[libc] Add implementations for isblank, iscntrl, isgraph, ispunct.

[sivachandra]

libc/src/ctype/ctype_utils.h

29

I should have said said this earlier: seems to me like all the functions in cytype_utils.h can take const unsigned char argument.

33

Wouldn't it read better this way:

return 0x20 < ch && ch < 0x7f;

Obvious to check that it matches exactly what the standard says.

[cgyurgyik]

Address siva's comments.

[cgyurgyik]

libc/src/ctype/ctype_utils.h

29

Yeah much cleaner thanks.

libc/config/linux/aarch64/entrypoints.txt

 set(TARGET_LIBC_ENTRYPOINTS
     # ctype.h entrypoints
     libc.src.ctype.isalnum
     libc.src.ctype.isalpha
+    libc.src.ctype.isblank
+    libc.src.ctype.iscntrl
     libc.src.ctype.isdigit
+    libc.src.ctype.isgraph
     libc.src.ctype.islower
+    libc.src.ctype.ispunct
     libc.src.ctype.isupper
     
     # errno.h entrypoints
     libc.src.errno.__errno_location
 
     # string.h entrypoints
     libc.src.string.bzero
     libc.src.string.memchr

libc/config/linux/api.td

let Macros = [
     StaticAssertMacro,
   ];
 }
 
 def CTypeAPI : PublicAPI<"ctype.h"> {
   let Functions = [
     "isalnum",
     "isalpha",
+    "isblank",
+    "iscntrl",
     "isdigit",
+    "isgraph",
     "islower",
+    "ispunct",
     "isupper",
   ];
 }
 
 def MathErrHandlingMacro : MacroDef<"math_errhandling"> {
   let Defn = [{
     #ifndef math_errhandling
     #ifdef __FAST_MATH__

libc/config/linux/x86_64/entrypoints.txt

 set(TARGET_LIBC_ENTRYPOINTS
     # assert.h entrypoints
     libc.src.assert.__assert_fail
     
     # ctype.h entrypoints
     libc.src.ctype.isalnum
     libc.src.ctype.isalpha
+    libc.src.ctype.isblank
+    libc.src.ctype.iscntrl
     libc.src.ctype.isdigit
+    libc.src.ctype.isgraph
     libc.src.ctype.islower
+    libc.src.ctype.ispunct
     libc.src.ctype.isupper
 
     # errno.h entrypoints
     libc.src.errno.__errno_location
 
     # signal.h entrypoints
     libc.src.signal.raise
     libc.src.signal.sigaction

libc/spec/stdc.td

HeaderSpec CType = HeaderSpec<
               [ArgSpec<IntType>]
           >,
           FunctionSpec<
               "isalpha",
               RetValSpec<IntType>,
               [ArgSpec<IntType>]
           >,
           FunctionSpec<
+              "isblank",
+              RetValSpec<IntType>,
+              [ArgSpec<IntType>]
+          >,
+          FunctionSpec<
+              "iscntrl",
+              RetValSpec<IntType>,
+              [ArgSpec<IntType>]
+          >,
+          FunctionSpec<
               "isdigit",
               RetValSpec<IntType>,
               [ArgSpec<IntType>]
           >,
           FunctionSpec<
+              "isgraph",
+              RetValSpec<IntType>,
+              [ArgSpec<IntType>]
+          >,
+          FunctionSpec<
               "islower",
               RetValSpec<IntType>,
               [ArgSpec<IntType>]
           >,
           FunctionSpec<
+              "ispunct",
+              RetValSpec<IntType>,
+              [ArgSpec<IntType>]
+          >,
+          FunctionSpec<
               "isupper",
               RetValSpec<IntType>,
               [ArgSpec<IntType>]
           >,
       ]
   >;
      
