[libc] Support %lc in printf

Add support for %lc in printf by calling internal
wcrtomb function and relevant end-to-end snprintf
test. Additionally, printf parser can also
recognize length modifier. Added a flag to disable
wchar support on windows platform.

Author: shubh@DOE

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.

libc/src/stdio/printf_core/CMakeLists.txt

@@ -43,6 +43,14 @@ if(NOT TARGET ${target_error_mapper})
     set(target_error_mapper libc.src.stdio.printf_core.generic.error_mapper)
 endif()
 
+set(wchar_deps libc.src.__support.wchar.wcrtomb)
+set(wchar_compile_flags "")
+
+if(${LIBC_TARGET_OS} STREQUAL "windows" OR WIN32)
+  set(wchar_deps "")
+  list(APPEND wchar_compile_flags "-DLIBC_COPT_PRINTF_DISABLE_WIDE")
+endif()
+
 add_header_library(
   printf_config
   HDRS
@@ -67,6 +75,7 @@ add_header_library(
     parser.h
   DEPENDS
     .core_structs
+    libc.hdr.types.wint_t
     libc.src.__support.arg_list
     libc.src.__support.ctype_utils
     libc.src.__support.str_to_integer
@@ -106,11 +115,16 @@ add_header_library(
     float_dec_converter.h
     fixed_converter.h #TODO: Check if this should be disabled when fixed unavail
     strerror_converter.h
+  COMPILE_OPTIONS
+    ${wchar_compile_flags}
   DEPENDS
     .core_structs
     .printf_config
     .writer
     libc.include.inttypes
+    libc.hdr.types.wchar_t
+    libc.hdr.types.wint_t
+    libc.hdr.wchar_macros
     libc.src.__support.big_int
     libc.src.__support.common
     libc.src.__support.CPP.limits
@@ -123,8 +137,10 @@ add_header_library(
     libc.src.__support.integer_to_string
     libc.src.__support.libc_assert
     libc.src.__support.uint128
+    libc.src.__support.wchar.mbstate
     libc.src.__support.StringUtil.error_to_string
     libc.src.string.memory_utils.inline_memcpy
+    ${wchar_deps}
 )
 
 add_header_library(

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
 
+#include "hdr/types/wchar_t.h"
+#include "hdr/types/wint_t.h"
+#include "hdr/wchar_macros.h"
+#include "src/__support/wchar/mbstate.h"
+
+#ifndef LIBC_COPT_PRINTF_DISABLE_WIDE
+#include "src/__support/wchar/wcrtomb.h"
+#endif // LIBC_COPT_PRINTF_DISABLE_WIDE
+
 #include "src/__support/macros/config.h"
 #include "src/stdio/printf_core/converter_utils.h"
 #include "src/stdio/printf_core/core_structs.h"
@@ -33,7 +42,28 @@ LIBC_INLINE int convert_char(Writer<write_mode> *writer,
     RET_IF_RESULT_NEGATIVE(writer->write(' ', padding_spaces));
   }
 
-  RET_IF_RESULT_NEGATIVE(writer->write(c));
+  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 -1;
+    }
+
+    char mb_str[MB_LEN_MAX];
+    internal::mbstate mbstate;
+    wchar_t wc = static_cast<wchar_t>(wi);
+
+    auto ret = internal::wcrtomb(mb_str, wc, &mbstate);
+    if (!ret.has_value()) {
+      return -1;
+    }
+
+    RET_IF_RESULT_NEGATIVE(writer->write({mb_str, ret.value()}));
+#endif // LIBC_COPT_PRINTF_DISABLE_WIDE
+  } else {
+    RET_IF_RESULT_NEGATIVE(writer->write(c));
+  }
 
   // If the padding is on the right side, write the spaces last.
   if (padding_spaces > 0 &&

libc/src/stdio/printf_core/parser.h

@@ -9,6 +9,7 @@
 #ifndef LLVM_LIBC_SRC_STDIO_PRINTF_CORE_PARSER_H
 #define LLVM_LIBC_SRC_STDIO_PRINTF_CORE_PARSER_H
 
+#include "hdr/types/wint_t.h"
 #include "include/llvm-libc-macros/stdfix-macros.h"
 #include "src/__support/CPP/algorithm.h" // max
 #include "src/__support/CPP/limits.h"
@@ -73,9 +74,9 @@ 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 +174,11 @@ 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) {
+          WRITE_ARG_VAL_SIMPLEST(section.conv_val_raw, wint_t, conv_index);
+        } else {
+          WRITE_ARG_VAL_SIMPLEST(section.conv_val_raw, int, conv_index);
+        }
         break;
       case ('d'):
       case ('i'):

libc/test/src/stdio/CMakeLists.txt

@@ -137,6 +137,11 @@ if(LIBC_CONF_PRINTF_DISABLE_STRERROR)
   list(APPEND sprintf_test_copts "-DLIBC_COPT_PRINTF_DISABLE_STRERROR")
 endif()
 
