Skip to content

Conversation

charles-zablit
Copy link
Contributor

This is a relanding of #149493. The tests were failing because we were interpreting a proper file descriptor as a console file descriptor.

This patch uses the Windows APIs to print to the Windows Console, through llvm::raw_fd_ostream.

This fixes a rendering issue where the characters defined in DiagnosticsRendering.cpp ("╰" for instance) are not rendered properly on Windows out of the box, because the default codepage is not utf-8.

This solution is based on this patch downstream.

rdar://156064500

Copy link
Contributor

@Nerixyz Nerixyz left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

(Windows 11, x86) Rendering in the terminal works as expected and the previously failing tests now pass for me.

@charles-zablit
Copy link
Contributor Author

Updating the branch to hopefully get the CI job to pass.

@charles-zablit charles-zablit merged commit 1cee0e7 into llvm:main Sep 2, 2025
9 checks passed
charles-zablit added a commit to charles-zablit/llvm-project that referenced this pull request Sep 3, 2025
This is a relanding of llvm#149493.
The tests were failing because we were interpreting a proper file
descriptor as a console file descriptor.

This patch uses the Windows APIs to print to the Windows Console,
through `llvm::raw_fd_ostream`.

This fixes a rendering issue where the characters defined in
`DiagnosticsRendering.cpp` ("╰" for instance) are not rendered properly
on Windows out of the box, because the default codepage is not `utf-8`.

This solution is based on [this patch
downstream](https://github.com/swiftlang/swift/pull/40632/files#diff-e948e4bd7a601e3ca82d596058ccb39326459a4751470eec4d393adeaf516977R37-R38).

rdar://156064500
@llvmbot
Copy link
Member

llvmbot commented Sep 3, 2025

@llvm/pr-subscribers-lldb

@llvm/pr-subscribers-platform-windows

Author: Charles Zablit (charles-zablit)

Changes

This is a relanding of #149493. The tests were failing because we were interpreting a proper file descriptor as a console file descriptor.

This patch uses the Windows APIs to print to the Windows Console, through llvm::raw_fd_ostream.

This fixes a rendering issue where the characters defined in DiagnosticsRendering.cpp ("╰" for instance) are not rendered properly on Windows out of the box, because the default codepage is not utf-8.

This solution is based on this patch downstream.

rdar://156064500


Full diff: https://github.com/llvm/llvm-project/pull/156469.diff

2 Files Affected:

  • (modified) lldb/include/lldb/Host/File.h (+7-9)
  • (modified) lldb/source/Host/common/File.cpp (+34)
diff --git a/lldb/include/lldb/Host/File.h b/lldb/include/lldb/Host/File.h
index 9e2d0abe0b1af..7402a2231735a 100644
--- a/lldb/include/lldb/Host/File.h
+++ b/lldb/include/lldb/Host/File.h
@@ -382,15 +382,11 @@ class NativeFile : public File {
     Unowned = false,
   };
 
-  NativeFile() : m_descriptor(kInvalidDescriptor), m_stream(kInvalidStream) {}
+  NativeFile();
 
-  NativeFile(FILE *fh, bool transfer_ownership)
-      : m_descriptor(kInvalidDescriptor), m_own_descriptor(false), m_stream(fh),
-        m_options(), m_own_stream(transfer_ownership) {}
+  NativeFile(FILE *fh, bool transfer_ownership);
 
-  NativeFile(int fd, OpenOptions options, bool transfer_ownership)
-      : m_descriptor(fd), m_own_descriptor(transfer_ownership),
-        m_stream(kInvalidStream), m_options(options), m_own_stream(false) {}
+  NativeFile(int fd, OpenOptions options, bool transfer_ownership);
 
   ~NativeFile() override { Close(); }
 
@@ -444,17 +440,19 @@ class NativeFile : public File {
     return ValueGuard(m_stream_mutex, StreamIsValidUnlocked());
   }
 
