[libc] add scanf reader

[michaelrj]

Summary

This is the interface that will be used to read from a file or string in
scanf. This patch also adds the string and file implementations of the
reader, although the file reader is not yet complete since ungetc has
not yet been implemented.

[michaelrj]

add guards for systems that don't support FILE

[sivachandra]

The code LGTM but hard to comment on the design without seeing how they are used.

[michaelrj]

adjust unget behavior to match the standard method.

[sivachandra]

LGTM with minor fixes.

libc/src/stdio/scanf_core/reader.h

38

There could be a valid \0 character in the input stream before EOF. Will this cause any problems? Either way, this seems to me like the method should return a cpp::optional value?

42

This comment should be next to where the ignore is actually happening and explain why.

[michaelrj]

add comment explaining ungetc behavior

libc/src/stdio/scanf_core/reader.h

38

a \0 in a file will be treated the same as a \0 in a regular string; no scanf formats are allowed to cross it. As for if this should be optional, I don't think there should be any difference between the end of a file and the end of a string, so representing them both with a '\0' seems a lot simpler than adding machinery everywhere characters are read to handle optionals.

[michaelrj]

adjust comment before landing

libc/src/stdio/scanf_core/CMakeLists.txt

DEPENDS
     .core_structs
     libc.src.__support.arg_list
     libc.src.__support.ctype_utils
     libc.src.__support.str_to_integer
     libc.src.__support.CPP.bit
     libc.src.__support.CPP.bitset
     libc.src.__support.CPP.string_view
 )
+
+if(NOT (TARGET libc.src.__support.File.file))
+  # Not all platforms have a file implementation. If file is unvailable,
+  # then we must skip all the parts that need file.
+  return()
+endif()
+
+add_object_library(
+  string_reader
+  SRCS
+    string_reader.cpp
+  HDRS
+    string_reader.h
+)
+
+add_object_library(
+  file_reader
+  SRCS
+    file_reader.cpp
+  HDRS
+    file_reader.h
+  DEPENDS
+    libc.src.__support.File.file
+)
+
+add_object_library(
+  reader
+  SRCS
+    reader.cpp
+  HDRS
+    reader.h
+  DEPENDS
+    .string_reader
+    .file_reader
+)

libc/src/stdio/scanf_core/file_reader.h

+//===-- FILE Reader definition for scanf ------------------------*- 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_LIBC_SRC_STDIO_SCANF_CORE_FILE_READER_H
+#define LLVM_LIBC_SRC_STDIO_SCANF_CORE_FILE_READER_H
+
+#include "src/__support/File/file.h"
+
+#include <stddef.h>
+#include <stdio.h>
+
+namespace __llvm_libc {
+namespace scanf_core {
+
+class FileReader {
+  __llvm_libc::File *file;
+
+public:
+  FileReader(::FILE *init_file) {
+    file = reinterpret_cast<__llvm_libc::File *>(init_file);
+    file->lock();
+  }
+
+  ~FileReader() { file->unlock(); }
+
+  char get_char();
+  void unget_char(char c);
+};
+
+} // namespace scanf_core
+} // namespace __llvm_libc
+
+#endif // LLVM_LIBC_SRC_STDIO_SCANF_CORE_FILE_READER_H

libc/src/stdio/scanf_core/file_reader.cpp

+//===-- FILE Reader implementation for scanf --------------------*- 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
+//
+//===----------------------------------------------------------------------===//
+
+#include "src/stdio/scanf_core/file_reader.h"
+#include "src/__support/File/file.h"
+#include <stddef.h>
+
+namespace __llvm_libc {
+namespace scanf_core {
+
+char FileReader::get_char() {
+  char tiny_buff = 0;
+  if (file->read_unlocked(&tiny_buff, 1) != 1)
+    return 0;
+  return tiny_buff;
+}
+
+void FileReader::unget_char(char c) { file->ungetc_unlocked(c); }
+
+} // namespace scanf_core
+} // namespace __llvm_libc

libc/src/stdio/scanf_core/reader.h

+//===-- Reader definition for scanf -----------------------------*- 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_LIBC_SRC_STDIO_SCANF_CORE_READER_H
+#define LLVM_LIBC_SRC_STDIO_SCANF_CORE_READER_H
+
+#include "src/stdio/scanf_core/file_reader.h"
+#include "src/stdio/scanf_core/string_reader.h"
+#include <stddef.h>
+
+namespace __llvm_libc {
+namespace scanf_core {
+
+enum class ReaderType { String, File };
+
+class Reader final {
+  union {
+    StringReader *string_reader;
+    FileReader *file_reader;
+  };
+
+  const ReaderType reader_type;
+
+public:
+  Reader(StringReader *init_string_reader)
+      : string_reader(init_string_reader), reader_type(ReaderType::String) {}
+
+  Reader(FileReader *init_file_reader)
+      : file_reader(init_file_reader), reader_type(ReaderType::File) {}
+
+  // This returns the next character from the input and advances it by one
+  // character. When it hits the end of the string or file it returns '\0' to
+  // signal to stop parsing.
+  char getc();
+
+  // This moves the input back by one character, placing c into the buffer if
+  // this is a file reader, else c is ignored.
+  void ungetc(char c);
+};
+
+} // namespace scanf_core
+} // namespace __llvm_libc
+
+#endif // LLVM_LIBC_SRC_STDIO_SCANF_CORE_READER_H

