g++中宏NULL究竟是什么?
NULL是个指针,还是个整数?0?或(void*)0?答案是和g++版本有关。g++ 4.6支持C++11,引入了nullptr,也许会发生变化。
可以写段简单代码求证一下:
#include #include
int main() { printf("NULL: %d\n", NULL); printf("sizeof(NULL): %d\n", sizeof(NULL)); printf("typeid(__null).name(): %s\n", typeid(__null).name()); printf("typeid(0).name(): %s\n", typeid(0).name()); return 0; } |
使用g++ 4.1.2(SuSE 10.1)编译,输出结果如下:
NULL: 0 sizeof(NULL): 8 typeid(__null).name(): l typeid(0).name(): i |
从输出结果,可以看到NULL是long类型的整数,定义应当是0L或0LL。
下面再借助gdb,来看一看它的真面目(博文:《GDB高级技巧》有介绍gdb的高级使用)。假设源文件名为x.cpp,使用如下方法编译:
g++ -g3 -gdwarf-2 -o x x.cpp |
注意这里使用了参数“-g3 -gdwarf-2”。其中“-g3”不能是“-g”。
编译成功后,在gdb中跟踪运行程序,在main函数处打断点即可:
adoop@VM-39-166-sles10-64:~/current> gdb ./x GNU gdb 6.6 Copyright (C) 2006 Free Software Foundation, Inc. GDB is free software, covered by the GNU General Public License, and you are welcome to change it and/or distribute copies of it under certain conditions. Type "show copying" to see the conditions. There is absolutely no warranty for GDB. Type "show warranty" for details. This GDB was configured as "x86_64-suse-linux"... Using host libthread_db library "/lib64/libthread_db.so.1". (gdb) b main Breakpoint 1 at 0x4005dc: file x.cpp, line 6. (gdb) r Starting program: /data/hadoop/hadoop-2.4.0/x
Breakpoint 1, main () at x.cpp:6 6 printf("NULL: %d\n", NULL); (gdb) macro expand NULL expands to: __null (gdb) p NULL No symbol "__null" in current context. (gdb) ptype NULL No symbol "__null" in current context. (gdb) p __null No symbol "__null" in current context. (gdb) ptype __null No symbol "__null" in current context. (gdb) |
从gdb的跟踪结果,不难看到NULL的真身是__null,但__null又是什么了?试图从/usr/include中找到答案:
hadoop@VM-39-166-sles10-64:~/current> grep __null /usr/include/* /usr/include/libio.h:# define NULL (__null) hadoop@VM-39-166-sles10-64:~/current> grep __null /usr/include/*/* /usr/include/asm-i386/vm86.h: long __null_ds; /usr/include/asm-i386/vm86.h: long __null_es; /usr/include/asm-i386/vm86.h: long __null_fs; /usr/include/asm-i386/vm86.h: long __null_gs; /usr/include/asm-i386/vm86.h: long __null_ds; /usr/include/asm-i386/vm86.h: long __null_es; |
未能找到满意的答案,那么__null只能是g++内置定义的,所以未出现在任何头文件中,事实证明也如此,在代码中可以直接使用__null(尽管如此,但这个不是个好主意):
#include #include
int main() { printf("NULL: %d\n", NULL); printf("NULL: %d\n", __null); printf("sizeof(NULL): %d\n", sizeof(NULL)); printf("typeid(__null).name(): %s\n", typeid(__null).name()); printf("typeid(0).name(): %s\n", typeid(0).name()); return 0; } |
执行man g++,然后搜索“__null”,找到一个有关的条目:
-Wstrict-null-sentinel (C++ only) Warn also about the use of an uncasted "NULL" as sentinel. When compiling only with GCC this is a valid sentinel, as "NULL" is defined to "__null". Although it is a null pointer constant not a null pointer, it is guaranteed to of the same size as a pointer(保证和指针相同的大小). But this use is not portable across different compilers. |