-  int m_descriptor;
+  int m_descriptor = kInvalidDescriptor;
   bool m_own_descriptor = false;
   mutable std::mutex m_descriptor_mutex;
 
-  FILE *m_stream;
+  FILE *m_stream = kInvalidStream;
   mutable std::mutex m_stream_mutex;
 
   OpenOptions m_options{};
   bool m_own_stream = false;
   std::mutex offset_access_mutex;
 
+  bool is_windows_console = false;
+
 private:
   NativeFile(const NativeFile &) = delete;
   const NativeFile &operator=(const NativeFile &) = delete;
diff --git a/lldb/source/Host/common/File.cpp b/lldb/source/Host/common/File.cpp
index 23b6dc9fe850d..8fd1ca069dc01 100644
--- a/lldb/source/Host/common/File.cpp
+++ b/lldb/source/Host/common/File.cpp
@@ -36,6 +36,7 @@
 #include "llvm/Support/Errno.h"
 #include "llvm/Support/FileSystem.h"
 #include "llvm/Support/Process.h"
+#include "llvm/Support/raw_ostream.h"
 
 using namespace lldb;
 using namespace lldb_private;
@@ -247,6 +248,32 @@ uint32_t File::GetPermissions(Status &error) const {
   return file_stats.st_mode & (S_IRWXU | S_IRWXG | S_IRWXO);
 }
 
+NativeFile::NativeFile() = default;
+
+NativeFile::NativeFile(FILE *fh, bool transfer_ownership)
+    : m_stream(fh), m_own_stream(transfer_ownership) {
+#ifdef _WIN32
+  // In order to properly display non ASCII characters in Windows, we need to
+  // use Windows APIs to print to the console. This is only required if the
+  // stream outputs to a console.
+  int fd = _fileno(fh);
+  is_windows_console =
+      ::GetFileType((HANDLE)::_get_osfhandle(fd)) == FILE_TYPE_CHAR;
+#endif
+}
+
+NativeFile::NativeFile(int fd, OpenOptions options, bool transfer_ownership)
+    : m_descriptor(fd), m_own_descriptor(transfer_ownership),
+      m_options(options) {
+#ifdef _WIN32
+  // In order to properly display non ASCII characters in Windows, we need to
+  // use Windows APIs to print to the console. This is only required if the
+  // file outputs to a console.
+  is_windows_console =
+      ::GetFileType((HANDLE)::_get_osfhandle(fd)) == FILE_TYPE_CHAR;
+#endif
+}
+
 bool NativeFile::IsValid() const {
   std::scoped_lock<std::mutex, std::mutex> lock(m_descriptor_mutex, m_stream_mutex);
   return DescriptorIsValidUnlocked() || StreamIsValidUnlocked();
@@ -629,6 +656,13 @@ Status NativeFile::Write(const void *buf, size_t &num_bytes) {
   }
 
   if (ValueGuard stream_guard = StreamIsValid()) {
+#ifdef _WIN32
+    if (is_windows_console) {
+      llvm::raw_fd_ostream(_fileno(m_stream), false)
+          .write((char *)buf, num_bytes);
+      return error;
+    }
+#endif
     bytes_written = ::fwrite(buf, 1, num_bytes, m_stream);
 
     if (bytes_written == 0) {

charles-zablit added a commit to charles-zablit/llvm-project that referenced this pull request Sep 4, 2025
This is a relanding of llvm#149493.
The tests were failing because we were interpreting a proper file
descriptor as a console file descriptor.

This patch uses the Windows APIs to print to the Windows Console,
through `llvm::raw_fd_ostream`.

This fixes a rendering issue where the characters defined in
`DiagnosticsRendering.cpp` ("╰" for instance) are not rendered properly
on Windows out of the box, because the default codepage is not `utf-8`.

This solution is based on [this patch
downstream](https://github.com/swiftlang/swift/pull/40632/files#diff-e948e4bd7a601e3ca82d596058ccb39326459a4751470eec4d393adeaf516977R37-R38).

rdar://156064500
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants