/clang/lib/Lex/Preprocessor.cpp
+++ b/clang/lib/Lex/Preprocessor.cpp
@@ -649,8 +649,9 @@ void Preprocessor::EnterMainSourceFile() {
<< PPOpts.PCHThroughHeader;
return;
}
- setPCHThroughHeaderFileID(
- SourceMgr.createFileID(*File, SourceLocation(), SrcMgr::C_User));
+ // FIXME: Figure out character-encoding converter treatment.
+ setPCHThroughHeaderFileID(SourceMgr.createFileID(
+ *File, SourceLocation(), SrcMgr::C_User, /*Converter=*/nullptr));
}
// Skip tokens from the Predefines and if needed the main file.
diff --git a/clang/lib/Lex/TextEncodingConfig.cpp b/clang/lib/Lex/TextEncodingConfig.cpp
index bb3f0d4b4abec..ed426c304423b 100644
--- a/clang/lib/Lex/TextEncodingConfig.cpp
+++ b/clang/lib/Lex/TextEncodingConfig.cpp
@@ -14,6 +14,8 @@ using namespace llvm;
llvm::TextEncodingConverter *
TextEncodingConfig::getConverter(ConversionAction Action) const {
switch (Action) {
+ case CA_FromInputEncoding:
+ return FromInputEncodingConverter.get();
default:
return nullptr;
}
diff --git a/clang/lib/Serialization/ASTReader.cpp b/clang/lib/Serialization/ASTReader.cpp
index 74a7b51368c28..7bc28a2791067 100644
--- a/clang/lib/Serialization/ASTReader.cpp
+++ b/clang/lib/Serialization/ASTReader.cpp
@@ -2002,8 +2002,11 @@ bool ASTReader::ReadSLocEntry(int ID) {
}
SrcMgr::CharacteristicKind
FileCharacter = (SrcMgr::CharacteristicKind)Record[2];
- FileID FID = SourceMgr.createFileID(*File, IncludeLoc, FileCharacter, ID,
- BaseOffset + Record[0]);
+ // Note: If conversion was originally necessary, OverriddenBuffer should be
+ // true and the associated handling will trigger.
+ FileID FID = SourceMgr.createFileID(*File, IncludeLoc, FileCharacter,
+ /*Converter=*/nullptr, ID,
+ BaseOffset + Record[0]);
SrcMgr::FileInfo &FileInfo = SourceMgr.getSLocEntry(FID).getFile();
FileInfo.NumCreatedFIDs = Record[5];
if (Record[3])
>From 51adf4f5b68b461d667507bb1e568d87cab8e23e Mon Sep 17 00:00:00 2001
From: alisonzhang <alisonzhang at ibm.com>
Date: Tue, 16 Jun 2026 15:34:51 -0400
Subject: [PATCH 3/4] Add wrapper function for getzOSFileTag
---
llvm/include/llvm/Support/AutoConvert.h | 17 +++++++++++++++++
1 file changed, 17 insertions(+)
diff --git a/llvm/include/llvm/Support/AutoConvert.h b/llvm/include/llvm/Support/AutoConvert.h
index d68b0e8b515e0..9586d41e10c2f 100644
--- a/llvm/include/llvm/Support/AutoConvert.h
+++ b/llvm/include/llvm/Support/AutoConvert.h
@@ -105,6 +105,23 @@ inline ErrorOr<bool> needConversion(const Twine &FileName, const int FD = -1) {
return false;
}
+#ifdef __MVS__
+inline ErrorOr<__ccsid_t> getFileTag(const Twine &FileName, const int FD = -1) {
+ ErrorOr<__ccsid_t> Ccsid = getzOSFileTag(FileName, FD);
+ if (!Ccsid)
+ return Ccsid;
+ // Assume untagged files to be IBM-1047 encoded
+ if (*Ccsid == 0)
+ return CCSID_IBM_1047;
+ return Ccsid;
+}
+#else
+inline ErrorOr<int> getFileTag(const Twine &FileName, const int FD = -1) {
+ // On non-z/OS platforms, return 0 to indicate no file tag
+ return 0;
+}
+#endif
+
} /* namespace llvm */
#endif /* __cplusplus */
>From b8b8c2945182007ee9f02ef74358acb4b40ea0f3 Mon Sep 17 00:00:00 2001
From: alisonzhang <alisonzhang at ibm.com>
Date: Tue, 16 Jun 2026 15:52:52 -0400
Subject: [PATCH 4/4] Changes from Bob
---
clang/include/clang/Lex/TextEncodingConfig.h | 8 +++++
clang/lib/Basic/SourceManager.cpp | 7 +++-
clang/lib/Lex/TextEncodingConfig.cpp | 38 ++++++++++++++++++++
llvm/include/llvm/Support/TextEncoding.h | 22 ++++++++++++
llvm/lib/Support/TextEncoding.cpp | 36 +++++++++++++++++++