搜索gcc 4.8.2源码中的FSFChangeLog.11文件:
Wed Aug 7 14:10:07 1996 Jason Merrill * ginclude/stddef.h (NULL): Use __null for G++. |
继续搜索gcc源代码,可以得到更多信息:
VM-39-166-sles10-64:/data/gcc-4.8.2/gcc # grep __null */* c-family/c-common.c: { "__null", RID_NULL, 0 }, c-family/c-common.c: /* Create the built-in __null node. It is important that this is c-family/c-common.c: /* Although __null (in C++) is only an integer we allow it c-family/c-common.h:/* The node for C++ `__null'. */ cp/call.c: /* If __null has been converted to an integer type, we do not cp/NEWS: is now defined as __null, a magic constant of type (void *) cp/parser.c: __null cp/parser.c: /* The `__null' literal. */ |
进一步查看c-family/c-common.c文件:
const struct c_common_resword c_common_reswords[] = { { "__is_union", RID_IS_UNION, D_CXXONLY }, { "__label__", RID_LABEL, 0 }, { "__null", RID_NULL, 0 }, { "__real", RID_REALPART, 0 }, { "__real__", RID_REALPART, 0 }, { "__restrict", RID_RESTRICT, 0 }, |
结构体c_common_resword的定义在c-family/c-common.h中:
/* The node for C++ `__null'. */ #define null_node c_global_trees[CTI_NULL]
/* An entry in the reserved keyword table. */ struct c_common_resword { const char *const word; ENUM_BITFIELD(rid) const rid : 16; const unsigned int disable : 16; };
enum c_tree_index { 。。。。。。 CTI_VOID_ZERO, CTI_NULL, CTI_MAX }; |
查看文件cp/parser.c,parser.c是C/C++语法解析器的实现文件:
static tree cp_parser_primary_expression (cp_parser *parser, bool address_p, bool cast_p, bool template_arg_p, bool decltype_p, cp_id_kind *idk) { 。。。。。。 case CPP_KEYWORD: switch (token->keyword) { /* These two are the boolean literals. */ case RID_TRUE: cp_lexer_consume_token (parser->lexer); return boolean_true_node; case RID_FALSE: cp_lexer_consume_token (parser->lexer); return boolean_false_node;
/* The `__null' literal. */ case RID_NULL: cp_lexer_consume_token (parser->lexer); return null_node;
/* The `nullptr' literal. */ case RID_NULLPTR: cp_lexer_consume_token (parser->lexer); return nullptr_node; |

低调大师中文资讯倾力打造互联网数据资讯、行业资源、电子商务、移动互联网、网络营销平台。
持续更新报道IT业界、互联网、市场资讯、驱动更新,是最及时权威的产业资讯及硬件资讯报道平台。
转载内容版权归作者及来源网站所有,本站原创内容转载请注明来源。
- 上一篇
Hadoop-2.4.0中HDFS文件块大小默认为128M
134217728 / 1024 = 131072 / 1024 = 128
- 下一篇
在Linux上编译Hadoop-2.4.0
Linux上编译Hadoop-2.4.0.pdf 目录 目录 1 1.前言 1 2.安装依赖 1 2.1.安装ProtocolBuffer 2 2.2.安装CMake 2 2.3.安装JDK 2 2.4.安装Maven 3 3.编译Hadoop源代码 3 附1:无联网环境编译 5 附2编译环境 6 附3:版本信息 6 附4:常见错误 6 1)unexpectedendtag: 6 附5:相关文档 7 1.前言 Hadoop-2.4.0的源码目录下有个BUILDING.txt文件,它介绍了如何在Linux和Windows下编译源代码,本文基本是遵照BUILDING.txt指示来操作的,这里再做一下简单的提炼。 第一次编译要求能够访问互联网,Hadoop的编译依赖非常多的东西,一定要保证机器可访问互联网,否则难逐一解决所有的编译问题,但第一次之后的编译则不用再下载了。 2.安装依赖 在编译Hadoop2.4.0源码之前,需要将下列几个依赖的东西安装好: 1)JDK1.6或更新版本(本文使用JDK1.7,请不要安装JDK1.8版本,JDK1.8和Hadoop2.4.0不匹配,编译Hadoo...
相关文章
文章评论
共有0条评论来说两句吧...