   HeaderSpec String = HeaderSpec<

libc/src/ctype/CMakeLists.txt

SRCS
     isalpha.cpp
   HDRS
     isalpha.h
   DEPENDS
     .ctype_utils
 )
 
 add_entrypoint_object(
+  isblank
+  SRCS
+    isblank.cpp
+  HDRS
+    isblank.h
+)
+
+add_entrypoint_object(
+  iscntrl
+  SRCS
+    iscntrl.cpp
+  HDRS
+    iscntrl.h
+)
+
+add_entrypoint_object(
   isdigit
   SRCS
     isdigit.cpp
   HDRS
     isdigit.h
   DEPENDS
     .ctype_utils
 )
 
 add_entrypoint_object(
+  isgraph
+  SRCS
+    isgraph.cpp
+  HDRS
+    isgraph.h
+  DEPENDS
+    .ctype_utils
+)
+
+add_entrypoint_object(
   islower
   SRCS
     islower.cpp
   HDRS
     islower.h
 )
 
 add_entrypoint_object(
+  ispunct
+  SRCS
+    ispunct.cpp
+  HDRS
+    ispunct.h
+  DEPENDS
+    .ctype_utils
+)
+
+add_entrypoint_object(
   isupper
   SRCS
     isupper.cpp
   HDRS
     isupper.h
 )

libc/src/ctype/ctype_utils.h

 namespace internal {
 
 // ------------------------------------------------------
 // Rationale: Since these classification functions are
 // called in other functions, we will avoid the overhead
 // of a function call by inlining them.
 // ------------------------------------------------------
 
-static inline int isdigit(int c) {
-  const unsigned ch = c;
-  return (ch - '0') < 10;
-}
-
-static inline int isalpha(int c) {
-  const unsigned ch = c;
-  return (ch | 32) - 'a' < 26;
-}
+static inline int isdigit(unsigned ch) { return (ch - '0') < 10; }
+
+static inline int isalpha(unsigned ch) { return (ch | 32) - 'a' < 26; }
+
+static inline int isalnum(unsigned ch) { return isalpha(ch) || isdigit(ch); }
+
+static inline int isgraph(unsigned ch) { return 0x20 < ch && ch < 0x7f; }
 
 } // namespace internal
 } // namespace __llvm_libc
 
 #endif //  LLVM_LIBC_SRC_CTYPE_CTYPE_UTILS_H

libc/src/ctype/isalnum.cpp

 #include "src/ctype/ctype_utils.h"
 
 #include "src/__support/common.h"
 
 namespace __llvm_libc {
 
 // TODO: Currently restricted to default locale.
 // These should be extended using locale information.
-int LLVM_LIBC_ENTRYPOINT(isalnum)(int c) {
-  return internal::isalpha(c) || internal::isdigit(c);
-}
+int LLVM_LIBC_ENTRYPOINT(isalnum)(int c) { return internal::isalnum(c); }
 
 } // namespace __llvm_libc

libc/src/ctype/isblank.h

+//===-- Implementation header for isblank -------------------------*-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_CTYPE_ISBLANK_H
+#define LLVM_LIBC_SRC_CTYPE_ISBLANK_H
+
+namespace __llvm_libc {
+
+int isblank(int c);
+
+} // namespace __llvm_libc
+
+#endif //  LLVM_LIBC_SRC_CTYPE_ISBLANK_H

libc/src/ctype/isblank.cpp

-//===-- Implementation of isalnum------------------------------------------===//
+//===-- Implementation of isblank------------------------------------------===//
 //
 // 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/ctype/isalnum.h"
-#include "src/ctype/ctype_utils.h"
+#include "src/ctype/isblank.h"
 
 #include "src/__support/common.h"
 
 namespace __llvm_libc {
 
 // TODO: Currently restricted to default locale.
 // These should be extended using locale information.
-int LLVM_LIBC_ENTRYPOINT(isalnum)(int c) {
-  return internal::isalpha(c) || internal::isdigit(c);
+int LLVM_LIBC_ENTRYPOINT(isblank)(int c) {
+  const unsigned char ch = c;
+  return ch == ' ' || ch == '\t';
 }
 
 } // namespace __llvm_libc

libc/src/ctype/iscntrl.h

+//===-- Implementation header for iscntrl -------------------------*-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_CTYPE_ISCNTRL_H
+#define LLVM_LIBC_SRC_CTYPE_ISCNTRL_H
+
+namespace __llvm_libc {
+
+int iscntrl(int c);
+
+} // namespace __llvm_libc
+
+#endif //  LLVM_LIBC_SRC_CTYPE_ISCNTRL_H

libc/src/ctype/iscntrl.cpp

-//===-- Implementation of isalnum------------------------------------------===//
+//===-- Implementation of iscntrl------------------------------------------===//
 //
 // 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/ctype/isalnum.h"
-#include "src/ctype/ctype_utils.h"
+#include "src/ctype/iscntrl.h"
 
