[libc-commits] [libc] [libc] Support %lc in printf (PR #169983)
Shubh Pachchigar via libc-commits
libc-commits at lists.llvm.org
Tue Jan 6 21:34:46 PST 2026
https://github.com/shubhe25p updated https://github.com/llvm/llvm-project/pull/169983
>From 08e5007c915ead02e797a682e8bd8042747b1961 Mon Sep 17 00:00:00 2001
From: "shubh at DOE" <shubhp at mbm3a24.local>
Date: Tue, 6 Jan 2026 21:34:07 -0800
Subject: [PATCH] [libc] Support %lc in printf
Add support for %lc in printf by calling internal
wcrtomb function and relevant end-to-end sprintf
test. Additionally, modified printf parser for
recognizing length modifier and added two internal
error codes. Also added a flag to disable wchar
support on windows platform.
---
.../modules/LLVMLibCCompileOptionRules.cmake | 4 +
libc/config/config.json | 4 +
libc/config/windows/config.json | 8 ++
libc/docs/configure.rst | 1 +
libc/docs/dev/printf_behavior.rst | 7 ++
libc/src/stdio/printf_core/CMakeLists.txt | 19 ++++
libc/src/stdio/printf_core/char_converter.h | 49 +++++++++-
libc/src/stdio/printf_core/core_structs.h | 2 +
.../stdio/printf_core/linux/error_mapper.h | 3 +
libc/src/stdio/printf_core/parser.h | 32 +++++-
libc/test/src/stdio/CMakeLists.txt | 10 ++
libc/test/src/stdio/sprintf_test.cpp | 98 ++++++++++++++++++-
12 files changed, 226 insertions(+), 11 deletions(-)
create mode 100644 libc/config/windows/config.json
diff --git a/libc/cmake/modules/LLVMLibCCompileOptionRules.cmake b/libc/cmake/modules/LLVMLibCCompileOptionRules.cmake
index a719f20d532e0..638cfc122d7de 100644
--- a/libc/cmake/modules/LLVMLibCCompileOptionRules.cmake
+++ b/libc/cmake/modules/LLVMLibCCompileOptionRules.cmake
@@ -135,6 +135,10 @@ function(_get_compile_options_from_config output_var)
endif()
endif()
+ if (LIBC_CONF_PRINTF_DISABLE_WIDE)
+ list(APPEND config_options "-DLIBC_COPT_PRINTF_DISABLE_WIDE")
+ endif()
+
set(${output_var} ${config_options} PARENT_SCOPE)
endfunction(_get_compile_options_from_config)
diff --git a/libc/config/config.json b/libc/config/config.json
index f981c433b2c7c..296d2e539c23d 100644
--- a/libc/config/config.json
+++ b/libc/config/config.json
@@ -52,6 +52,10 @@
"LIBC_CONF_PRINTF_RUNTIME_DISPATCH": {
"value": true,
"doc": "Use dynamic dispatch for the output mechanism to reduce code size."
+ },
+ "LIBC_CONF_PRINTF_DISABLE_WIDE": {
+ "value": false,
+ "doc": "Disable handling wide characters for printf and friends."
}
},
"scanf": {
diff --git a/libc/config/windows/config.json b/libc/config/windows/config.json
new file mode 100644
index 0000000000000..6bb09db9d1aab
--- /dev/null
+++ b/libc/config/windows/config.json
@@ -0,0 +1,8 @@
+{
+ "printf": {
+ "LIBC_CONF_PRINTF_DISABLE_WIDE": {
+ "value": "true",
+ "doc": "Disable handling wide characters for printf and friends."
+ }
+ }
+}
diff --git a/libc/docs/configure.rst b/libc/docs/configure.rst
index 81888bb07153c..8f5074ec23475 100644
--- a/libc/docs/configure.rst
+++ b/libc/docs/configure.rst
@@ -49,6 +49,7 @@ to learn about the defaults for your platform and target.
- ``LIBC_CONF_PRINTF_FLOAT_TO_STR_USE_FLOAT320``: Use an alternative printf float implementation based on 320-bit floats
- ``LIBC_CONF_PRINTF_FLOAT_TO_STR_USE_MEGA_LONG_DOUBLE_TABLE``: Use large table for better printf long double performance.
- ``LIBC_CONF_PRINTF_RUNTIME_DISPATCH``: Use dynamic dispatch for the output mechanism to reduce code size.
+ - ``LIBC_CONF_PRINTF_DISABLE_WIDE``: Disable handling of %lc and %ls on unsupported platforms.
