一. Linux xxd -i功能
Linux系统xxd命令使用二进制或十六进制格式显示文件内容。若未指定outfile参数,则将结果显示在终端屏幕上;否则输出到outfile中。详细的用法可参考linux命令xxd。
本文主要关注xxd命令-i选项。使用该选项可输出以inputfile为名的C语言数组定义。例如,执行echo 12345 > test和xxd -i test命令后,输出为:
unsigned char test[] = {0x31, 0x32, 0x33, 0x34, 0x35, 0x0a};unsigned int test_len = 6;
可见,数组名即输入文件名(若有后缀名则点号替换为下划线)。注意,0x0a表示换行符LF,即'/n'。
二. xxd -i常见用途
当设备没有文件系统或不支持动态内存管理时,有时会将二进制文件(如引导程序和固件)内容存储在C代码静态数组内。此时,借助xxd命令就可自动生成版本数组。举例如下:
1) 使用Linux命令xdd将二进制文件VdslBooter.bin转换为16进制文件DslBooter.txt:
xxd -i < VdslBooter.bin > DslBooter.txt
其中,'-i'选项表示输出为C包含文件的风格(数组方式)。重定向符号'<'将VdslBooter.bin文件内容重定向到标准输入,该处理可剔除数组声明和长度变量定义,使输出仅包含16进制数值。
2) 在C代码源文件内定义相应的静态数组:
static const uint8 bootImageArray[] = {#include " ../../DslBooter.txt"};TargetImage bootImage = {(uint8 *) bootImageArray,sizeof(bootImageArray) / sizeof(bootImageArray[0])};
编译源码时,DslBooter.txt文件的内容会自动展开到上述数组内。通过巧用#include预处理指令,可免去手工拷贝数组内容的麻烦。
三. 类xxd -i功能的Python实现
本节将使用Python2.7语言实现类似xxd -i的功能。
因为作者处于学习阶段,代码中存在许多写法不同但功能相同或相近的地方,旨在提供不同的语法参考,敬请谅解。
首先,请看一段短小却完整的程序(保存为xddi.py):
#!/usr/bin/python#coding=utf-8#判断是否C语言关键字CKeywords = ("auto", "break", "case", "char", "const", "continue", "default","do","double","else","enum","extern","float","for","goto","if","int","long","register","return","short","signed","static","sizeof","struct","switch","typedef","union","unsigned","void","volatile","while", "_Bool") #_Bool为C99新关键字def IsCKeywords(name):for x in CKeywords:if cmp(x, name) == 0:return Truereturn Falseif __name__ == '__main__':print IsCKeywords('const')#Xxdi()
这段代码判断给定的字符串是否为C语言关键字。在Windows系统cmd命令提示符下输入E:/PyTest>python xxdi.py,执行结果为True。
接下来的代码片段将省略头部的脚本和编码声明,以及尾部的'main'段。
生成C数组前,应确保数组名合法。C语言标识符只能由字母、数字和下划线组成,且不能以数字开头。此外,关键字不能用作标识符。所有,需要对非法字符做处理,其规则参见代码注释:
import redef GenerateCArrayName(inFile):#字母数字下划线以外的字符均转为下划线#'int $=5;'的定义在Gcc 4.1.2可编译通过,但此处仍视为非法标识符inFile = re.sub('[^0-9a-zA-Z/_]', '_', inFile) #'_'改为''可剔除非法字符#数字开头加双下划线if inFile[0].isdigit() == True:inFile = '__' + inFile#若输入文件名为C语言关键字,则将其大写并加下划线后缀作为数组名#不能仅仅大写或加下划线前,否则易于用户自定义名冲突if IsCKeywords(inFile) is True:inFile = '%s_' %inFile.upper()return inFile
以print GenerateCArrayName('1a$if1#1_4.txt')执行时,入参字符串将被转换为__1a_if1_1_4_txt。类似地,_Bool被转换为_BOOL_。
为了尽可能模拟Linux命令风格,还需提供命令行选项和参数。解析模块选用optionparser,其用法详见python命令行解析。类xxd -i功能的命令行实现如下:
#def ParseOption(base, cols, strip, inFile, outFile):def ParseOption(base = 16, cols = 12, strip = False, inFile = '', outFile = None):from optparse import OptionParsercustUsage = '/n xxdi(.py) [options] inFile [outFile]'parser = OptionParser(usage=custUsage)parser.add_option('-b', '--base', dest='base',help='represent values according to BASE(default:16)')parser.add_option('-c', '--column', dest='col',help='COL octets per line(default:12)')parser.add_option('-s', '--strip', action='store_true', dest='strip',help='only output C array elements')(options, args) = parser.parse_args()if options.base is not None:base = int(options.base)if options.col is not None:cols = int(options.col)if options.strip is not None:strip = Trueif len(args) == 0:print 'No argument, at least one(inFile)!/nUsage:%s' %custUsageif len(args) >= 1:inFile = args[0]if len(args) >= 2:outFile = args[1]return ([base, cols, strip], [inFile, outFile])
被注释掉的def ParseOption(...)原本是以下面的方式调用:
base = 16; cols = 12; strip = False; inFile = ''; outFile = ''([base, cols, strip], [inFile, outFile]) = ParseOption(base,cols, strip, inFile, outFile)
其意图是同时修改base、cols、strip等参数值。但这种写法非常别扭,改用缺省参数的函数定义方式,调用时只需要写ParseOption()即可。若读者知道更好的写法,望不吝赐教。
以-h选项调出命令提示,可见非常接近Linux风格:
