使用awk处理带标签文本

 假设有"123abc456def789ghi"这么一个字符串
题中的字符串，要截取长度5，则返回的字符串应该为:123ab,要截取长度8，应返回123abc45。
附：
1 和标记不得计算在长度之内。
2 截取后的字符串，要保留原有标签，不过如果最后有一个标签没有闭合，则去掉其开始标签。

代码：如下

#!/bin/bash
a='123abc456def789ghi'
echo "test text is '$a'"
read -p "please input a number little than `echo $a|awk '{gsub("<em>|</em>","",$0);printf length($0)}'`: " num

echo $a | awk -vnum="$num" -vFS='' '{for(i=1;i<=NF;i++){orisum=sum;sum+=length($i)-4;if(sum<=num){strlist=strlist ~ /./ ? strlist""FS""$i : $i}else{sub("","",$i);strlist=strlist ~ /./ ? strlist""FS""substr($i,1,num-orisum) : substr($i,1,num-orisum);break}}print strlist}'

测试结果：

[root@localhost ~]# bash 1.sh
test text is '123abc456def789ghi'
please input a number little than 18: 2
12
[root@localhost ~]# bash 1.sh
test text is '123abc456def789ghi'
please input a number little than 18: 5
123ab
[root@localhost ~]# bash 1.sh
test text is '123abc456def789ghi'
please input a number little than 18: 7
123<em>abc</em>4
[root@localhost ~]# bash 1.sh
test text is '123abc456def789ghi'
please input a number little than 18: 10
123<em>abc</em>456d
[root@localhost ~]# bash 1.sh
test text is '123abc456def789ghi'
please input a number little than 18: 9
123<em>abc</em>456
[root@localhost ~]# bash 1.sh
test text is '123abc456def789ghi'
please input a number little than 18: 14
123<em>abc</em>456<em>def</em>78
[root@localhost ~]# bash 1.sh
test text is '123abc456def789ghi'
please input a number little than 18: 17
123<em>abc</em>456<em>def</em>789gh
[root@localhost ~]# bash 1.sh
test text is '123abc456def789ghi'
please input a number little than 18: 18
123<em>abc</em>456<em>def</em>789<em>ghi</em>
[root@localhost ~]#