* **"pthread" options**
- ``LIBC_CONF_RAW_MUTEX_DEFAULT_SPIN_COUNT``: Default number of spins before blocking if a mutex is in contention (default to 100).
- ``LIBC_CONF_RWLOCK_DEFAULT_SPIN_COUNT``: Default number of spins before blocking if a rwlock is in contention (default to 100).
diff --git a/libc/docs/dev/printf_behavior.rst b/libc/docs/dev/printf_behavior.rst
index 01ab128a1f238..ba0578aee3fd8 100644
--- a/libc/docs/dev/printf_behavior.rst
+++ b/libc/docs/dev/printf_behavior.rst
@@ -71,6 +71,13 @@ conversions (%r, %k); any fixed point number conversion will be treated as
invalid. This reduces code size. This has no effect if the current compiler does
not support fixed point numbers.
+LIBC_COPT_PRINTF_DISABLE_WIDE
+--------------------------------
+When set, this flag disables support for wide characters (%lc and %ls). Any
+conversions will be ignored. This reduces code size. This will be set by default
+on windows platforms as current printf implementation does not support UTF-16 wide
+characters.
+
LIBC_COPT_PRINTF_NO_NULLPTR_CHECKS
----------------------------------
When set, this flag disables the nullptr checks in %n and %s.
diff --git a/libc/src/stdio/printf_core/CMakeLists.txt b/libc/src/stdio/printf_core/CMakeLists.txt
index 624129b2b36e7..ae93cc754299b 100644
--- a/libc/src/stdio/printf_core/CMakeLists.txt
+++ b/libc/src/stdio/printf_core/CMakeLists.txt
@@ -43,6 +43,22 @@ if(NOT TARGET ${target_error_mapper})
set(target_error_mapper libc.src.stdio.printf_core.generic.error_mapper)
endif()
+if(LIBC_CONF_PRINTF_DISABLE_WIDE)
+ set(wchar_deps "")
+ set(parser_wchar_deps "")
+else()
+ set(wchar_deps
+ libc.hdr.types.wchar_t
+ libc.hdr.types.wint_t
+ libc.hdr.wchar_macros
+ libc.src.__support.wchar.wcrtomb
+ libc.src.__support.wchar.mbstate
+ )
+ set(parser_wchar_deps
+ libc.hdr.types.wint_t
+ )
+endif()
+
add_header_library(
printf_config
HDRS
@@ -76,6 +92,7 @@ add_header_library(
libc.src.__support.CPP.string_view
libc.src.__support.CPP.type_traits
libc.src.__support.common
+ ${parser_wchar_deps}
)
add_header_library(
@@ -111,6 +128,7 @@ add_header_library(
.printf_config
.writer
libc.include.inttypes
+ libc.hdr.limits_macros
libc.src.__support.big_int
libc.src.__support.common
libc.src.__support.CPP.limits
@@ -125,6 +143,7 @@ add_header_library(
libc.src.__support.uint128
libc.src.__support.StringUtil.error_to_string
libc.src.string.memory_utils.inline_memcpy
+ ${wchar_deps}
)
add_header_library(
diff --git a/libc/src/stdio/printf_core/char_converter.h b/libc/src/stdio/printf_core/char_converter.h
index fd2eb2553887a..a5512d172c1b2 100644
--- a/libc/src/stdio/printf_core/char_converter.h
+++ b/libc/src/stdio/printf_core/char_converter.h
@@ -1,4 +1,4 @@
-//===-- String Converter for printf -----------------------------*- C++ -*-===//
+//===-- Character Converter for printf --------------------------*- C++ -*-===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
@@ -9,6 +9,15 @@
#ifndef LLVM_LIBC_SRC_STDIO_PRINTF_CORE_CHAR_CONVERTER_H
#define LLVM_LIBC_SRC_STDIO_PRINTF_CORE_CHAR_CONVERTER_H
+#ifndef LIBC_COPT_PRINTF_DISABLE_WIDE
+#include "hdr/types/wchar_t.h"
+#include "hdr/types/wint_t.h"
+#include "hdr/wchar_macros.h"
+#include "src/__support/wchar/mbstate.h"
+#include "src/__support/wchar/wcrtomb.h"
+#endif // LIBC_COPT_PRINTF_DISABLE_WIDE
+
+#include "hdr/limits_macros.h"
#include "src/__support/macros/config.h"
#include "src/stdio/printf_core/converter_utils.h"
#include "src/stdio/printf_core/core_structs.h"
@@ -20,12 +29,42 @@ namespace printf_core {
template <WriteMode write_mode>
LIBC_INLINE int convert_char(Writer<write_mode> *writer,
const FormatSection &to_conv) {
- char c = static_cast<char>(to_conv.conv_val_raw);
- constexpr int STRING_LEN = 1;
+ char buffer[MB_LEN_MAX];
+ size_t write_size = 0;
+
+ if (to_conv.length_modifier == LengthModifier::l) {
+#ifndef LIBC_COPT_PRINTF_DISABLE_WIDE
+ wint_t wi = static_cast<wint_t>(to_conv.conv_val_raw);
+
+ if (wi == WEOF) {
+ return ILLEGAL_WIDE_CHAR;
+ }
+
+ internal::mbstate mbstate;
+ wchar_t wc = static_cast<wchar_t>(wi);
+ auto ret = internal::wcrtomb(buffer, wc, &mbstate);
+
+ if (!ret.has_value()) {
+ return MB_CONVERSION_ERROR;
+ }
+
+ write_size = ret.value();
+#else
+ // If wide characters are disabled, treat the 'l' modifier as a no-op.