 #include "src/__support/common.h"
 
 namespace __llvm_libc {
 
 // TODO: Currently restricted to default locale.
 // These should be extended using locale information.
-int LLVM_LIBC_ENTRYPOINT(isalnum)(int c) {
-  return internal::isalpha(c) || internal::isdigit(c);
+int LLVM_LIBC_ENTRYPOINT(iscntrl)(int c) {
+  const unsigned char ch = c;
+  return ch < 0x20 || ch == 0x7f;
 }
 
 } // namespace __llvm_libc

libc/src/ctype/isgraph.h

+//===-- Implementation header for isgraph -------------------------*-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_CTYPE_ISGRAPH_H
+#define LLVM_LIBC_SRC_CTYPE_ISGRAPH_H
+
+namespace __llvm_libc {
+
+int isgraph(int c);
+
+} // namespace __llvm_libc
+
+#endif //  LLVM_LIBC_SRC_CTYPE_ISGRAPH_H

libc/src/ctype/isgraph.cpp

-//===-- Implementation of isalnum------------------------------------------===//
+//===-- Implementation of isgraph------------------------------------------===//
 //
 // 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/ctype/isalnum.h"
-#include "src/ctype/ctype_utils.h"
+#include "src/ctype/isgraph.h"
 
 #include "src/__support/common.h"
+#include "src/ctype/ctype_utils.h"
 
 namespace __llvm_libc {
 
 // TODO: Currently restricted to default locale.
 // These should be extended using locale information.
-int LLVM_LIBC_ENTRYPOINT(isalnum)(int c) {
-  return internal::isalpha(c) || internal::isdigit(c);
-}
+int LLVM_LIBC_ENTRYPOINT(isgraph)(int c) { return internal::isgraph(c); }
 
 } // namespace __llvm_libc

libc/src/ctype/ispunct.h

+//===-- Implementation header for ispunct -------------------------*-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_CTYPE_ISPUNCT_H
+#define LLVM_LIBC_SRC_CTYPE_ISPUNCT_H
+
+namespace __llvm_libc {
+
+int ispunct(int c);
+
+} // namespace __llvm_libc
+
+#endif //  LLVM_LIBC_SRC_CTYPE_ISPUNCT_H

libc/src/ctype/ispunct.cpp

-//===-- Implementation of isalnum------------------------------------------===//
+//===-- Implementation of ispunct------------------------------------------===//
 //
 // 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/ctype/isalnum.h"
-#include "src/ctype/ctype_utils.h"
+#include "src/ctype/ispunct.h"
 
 #include "src/__support/common.h"
+#include "src/ctype/ctype_utils.h"
 
 namespace __llvm_libc {
 
 // TODO: Currently restricted to default locale.
 // These should be extended using locale information.
-int LLVM_LIBC_ENTRYPOINT(isalnum)(int c) {
-  return internal::isalpha(c) || internal::isdigit(c);
+int LLVM_LIBC_ENTRYPOINT(ispunct)(int c) {
+  return !internal::isalnum(c) && internal::isgraph(c);
 }
 
 } // namespace __llvm_libc

libc/test/src/ctype/CMakeLists.txt

SUITE
     libc_ctype_unittests
   SRCS
     isalpha_test.cpp
   DEPENDS
     libc.src.ctype.isalpha
 )
 
 add_libc_unittest(
+  isblank
+  SUITE
+    libc_ctype_unittests
+  SRCS
+    isblank_test.cpp
+  DEPENDS
+    libc.src.ctype.isblank
+)
+
+add_libc_unittest(
+  iscntrl
+  SUITE
+    libc_ctype_unittests
+  SRCS
+    iscntrl_test.cpp
+  DEPENDS
+    libc.src.ctype.iscntrl
+)
+
+add_libc_unittest(
   isdigit
   SUITE
     libc_ctype_unittests
   SRCS
     isdigit_test.cpp
   DEPENDS
     libc.src.ctype.isdigit
 )
 