+set(snprintf_test_flags "")
+if(${LIBC_TARGET_OS} STREQUAL "windows" OR WIN32)
+  list(APPEND snprintf_test_flags "-DLIBC_COPT_PRINTF_DISABLE_WIDE")
+endif()
+
 add_fp_unittest(
   sprintf_test
   UNIT_TEST_ONLY
@@ -158,7 +163,11 @@ add_libc_test(
     libc_stdio_unittests
   SRCS
     snprintf_test.cpp
+  COMPILE_FLAGS
+    ${snprintf_test_flags}
   DEPENDS
+    libc.hdr.types.wchar_t
+    libc.hdr.wchar_macros
     libc.src.stdio.snprintf
 )
 

libc/test/src/stdio/printf_core/CMakeLists.txt

@@ -1,3 +1,8 @@
+set(printf_test_flags "")
+if(${LIBC_TARGET_OS} STREQUAL "windows" OR WIN32)
+  list(APPEND printf_test_flags "-DLIBC_COPT_PRINTF_DISABLE_WIDE")
+endif()
+
 add_libc_unittest(
   parser_test
   SUITE
@@ -6,7 +11,10 @@ add_libc_unittest(
     parser_test.cpp
   LINK_LIBRARIES
     LibcPrintfHelpers
+  COMPILE_FLAGS
+    ${printf_test_flags}
   DEPENDS
+    libc.hdr.types.wchar_t
     libc.src.stdio.printf_core.parser
     libc.src.stdio.printf_core.core_structs
     libc.src.__support.CPP.string_view
@@ -19,6 +27,8 @@ add_libc_unittest(
     libc_stdio_unittests
   SRCS
     writer_test.cpp
+  COMPILE_FLAGS
+    ${printf_test_flags}
   DEPENDS
     libc.src.stdio.printf_core.writer
     libc.src.string.memory_utils.inline_memcpy
@@ -31,6 +41,8 @@ add_libc_unittest(
     libc_stdio_unittests
   SRCS
     converter_test.cpp
+  COMPILE_FLAGS
+    ${printf_test_flags}
   DEPENDS
     libc.src.stdio.printf_core.converter
     libc.src.stdio.printf_core.writer

libc/test/src/stdio/printf_core/parser_test.cpp

@@ -6,6 +6,7 @@
 //
 //===----------------------------------------------------------------------===//
 
+#include "hdr/types/wchar_t.h"
 #include "src/__support/CPP/bit.h"
 #include "src/__support/CPP/string_view.h"
 #include "src/__support/arg_list.h"
@@ -370,6 +371,24 @@ TEST(LlvmLibcPrintfParserTest,
   ASSERT_PFORMAT_EQ(expected, format_arr[0]);
 }
 
+TEST(LlvmLibcPrintfParserTest, EvalOneArgWithWideCharacter) {
+  LIBC_NAMESPACE::printf_core::FormatSection format_arr[2];
+  const char *str = "%lc";
+  wchar_t arg1 = L'€';
+  evaluate(format_arr, str, arg1);
+
+  LIBC_NAMESPACE::printf_core::FormatSection expected;
+  expected.has_conv = true;
+
+  expected.raw_string = {str, 3};
+  expected.length_modifier = LIBC_NAMESPACE::printf_core::LengthModifier::l;
+  expected.conv_val_raw =
+      static_cast<LIBC_NAMESPACE::fputil::FPBits<double>::StorageType>(arg1);
+  expected.conv_name = 'c';
+
+  ASSERT_PFORMAT_EQ(expected, format_arr[0]);
+}
+
 #ifndef LIBC_COPT_PRINTF_DISABLE_INDEX_MODE
 
 TEST(LlvmLibcPrintfParserTest, IndexModeOneArg) {

libc/test/src/stdio/snprintf_test.cpp

@@ -6,6 +6,8 @@
 //
 //===----------------------------------------------------------------------===//
 
+#include "hdr/types/wchar_t.h"
+#include "hdr/wchar_macros.h"
 #include "src/stdio/snprintf.h"
 
 #include "test/UnitTest/ErrnoCheckingTest.h"
@@ -74,3 +76,45 @@ TEST(LlvmLibcSNPrintfTest, CharsWrittenOverflow) {
   EXPECT_LT(written, 0);
   ASSERT_ERRNO_FAILURE();
 }
+
+#ifndef LIBC_COPT_PRINTF_DISABLE_WIDE
+TEST(LlvmLibcSNPrintfTest, WideCharConversion) {
+  char buff[16];
+  int written;
+
+  written = LIBC_NAMESPACE::snprintf(buff, sizeof(buff), "%lc",
+                                     static_cast<wchar_t>(L'€'));
+  EXPECT_EQ(written, 3);
+  ASSERT_STREQ(buff, "€");
+}
+
+TEST(LlvmLibcSNPrintfTest, WideCharConversionLeftJustified) {
+  char buff[16];
+  int written;
+
+  written = LIBC_NAMESPACE::snprintf(buff, sizeof(buff), "%-4lc",
+                                     static_cast<wchar_t>(L'€'));
+  EXPECT_EQ(written, 6);
+  ASSERT_STREQ(buff, "€   ");
+}
+
+TEST(LlvmLibcSNPrintfTest, WideCharConversionRightJustified) {
+  char buff[16];
+  int written;
+
+  written = LIBC_NAMESPACE::snprintf(buff, sizeof(buff), "%4lc",
+                                     static_cast<wchar_t>(L'€'));
+  EXPECT_EQ(written, 6);
+  ASSERT_STREQ(buff, "   €");
+}
+
+TEST(LlvmLibcSNPrintfTest, WideCharWEOFConversion) {
+  char buff[16];
+  int written;
+
+  written = LIBC_NAMESPACE::snprintf(buff, sizeof(buff), "%lc",
+                                     static_cast<wchar_t>(WEOF));
+  EXPECT_EQ(written, -1);
+  ASSERT_ERRNO_FAILURE();
+}
+#endif // LIBC_COPT_PRINTF_DISABLE_WIDE