+ buffer[0] = static_cast<char>(to_conv.conv_val_raw);
+ write_size = 1;
+
+#endif // LIBC_COPT_PRINTF_DISABLE_WIDE
+ } else {
+ buffer[0] = static_cast<char>(to_conv.conv_val_raw);
+ write_size = 1;
+ }
size_t padding_spaces =
- to_conv.min_width > STRING_LEN ? to_conv.min_width - STRING_LEN : 0;
+ to_conv.min_width > static_cast<int>(write_size) +
+ ? to_conv.min_width - static_cast<int>(write_size) +
+ : 0;
// If the padding is on the left side, write the spaces first.
if (padding_spaces > 0 &&
@@ -33,7 +72,7 @@ LIBC_INLINE int convert_char(Writer<write_mode> *writer,
RET_IF_RESULT_NEGATIVE(writer->write(' ', padding_spaces));
}
- RET_IF_RESULT_NEGATIVE(writer->write(c));
+ RET_IF_RESULT_NEGATIVE(writer->write({buffer, write_size}));
// If the padding is on the right side, write the spaces last.
if (padding_spaces > 0 &&
diff --git a/libc/src/stdio/printf_core/core_structs.h b/libc/src/stdio/printf_core/core_structs.h
index 0d41f2244d8da..d93fa962db90e 100644
--- a/libc/src/stdio/printf_core/core_structs.h
+++ b/libc/src/stdio/printf_core/core_structs.h
@@ -142,6 +142,8 @@ constexpr int INT_CONVERSION_ERROR = -1004;
constexpr int FIXED_POINT_CONVERSION_ERROR = -1005;
constexpr int ALLOCATION_ERROR = -1006;
constexpr int OVERFLOW_ERROR = -1007;
+constexpr int ILLEGAL_WIDE_CHAR = -1008;
+constexpr int MB_CONVERSION_ERROR = -1009;
} // namespace printf_core
} // namespace LIBC_NAMESPACE_DECL
diff --git a/libc/src/stdio/printf_core/linux/error_mapper.h b/libc/src/stdio/printf_core/linux/error_mapper.h
index 3c2fe663072d0..3449f12593469 100644
--- a/libc/src/stdio/printf_core/linux/error_mapper.h
+++ b/libc/src/stdio/printf_core/linux/error_mapper.h
@@ -40,6 +40,9 @@ LIBC_INLINE static int internal_error_to_errno(int internal_error) {
return ENOMEM;
case OVERFLOW_ERROR:
return EOVERFLOW;
+ case ILLEGAL_WIDE_CHAR:
+ case MB_CONVERSION_ERROR:
+ return EILSEQ;
default:
LIBC_ASSERT(
false &&
diff --git a/libc/src/stdio/printf_core/parser.h b/libc/src/stdio/printf_core/parser.h
index cef9b1ae58fa0..a3b62991bcec9 100644
--- a/libc/src/stdio/printf_core/parser.h
+++ b/libc/src/stdio/printf_core/parser.h
@@ -27,6 +27,9 @@
#ifndef LIBC_COPT_PRINTF_DISABLE_STRERROR
#include "src/__support/libc_errno.h"
#endif // LIBC_COPT_PRINTF_DISABLE_STRERROR
+#ifndef LIBC_COPT_PRINTF_DISABLE_WIDE
+#include "hdr/types/wint_t.h"
+#endif // LIBC_COPT_PRINTF_DISABLE_WIDE
namespace LIBC_NAMESPACE_DECL {
namespace printf_core {
@@ -73,9 +76,9 @@ template <typename ArgProvider> class Parser {
ArgProvider args_cur;
#ifndef LIBC_COPT_PRINTF_DISABLE_INDEX_MODE
- // args_start stores the start of the va_args, which is allows getting the
- // value of arguments that have already been passed. args_index is tracked so
- // that we know which argument args_cur is on.