 add_libc_unittest(
+  isgraph
+  SUITE
+    libc_ctype_unittests
+  SRCS
+    isgraph_test.cpp
+  DEPENDS
+    libc.src.ctype.isgraph
+)
+
+add_libc_unittest(
   islower
   SUITE
     libc_ctype_unittests
   SRCS
     islower_test.cpp
   DEPENDS
     libc.src.ctype.islower
 )
 
 add_libc_unittest(
+  ispunct
+  SUITE
+    libc_ctype_unittests
+  SRCS
+    ispunct_test.cpp
+  DEPENDS
+    libc.src.ctype.ispunct
+)
+
+add_libc_unittest(
   isupper
   SUITE
     libc_ctype_unittests
   SRCS
     isupper_test.cpp
   DEPENDS
     libc.src.ctype.isupper
 )

libc/test/src/ctype/isblank_test.cpp

+//===-- Unittests for isblank----------------------------------------------===//
+//
+// 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/ctype/isblank.h"
+#include "utils/UnitTest/Test.h"
+
+TEST(IsBlank, DefaultLocale) {
+  // Loops through all characters, verifying that space and horizontal tab
+  // return a non-zero integer and everything else returns zero.
+  for (int ch = 0; ch < 255; ++ch) {
+    if (ch == ' ' || ch == '\t')
+      EXPECT_NE(__llvm_libc::isblank(ch), 0);
+    else
+      EXPECT_EQ(__llvm_libc::isblank(ch), 0);
+  }
+}

libc/test/src/ctype/iscntrl_test.cpp

+//===-- Unittests for iscntrl----------------------------------------------===//
+//
+// 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/ctype/iscntrl.h"
+#include "utils/UnitTest/Test.h"
+
+TEST(IsCntrl, DefaultLocale) {
+  // Loops through all characters, verifying that control characters
+  // return a non-zero integer, all others return zero.
+  for (int ch = 0; ch < 255; ++ch) {
+    if ((0 <= ch && ch <= 0x1f /*US*/) || ch == 0x7f /*DEL*/)
+      EXPECT_NE(__llvm_libc::iscntrl(ch), 0);
+    else
+      EXPECT_EQ(__llvm_libc::iscntrl(ch), 0);
+  }
+}

libc/test/src/ctype/isgraph_test.cpp

+//===-- Unittests for isgraph----------------------------------------------===//
+//
+// 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/ctype/isgraph.h"
+#include "utils/UnitTest/Test.h"
+
+TEST(IsGraph, DefaultLocale) {
+  // Loops through all characters, verifying that graphical characters
+  // return a non-zero integer, everything else returns zero.
+  for (int ch = 0; ch < 255; ++ch) {
+    if ('!' <= ch && ch <= '~') // A-Z, a-z, 0-9, punctuation.
+      EXPECT_NE(__llvm_libc::isgraph(ch), 0);
+    else
+      EXPECT_EQ(__llvm_libc::isgraph(ch), 0);
+  }
+}

libc/test/src/ctype/ispunct_test.cpp

+//===-- Unittests for ispunct----------------------------------------------===//
+//
+// 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/ctype/ispunct.h"
+#include "utils/UnitTest/Test.h"
+
+// Helper function to mark the sections of the ASCII table that are
+// punctuation characters. These are listed below:
+//  Decimal    |         Symbol
+//  -----------------------------------------
+//  33 -  47   |  ! " $ % & ' ( ) * + , - . /
+//  58 -  64   |  : ; < = > ? @
+//  91 -  96   |  [ \ ] ^ _ `
+// 123 - 126   |  { | } ~
+static inline int is_punctuation_character(int c) {
+  return ('!' <= c && c <= '/') || (':' <= c && c <= '@') ||
+         ('[' <= c && c <= '`') || ('{' <= c && c <= '~');
+}
+
+TEST(IsPunct, DefaultLocale) {
+  // Loops through all characters, verifying that punctuation characters
+  // return a non-zero integer, and everything else returns zero.
+  for (int ch = 0; ch < 255; ++ch) {
+    if (is_punctuation_character(ch))
+      EXPECT_NE(__llvm_libc::ispunct(ch), 0);
+    else
+      EXPECT_EQ(__llvm_libc::ispunct(ch), 0);
+  }
+}