E:/PyTest>python xxdi.py -hUsage:xxdi(.py) [options] inFile [outFile]Options:-h, --help show this help message and exit-b BASE, --base=BASE represent values according to BASE(default:16)-c COL, --column=COL COL octets per line(default:12)-s, --strip only output C array elements
基于上述练习,接着完成本文的重头戏:
def Xxdi():#解析命令行选项及参数([base, cols, strip], [inFile, outFile]) = ParseOption()import osif os.path.isfile(inFile) is False:print ''''%s' is not a file!''' %inFilereturnwith open(inFile, 'rb') as file: #必须以'b'模式访问二进制文件#file = open(inFile, 'rb') #Python2.5以下版本不支持with...as语法#if True:#不用for line in file或readline(s),以免遇'0x0a'换行content = file.read()#将文件内容"打散"为字节数组if base is 16: #Hexadecimalcontent = map(lambda x: hex(ord(x)), content)elif base is 10: #Decimalcontent = map(lambda x: str(ord(x)), content)elif base is 8: #Octalcontent = map(lambda x: oct(ord(x)), content)else:print '[%s]: Invalid base or radix for C language!' %basereturn#构造数组定义头及长度变量cArrayName = GenerateCArrayName(inFile)if strip is False:cArrayHeader = 'unsigned char %s[] = {' %cArrayNameelse:cArrayHeader = ''cArrayTailer = '};/nunsigned int %s_len = %d;' %(cArrayName, len(content))if strip is True: cArrayTailer = ''#print会在每行输出后自动换行if outFile is None:print cArrayHeaderfor i in range(0, len(content), cols):line = ', '.join(content[i:i+cols])print ' ' + line + ','print cArrayTailerreturnwith open(outFile, 'w') as file:#file = open(outFile, 'w') #Python2.5以下版本不支持with...as语法#if True:file.write(cArrayHeader + '/n')for i in range(0, len(content), cols):line = reduce(lambda x,y: ', '.join([x,y]), content[i:i+cols])file.write(' %s,/n' %line)file.flush()file.write(cArrayTailer)
Python2.5以下版本不支持with...as语法,而作者调试所用的Linux系统仅装有Python2.4.3。因此,要在Linux系统中运行xddi.py,只能写为file = open(...。但这需要处理文件的关闭和异常,详见理解Python中的with…as…语法。注意,Python2.5中使用with...as语法时需要声明from __future__ import with_statement。
可通过platform.python_version()获取Python版本号。例如:
import platform#判断Python是否为major.minor及以上版本def IsForwardPyVersion(major, minor):#python_version()返回'major.minor.patchlevel',如'2.7.11'ver = platform.python_version().split('.')if int(ver[0]) >= major and int(ver[1]) >= minor:return Truereturn False
经过Windows和Linux系统双重检验后,Xddi()工作基本符合预期。以123456789ABCDEF.txt文件(内容为'123456789ABCDEF')为例,测试结果如下:
E:/PyTest>python xxdi.py -c 5 -b 2 -s 123456789ABCDEF.txt[2]: Invalid base or radix for C language!E:/Pytest>python xxdi.py -c 5 -b 10 -s 123456789ABCDEF.txt49, 50, 51, 52, 53,54, 55, 56, 57, 65,66, 67, 68, 69, 70,E:/PyTest>python xxdi.py -c 5 -b 10 123456789ABCDEF.txtunsigned char __123456789ABCDEF_txt[] = {49, 50, 51, 52, 53,54, 55, 56, 57, 65,66, 67, 68, 69, 70,};unsigned int __123456789ABCDEF_txt_len = 15;E:/PyTest>python xxdi.py -c 5 -b 8 123456789ABCDEF.txtunsigned char __123456789ABCDEF_txt[] = {061, 062, 063, 064, 065,066, 067, 070, 071, 0101,0102, 0103, 0104, 0105, 0106,};unsigned int __123456789ABCDEF_txt_len = 15;E:/PyTest>python xxdi.py 123456789ABCDEF.txtunsigned char __123456789ABCDEF_txt[] = {0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, 0x38, 0x39, 0x41, 0x42, 0x43,0x44, 0x45, 0x46,};unsigned int __123456789ABCDEF_txt_len = 15;
再以稍大的二级制文件为例,执行 python xxdi.py VdslBooter.bin booter.c后,booter.c文件内容如下(截取首尾):
unsigned char VdslBooter_bin[] = {0xff, 0x31, 0x0, 0xb, 0xff, 0x3, 0x1f, 0x5a, 0x0, 0x0, 0x0, 0x0,//... ... ... ...0x0, 0x0, 0x0, 0x0, 0xff, 0xff, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0,0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0,};unsigned int VdslBooter_bin_len = 53588;
综上可见,作者实现的xxdi模块与Linux xxd -i功能非常接近,且各有优劣。xxdi优点在于对数组名合法性校验更充分(关键字检查),数组内容表现形式更丰富(8进制和10进制);缺点在于不支持重定向,且数值宽度不固定(如0xb和0xff)。当然,这些缺点并不难消除。例如,用'0x%02x'%val代替hex(val)即可控制输出位宽。只是,再加完善难免提高代码复杂度,也许会事倍功半。
以上所述是小编给大家介绍的Python实现Linux命令xxd -i功能,希望对大家以上帮助!