+ // args_start stores the start of the va_args, which helps in getting the
+ // number of arguments that have already been passed. args_index is tracked
+ // so that we know which argument args_cur is on.
ArgProvider args_start;
size_t args_index = 1;
@@ -173,7 +176,17 @@ template <typename ArgProvider> class Parser {
section.has_conv = true;
break;
case ('c'):
- WRITE_ARG_VAL_SIMPLEST(section.conv_val_raw, int, conv_index);
+ if (section.length_modifier == LengthModifier::l) {
+#ifdef LIBC_COPT_PRINTF_DISABLE_WIDE
+ using WideCharArgType = int;
+#else
+ using WideCharArgType = wint_t;
+#endif // LIBC_COPT_PRINTF_DISABLE_WIDE
+ WRITE_ARG_VAL_SIMPLEST(section.conv_val_raw, WideCharArgType,
+ conv_index);
+ } else {
+ WRITE_ARG_VAL_SIMPLEST(section.conv_val_raw, int, conv_index);
+ }
break;
case ('d'):
case ('i'):
@@ -574,7 +587,16 @@ template <typename ArgProvider> class Parser {
conv_size = type_desc_from_type<void>();
break;
case ('c'):
- conv_size = type_desc_from_type<int>();
+ if (lm == LengthModifier::l) {
+#ifdef LIBC_COPT_PRINTF_DISABLE_WIDE
+ using WideCharArgType = int;
+#else
+ using WideCharArgType = wint_t;
+#endif // LIBC_COPT_PRINTF_DISABLE_WIDE
+ conv_size = type_desc_from_type<WideCharArgType>();
+ } else {
+ conv_size = type_desc_from_type<int>();
+ }
break;
case ('d'):
case ('i'):
diff --git a/libc/test/src/stdio/CMakeLists.txt b/libc/test/src/stdio/CMakeLists.txt
index a39428fb8d16c..fde2023caaac1 100644
--- a/libc/test/src/stdio/CMakeLists.txt
+++ b/libc/test/src/stdio/CMakeLists.txt
@@ -137,6 +137,15 @@ if(LIBC_CONF_PRINTF_DISABLE_STRERROR)
list(APPEND sprintf_test_copts "-DLIBC_COPT_PRINTF_DISABLE_STRERROR")
endif()
+if(LIBC_CONF_PRINTF_DISABLE_WIDE)
+ set(wchar_deps "")
+else()
+ set(wchar_deps
+ libc.hdr.types.wint_t
+ libc.hdr.wchar_macros
+ )
+endif()
+
add_fp_unittest(
sprintf_test
UNIT_TEST_ONLY
@@ -148,6 +157,7 @@ add_fp_unittest(
libc.src.stdio.sprintf
libc.src.__support.FPUtil.fp_bits
libc.include.inttypes
+ ${wchar_deps}
COMPILE_OPTIONS
${sprintf_test_copts}
)
diff --git a/libc/test/src/stdio/sprintf_test.cpp b/libc/test/src/stdio/sprintf_test.cpp
index 689a38a49f13c..78186abb9966e 100644
--- a/libc/test/src/stdio/sprintf_test.cpp
+++ b/libc/test/src/stdio/sprintf_test.cpp
@@ -9,8 +9,13 @@
#include "src/__support/macros/config.h"
#include "src/stdio/sprintf.h"
+#ifndef LIBC_COPT_PRINTF_DISABLE_WIDE
+#include "hdr/types/wint_t.h"
+#include "hdr/wchar_macros.h"
+#endif // LIBC_COPT_PRINTF_DISABLE_WIDE
+
#include "src/__support/FPUtil/FPBits.h"
-#include "src/__support/libc_errno.h"
+#include "test/UnitTest/ErrnoCheckingTest.h"
#include "test/UnitTest/RoundingModeUtils.h"
#include "test/UnitTest/Test.h"
#include <inttypes.h>
@@ -3487,3 +3492,94 @@ TEST(LlvmLibcSPrintfTest, IndexModeParsing) {
"why would u do this, this is such a pain. %");
}
#endif // LIBC_COPT_PRINTF_DISABLE_INDEX_MODE
+
+#ifndef LIBC_COPT_PRINTF_DISABLE_WIDE
+TEST(LlvmLibcSprintfTest, WideCharConversion) {
+ char buff[16];
+ int written;
+
+ // 1 byte UTF-8 character.
