更新ivy资源库的自动化脚本

为产品库添加一个新jar包需要下面几个步骤：

1. 根据jar包名字，从maven库里得到依赖定义字符串；

2. 将该字符串加上配置说明和下载脚本（参见笔记“依赖管理常用vim脚本”）；

3. 将已下载条目备份文件（installed.txt）另存为old.bpk；

4. 将build.xml中的下载条目插入到installed.txt中（按字母顺序排列）；

5. 用新的下载脚本替换build.xml中原有的下载脚本；

6. 运行ant命令将jar包下载到产品库中；

7. 在ivy.xml中加入依赖条目，解析之；

以下是上面1~5步的自动化脚本。

#-------------------------------------------------------------------------------

# Purpose: download jars to public repo

# Author: Li Chao

# Created: 28-06-2011

#

# Demonstration:

# install Apache Log4j 1.2

# 1. from mvnrepository.com we got the dep string:

#

# 2. modify dep_str in this file with above dep string:

# 3. run this file, copy modified dep string:

#

# 4. open a shell and run ant cmd to download jars of log4j

# 5. paste modified dep string to ivy.xml and resolve it

#-------------------------------------------------------------------------------

#!/usr/bin/env python

import os, re

dep_str = ''

cmd_pattern = '\(\{from.resolver\}" to="\\)\{to.resolver\}"/>'

script = 'build.xml'

def conv_str():

"""

covert dep_str to custom_ivy_dep string and ivy_install string;

"""

custom_ivy_dep = dep_str.replace('/>',' conf="compile->master"/>')

print 'Copy the following String:\n%s' %custom_ivy_dep

keys = re.findall(r'"\S+"', dep_str)

if len(keys)!=3:

raise Exception('Format wrong of input string!')

global ivy_install

ivy_install = '<ivy:install organisation=%s module=%s revision=%s '\

  'from="${from.resolver}" to="${to.resolver}"/>' %tuple(keys)

def do_bpk():

"""

create backup file(filename: installed.txt/old.bpk):

1 extract old install string from build.xml;

2 delete old.bpk if it exists;

3 rename installed.txt to old.bpk;

4 read old.bpk file, insert old install string and create installed.txt file;

"""

installed = 'installed.txt'

old_bpk ='old.bpk'

xml_file = open(script)

global xml_str

xml_str = xml_file.read()

xml_file.close()

old_install = re.findall(cmd_pattern, xml_str, re.M)[0]

print 'old install string is:\n%s' %old_install

if os.path.exists(old_bpk): os.remove(old_bpk)

if os.path.exists(installed):

os.rename(installed, old_bpk)

else:

raise Exception('%s does not exist!'%installed)

installed_file = open(old_bpk)

installed_items = installed_file.read()

installed_file.close()

all_items = installed_items.split('\n')

all_items.append(old_install)

all_items.sort()

bpk_file = open(installed,'w')

bpk_file.write( '\n'.join(all_items) )

bpk_file.close()

def build_script():

"""

modify install script(build.xml): replace old install string with new one

"""

xml_file = open(script,'w')

new_xml = re.sub(cmd_pattern, ivy_install, xml_str)

xml_file.write(new_xml)

xml_file.close()

def main():

conv_str()

do_bpk()

build_script()

if name == 'main':

main()

代码中的技术点包括：

1. 格式化字符串的生成方法：%后面必须接一个tuple；

2. 如何使用正则表达式进行文本的搜索（findall）和替换(sub)；

3. 字符串的join方法进行list的组合，也可以用reduce方法： total_str = reduce(lambda x,y:x+y, str_list)

但用join方法更简洁。