5 files changed, 110 insertions(+), 1 deletion(-)
diff --git a/clang/include/clang/Lex/TextEncodingConfig.h b/clang/include/clang/Lex/TextEncodingConfig.h
index 30e0fcf2ac919..e4f6595997888 100644
--- a/clang/include/clang/Lex/TextEncodingConfig.h
+++ b/clang/include/clang/Lex/TextEncodingConfig.h
@@ -17,9 +17,17 @@ enum ConversionAction { CA_NoConversion, CA_FromInputEncoding };
class TextEncodingConfig {
std::unique_ptr<llvm::TextEncodingConverter> FromInputEncodingConverter;
+llvm::StringMap<std::unique_ptr<llvm::TextEncodingConverter>> FromFiletagEncodingConverters;
public:
llvm::TextEncodingConverter *getConverter(ConversionAction Action) const;
+ static llvm::TextEncodingConverter *
+ getFromFiletagEncodingConverter(TextEncodingConfig &TEC,
+ llvm::StringRef FiletagEncoding);
+ static llvm::TextEncodingConverter *
+ createAndInsertFromFiletagEncodingConverter(TextEncodingConfig &TEC,
+ llvm::StringRef FiletagEncoding,
+ clang::DiagnosticsEngine &Diag);
};
#endif
diff --git a/clang/lib/Basic/SourceManager.cpp b/clang/lib/Basic/SourceManager.cpp
index c33ee69962864..25a7023b7db85 100644
--- a/clang/lib/Basic/SourceManager.cpp
+++ b/clang/lib/Basic/SourceManager.cpp
@@ -122,7 +122,10 @@ ContentCache::getBufferOrNone(DiagnosticsEngine &Diag, FileManager &FM,
// return paths.
IsBufferInvalid = true;
- auto BufferOrError = FM.getBufferForFile(*ContentsEntry, IsFileVolatile);
+ auto BufferOrError = FM.getBufferForFile(*ContentsEntry, IsFileVolatile,
+ /*RequiresNullTerminator=*/true,
+ /*MaybeLimit=*/std::nullopt,
+ /*IsText=*/false);
// If we were unable to open the file, then we are in an inconsistent
// situation where the content cache referenced a file which no longer
@@ -589,6 +592,8 @@ FileID SourceManager::createFileID(FileEntryRef SourceFile,
SrcMgr::ContentCache &IR = getOrCreateContentCache(SourceFile,
isSystem(FileCharacter));
+ llvm::ErrorOr<int> Ccsid = llvm::getFileTag(SourceFile.getName());
+ if (Ccsid && *Ccsid > 0) {
#ifndef NDEBUG
// Either the content cache has never been used for a FileID (and, if we are
// being asked to use a converter, there should be no valid buffer set up for
diff --git a/clang/lib/Lex/TextEncodingConfig.cpp b/clang/lib/Lex/TextEncodingConfig.cpp
index ed426c304423b..01c9bfcf607db 100644
--- a/clang/lib/Lex/TextEncodingConfig.cpp
+++ b/clang/lib/Lex/TextEncodingConfig.cpp
@@ -20,3 +20,41 @@ TextEncodingConfig::getConverter(ConversionAction Action) const {
return nullptr;
}
}
+
+llvm::TextEncodingConverter *
+TextEncodingConfig::getFromFiletagEncodingConverter(TextEncodingConfig &TEC,
+ StringRef FiletagEncoding) {
+ auto Iter = TEC.FromFiletagEncodingConverters.find(FiletagEncoding);
+ if (Iter != TEC.FromFiletagEncodingConverters.end()) {
+ return Iter->second.get();
+ }
+ return nullptr;
+}
+
+llvm::TextEncodingConverter *
+TextEncodingConfig::createAndInsertFromFiletagEncodingConverter(TextEncodingConfig &TEC,
+ StringRef FiletagEncoding,
+ clang::DiagnosticsEngine &Diag) {
+ llvm::TextEncodingConverter *Converter = getFromFiletagtEncodingConverter(TEC, FiletagEncoding);
+ if (Converter)
+ return Converter;
+
+ const char *UTF8 = "UTF-8";
+ // Create a converter between the input and internal encodings
+ if (FiletagEncoding != UTF8) {
+ ErrorOr<TextEncodingConverter> ErrorOrConverter =
+ llvm::TextEncodingConverter::create(FiletagEncoding, UTF8);
+ if (!ErrorOrConverter) {
+ Diag.Report(clang::diag::err_drv_invalid_value)
+ << "Filetag encoding" << FiletagEncoding;