+ written = LIBC_NAMESPACE::sprintf(buff, "%lc", L'A');
+ EXPECT_EQ(written, 1);
+ ASSERT_STREQ_LEN(written, buff, "A");
+
+ // 1 byte UTF-8 character left justified.
+ written = LIBC_NAMESPACE::sprintf(buff, "%-4lc", L'A');
+ EXPECT_EQ(written, 4);
+ ASSERT_STREQ_LEN(written, buff, "A ");
+
+ // 1 byte UTF-8 character right justified.
+ written = LIBC_NAMESPACE::sprintf(buff, "%4lc", L'A');
+ EXPECT_EQ(written, 4);
+ ASSERT_STREQ_LEN(written, buff, " A");
+
+ // 2 byte UTF-8 character.
+ written = LIBC_NAMESPACE::sprintf(buff, "%lc", L'¢');
+ EXPECT_EQ(written, 2);
+ ASSERT_STREQ_LEN(written, buff, "¢");
+
+ // 2 byte UTF-8 character left justified.
+ written = LIBC_NAMESPACE::sprintf(buff, "%-4lc", L'¢');
+ EXPECT_EQ(written, 4);
+ ASSERT_STREQ_LEN(written, buff, "¢ ");
+
+ // 2 byte UTF-8 character right justified.
+ written = LIBC_NAMESPACE::sprintf(buff, "%4lc", L'¢');
+ EXPECT_EQ(written, 4);
+ ASSERT_STREQ_LEN(written, buff, " ¢");
+
+ // Euro sign is a 3-byte UTF-8 character.
+ written = LIBC_NAMESPACE::sprintf(buff, "%lc", L'€');
+ EXPECT_EQ(written, 3);
+ ASSERT_STREQ_LEN(written, buff, "€");
+
+ // Euro sign left justified.
+ written = LIBC_NAMESPACE::sprintf(buff, "%-4lc", L'€');
+ EXPECT_EQ(written, 4);
+ ASSERT_STREQ_LEN(written, buff, "€ ");
+
+ // Euro sign right justified.
+ written = LIBC_NAMESPACE::sprintf(buff, "%4lc", L'€');
+ EXPECT_EQ(written, 4);
+ ASSERT_STREQ_LEN(written, buff, " €");
+
+ // Euro sign right justified.
+ written = LIBC_NAMESPACE::sprintf(buff, "%+4lc", L'€');
+ EXPECT_EQ(written, 4);
+ ASSERT_STREQ_LEN(written, buff, " €");
+
+ // Grinning face emoji is a 4-byte UTF-8 character.
+ written = LIBC_NAMESPACE::sprintf(buff, "%lc", L'😀');
+ EXPECT_EQ(written, 4);
+ ASSERT_STREQ_LEN(written, buff, "😀");
+
+ // Grinning face emoji left justified.
+ written = LIBC_NAMESPACE::sprintf(buff, "%-4lc", L'😀');
+ EXPECT_EQ(written, 4);
+ ASSERT_STREQ_LEN(written, buff, "😀");
+
+ // Grinning face emoji right justified.
+ written = LIBC_NAMESPACE::sprintf(buff, "%4lc", L'😀');
+ EXPECT_EQ(written, 4);
+ ASSERT_STREQ_LEN(written, buff, "😀");
+
+ // Grinning face emoji with smaller width, left justified.
+ written = LIBC_NAMESPACE::sprintf(buff, "%-3lc", L'😀');
+ EXPECT_EQ(written, 4);
+ ASSERT_STREQ_LEN(written, buff, "😀");
+
+ // Grinning face emoji with smaller width, right justified.
+ written = LIBC_NAMESPACE::sprintf(buff, "%3lc", L'😀');
+ EXPECT_EQ(written, 4);
+ ASSERT_STREQ_LEN(written, buff, "😀");
+
+ // WEOF test.
+ EXPECT_EQ(LIBC_NAMESPACE::sprintf(buff, "%lc", WEOF), -1);
+ ASSERT_ERRNO_EQ(EILSEQ);
+
+ // Invalid wide character test
+ EXPECT_EQ(LIBC_NAMESPACE::sprintf(buff, "%lc", static_cast<wint_t>(0x12ffff)),
+ -1);
+ ASSERT_ERRNO_EQ(EILSEQ);
+}
+#endif // LIBC_COPT_PRINTF_DISABLE_WIDE
More information about the libc-commits
mailing list