libc/src/stdio/scanf_core/reader.cpp

+//===-- Reader definition for scanf -----------------------------*- 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
+//
+//===----------------------------------------------------------------------===//
+
+#include "src/stdio/scanf_core/reader.h"
+#include <stddef.h>
+
+namespace __llvm_libc {
+namespace scanf_core {
+
+char Reader::getc() {
+  if (reader_type == ReaderType::String) {
+    return string_reader->get_char();
+  } else {
+    return file_reader->get_char();
+  }
+}
+
+void Reader::ungetc(char c) {
+  if (reader_type == ReaderType::String) {
+    // The string reader ignores the char c passed to unget since it doesn't
+    // need to place anything back into a buffer, and modifying the source
+    // string would be dangerous.
+    return string_reader->unget_char();
+  } else {
+    return file_reader->unget_char(c);
+  }
+}
+
+} // namespace scanf_core
+} // namespace __llvm_libc

libc/src/stdio/scanf_core/string_reader.h

+//===-- String Reader definition for scanf ----------------------*- 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_LIBC_SRC_STDIO_SCANF_CORE_STRING_READER_H
+#define LLVM_LIBC_SRC_STDIO_SCANF_CORE_STRING_READER_H
+
+#include <stddef.h>
+
+namespace __llvm_libc {
+namespace scanf_core {
+
+class StringReader {
+  const char *string;
+  size_t cur_index = 0;
+
+public:
+  StringReader(const char *init_string) { string = init_string; }
+
+  ~StringReader() {}
+
+  char get_char();
+  void unget_char();
+};
+
+} // namespace scanf_core
+} // namespace __llvm_libc
+
+#endif // LLVM_LIBC_SRC_STDIO_SCANF_CORE_STRING_READER_H

libc/src/stdio/scanf_core/string_reader.cpp

+//===-- String Reader implementation for scanf ------------------*- 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
+//
+//===----------------------------------------------------------------------===//
+
+#include "src/stdio/scanf_core/string_reader.h"
+#include <stddef.h>
+
+namespace __llvm_libc {
+namespace scanf_core {
+
+char StringReader::get_char() {
+  char cur_char = string[cur_index];
+  ++cur_index;
+  return cur_char;
+}
+
+void StringReader::unget_char() { --cur_index; }
+
+} // namespace scanf_core
+} // namespace __llvm_libc

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

 add_libc_unittest(
   parser_test
   SUITE
     libc_stdio_unittests
   SRCS
     parser_test.cpp
   LINK_LIBRARIES
     LibcScanfHelpers
   DEPENDS
     libc.src.stdio.scanf_core.parser
     libc.src.stdio.scanf_core.core_structs
     libc.src.__support.CPP.string_view
     libc.src.__support.arg_list
 )
+
+if(NOT (TARGET libc.src.__support.File.file))
+  # Not all platforms have a file implementation. If file is unvailable,
+  # then we must skip all the parts that need file.
+  return()
+endif()
+
+add_libc_unittest(
+  string_reader_test
+  SUITE
+    libc_stdio_unittests
+  SRCS
+    string_reader_test.cpp
+  DEPENDS
+    libc.src.stdio.scanf_core.reader
+    libc.src.stdio.scanf_core.string_reader
+    libc.src.__support.CPP.string_view
+)

libc/test/src/stdio/scanf_core/string_reader_test.cpp

+//===-- Unittests for the scanf String Reader -----------------------------===//
+//
+// 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 "src/__support/CPP/string_view.h"
+#include "src/stdio/scanf_core/reader.h"
+#include "src/stdio/scanf_core/string_reader.h"
+
+#include "utils/UnitTest/Test.h"
+
+TEST(LlvmLibcScanfStringReaderTest, Constructor) {
+  char str[10];
+  __llvm_libc::scanf_core::StringReader str_reader(str);
+  __llvm_libc::scanf_core::Reader reader(&str_reader);
+}
+
+TEST(LlvmLibcScanfStringReaderTest, SimpleRead) {
+  const char *str = "abc";
+  __llvm_libc::scanf_core::StringReader str_reader(str);
+  __llvm_libc::scanf_core::Reader reader(&str_reader);
+
+  for (size_t i = 0; i < sizeof(str); ++i) {
+    ASSERT_EQ(str[i], reader.getc());
+  }
+}
+
+TEST(LlvmLibcScanfStringReaderTest, ReadAndReverse) {
+  const char *str = "abcDEF123";
+  __llvm_libc::scanf_core::StringReader str_reader(str);
+  __llvm_libc::scanf_core::Reader reader(&str_reader);
+
+  for (size_t i = 0; i < 5; ++i) {
+    ASSERT_EQ(str[i], reader.getc());
+  }
+
+  // Move back by 3, cursor should now be on 2
+  reader.ungetc(str[4]);
+  reader.ungetc(str[3]);
+  reader.ungetc(str[2]);
+
+  for (size_t i = 2; i < 7; ++i) {
+    ASSERT_EQ(str[i], reader.getc());
+  }
+
+  // Move back by 2, cursor should now be on 5
+  reader.ungetc(str[6]);
+  reader.ungetc(str[5]);
+
+  for (size_t i = 5; i < 10; ++i) {
+    ASSERT_EQ(str[i], reader.getc());
+  }
+
+  // Move back by 10, which should be back to the start.
+  for (size_t i = 0; i < 10; ++i) {
+    reader.ungetc(str[9 - i]);
+  }
+
+  // Check the whole string.
+  for (size_t i = 0; i < sizeof(str); ++i) {
+    ASSERT_EQ(str[i], reader.getc());
+  }
+}