+ return nullptr;
+ } else {
+ auto FromFiletagEncodingConverter = std::make_unique<llvm::TextEncodingConverter>(
+ std::move(*ErrorOrConverter));
+ llvm::TextEncodingConverter *Result = FromFiletagEncodingConverter.get();
+ TEC.FromFiletagEncodingConverters.insert_or_assign(FiletagEncoding, std::move(FromFiletagEncodingConverter));
+ return Result;
+ }
+ }
+ return nullptr;
+}
diff --git a/llvm/include/llvm/Support/TextEncoding.h b/llvm/include/llvm/Support/TextEncoding.h
index 8a304910aa5dd..24a2e026ebc82 100644
--- a/llvm/include/llvm/Support/TextEncoding.h
+++ b/llvm/include/llvm/Support/TextEncoding.h
@@ -16,11 +16,13 @@
#define LLVM_SUPPORT_TEXT_ENCODING_H
#include "llvm/ADT/SmallString.h"
+#include "llvm/ADT/StringMap.h"
#include "llvm/ADT/StringRef.h"
#include "llvm/Config/config.h"
#include "llvm/Support/Compiler.h"
#include "llvm/Support/ErrorOr.h"
+#include <memory>
#include <string>
#include <system_error>
@@ -137,6 +139,26 @@ class TextEncodingConverter {
}
};
+/// Utility class to manage a cache of TextEncodingConverter instances.
+/// This is useful when you need to convert from multiple source encodings
+/// to a common target encoding (e.g., UTF-8).
+class TextEncodingConverterCache {
+ StringMap<std::unique_ptr<TextEncodingConverter>> Converters;
+
+public:
+ /// Get a converter from the cache, or nullptr if not found.
+ /// \param[in] SourceEncoding the source character encoding name
+ /// \return pointer to the converter or nullptr
+ TextEncodingConverter *getConverter(StringRef SourceEncoding) const;
+
+ /// Create and insert a converter into the cache.
+ /// \param[in] SourceEncoding the source character encoding name
+ /// \param[in] TargetEncoding the target character encoding name
+ /// \return pointer to the converter or nullptr on error
+ TextEncodingConverter *createAndInsertConverter(StringRef SourceEncoding,
+ StringRef TargetEncoding);
+};
+
} // namespace llvm
#endif
diff --git a/llvm/lib/Support/TextEncoding.cpp b/llvm/lib/Support/TextEncoding.cpp
index d36f02c1300b9..2a1d19935699e 100644
--- a/llvm/lib/Support/TextEncoding.cpp
+++ b/llvm/lib/Support/TextEncoding.cpp
@@ -356,3 +356,39 @@ ErrorOr<TextEncodingConverter> TextEncodingConverter::create(StringRef From,
return std::make_error_code(std::errc::invalid_argument);
#endif
}
+
+TextEncodingConverter *
+TextEncodingConverterCache::getConverter(StringRef SourceEncoding) const {
+ auto Iter = Converters.find(SourceEncoding);
+ if (Iter != Converters.end()) {
+ return Iter->second.get();
+ }
+ return nullptr;
+}
+
+TextEncodingConverter *
+TextEncodingConverterCache::createAndInsertConverter(StringRef SourceEncoding,
+ StringRef TargetEncoding) {
+ // Check if converter already exists
+ TextEncodingConverter *Converter = getConverter(SourceEncoding);
+ if (Converter)
+ return Converter;
+
+ // Don't create a converter if source and target are the same
+ if (SourceEncoding == TargetEncoding)
+ return nullptr;
+
+ // Create a new converter
+ ErrorOr<TextEncodingConverter> ErrorOrConverter =
+ TextEncodingConverter::create(SourceEncoding, TargetEncoding);
+ if (!ErrorOrConverter) {
+ return nullptr;
+ }
+
+ // Insert into cache and return pointer
+ auto NewConverter =
+ std::make_unique<TextEncodingConverter>(std::move(*ErrorOrConverter));
+ TextEncodingConverter *Result = NewConverter.get();
+ Converters.insert_or_assign(SourceEncoding, std::move(NewConverter));
+ return Result;
+}
More information about the cfe-commits
mailing list