BASH版本的base64

damcool

AWK版本的base64編解碼腳本有很多限制，特別是在busybox下，作為新年禮物寫了一些BASH版本的。如果busybox有hexdump的話。基本上沒什么問題了。速度還可以吧。
更新說明
1、去除xargs對(duì)腳本兼容性影響
2、加入對(duì)管道輸入的支持

1. #!/bin/bash
   
2. 
   
3. ################################################################################
   
4. # Function Name:   HELP_USAGE                                                   
   
5. # Description:     Function to display the usage of the script                  
   
6. # Parameters:      None                                                         
   
7. # Return:          Help messages                                                
   
8. # Called By:       Script Main Loop->Script Parameters' Handler                 
   
9. # History:         2013-Dec-31 Initial Edition                       RobinHoo  
   
10. ################################################################################
    
11. 
    
12. function help_usage(){
    
13. cat <<EOF
    
14. BASE64 ENCODE & DECODE BASH SCRIPT
    
15. Usage: $PROGNAME [OPTION]... [FILE]
    
16.   -d, --decode Decode the base64 encoded file
    
17.   -h, --help   Show current help message of the script usages
    
18.    
    
19. 
    
20. Please Report Script Bugs to $AUTHOR_MAIL
    
21. EOF
    
22. exit 1
    
23. }
    
24. 
    
25. function BASE64ENCODE()
    
26. {
    
27.         hexdump -ve '1/1 "%d "' < "$FNAME"|awk -v B64="$BASE64CODE" '{for(i=1;i<=NF;i+=3){t=0;t=$i*256*256+$(i+1)*256+$(i+2);for(j=3;j>=0;j--){c=(i+2-j<=NF)?substr(B64,int(t/2^(6*j))+1,1):"=";printf("%c",c);t%=2^(6*j);if (++k==76) printf"\n";k%=76}}}'
    
28. }
    
29. 
    
30. function BASE64DECODE()
    
31. {
    
32.         local buff=""
    
33.         for buff in $(cat < "$FNAME"|awk -v B64="$BASE64CODE" '{while(length()){split(substr($0,1,4),a,"");$0=substr($0,5);t=0;for(i=3;i>=0;i--) t=t+2^(6*i)*((a[4-i]=="=")?0:index(B64,a[4-i])-1);for (i=2;i>=0;i--) if (a[4-i]!="="){printf("\\x%x",t/2^(8*i));t=t%2^(8*i)}}printf"\n"}');do printf "$buff"; done
    
34. }
    
35. BASE_DIR=$(cd "$(dirname "$0")" && pwd)
    
36. PROGNAME=$(basename "$0")
    
37. AUTHOR_MAIL="robin.hoo@hotmail.com"
    
38. BASE64CODE="ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
    
39. DECODE=0;
    
40. HELP=0;
    
41. FNAME="";
    
42. while [ $# -gt 0 ]
    
43. do
    
44.     case "$1" in
    
45.     (-d)        DECODE=1;;
    
46.     (-h)        HELP=1;;
    
47.     (--decode)  DECODE=1;;
    
48.     (--help)    HELP=1;;
    
49.     (-*)        echo "$PROGNAME: error - unrecognized option or parameter $1" 1>&2; HELP=1;break;;
    
50.     (*)         [ "$FNAME" != "" ] && echo "$PROGNAME: error - more than one file name " 1>&2 && HELP=1 && break || FNAME="$1";;
    
51.     esac
    
52.     shift
    
53. done
    
54. [ $# -gt 1 ] && HELP=1
    
55. [ "$FNAME" == "" ] && FNAME="/dev/stdin"
    
56. [ $HELP -eq 1 ] && help_usage
    
57. 
    
58. [ $DECODE -eq 1 ] && BASE64DECODE || BASE64ENCODE
    

復(fù)制代碼

bikkuri

膜拜了！之前我用awk寫的base64編碼和解碼程序確實(shí)效率很低，在CPU 200MHz的Busybox上編碼或者解碼一個(gè)大約4000字節(jié)的二進(jìn)制文件需要大約一分鐘時(shí)間。
說實(shí)話，對(duì)您的腳本中的awk的部分我看得還不是很懂，需要時(shí)間消化。
對(duì)您的腳本我有三個(gè)問題：
1、貌似function help_usage(){ 這樣定義函數(shù)的方式是awk，在shell下應(yīng)該去掉function；
2、編碼的部分我在busybox下測試了沒有問題，但是解碼的部分在busybox下測試報(bào)錯(cuò)而且解碼文件比原文件還要大：

1. root@unknown:/tmp/test# ./base64.sh -d a.64 > b.gz
   
2. xargs: illegal option -- i
   
3. root@unknown:/tmp/test# ls -l
   
4. -rw-r--r--    1 root     root            36 Jan  1 02:01 a.64
   
5. -rw-r--r--    1 root     root            26 Jan  1 02:01 a.gz
   
6. -rw-r--r--    1 root     root            48 Jan  1 02:31 b.gz
   
7. -rwxr-xr-x    1 root     root          2183 Jan  1 02:29 base64.sh
   
8. root@unknown:/tmp/test# more b.gz
   
9. SDRzSUFPZDJ3MUlDQS9OSXpjbko1d0lBRmpXV01RWUFBQUE9root@unknown:/tmp/test#
   

復(fù)制代碼

3、解碼時(shí)在busybox下awk無法輸出NULL字符，不知道您是怎么解決的？

damcool

1.Please forget the AWK solution at all
2.Check out the xargs usage in busybox & adjust the decode line accordingly.

Happy New Year !

bikkuri

I found the xargs information in busybox as below:

1. xargs
   
2. xargs [OPTIONS] [PROG [ARGS]]
   
3. Run PROG on every item given by standard input
   
4. Options:
   
5. -p Ask user whether to run each command
   
6. -r Do not run command if input is empty
   
7. -0 Input is separated by NUL characters
   
8. -t Print the command on stderr before execution
   
9. -e[STR] STR stops input processing
   
10. -n N Pass no more than N args to PROG
    
11. -s N Pass command line of no more than N bytes
    
12. -x Exit if size is exceeded

復(fù)制代碼

which option should I use to replace the -i option?

damcool 發(fā)表于 2014-01-01 10:22
1.Please forget the AWK solution at all
2.Check out the xargs usage in busybox & adjust the decode  ...

damcool

Replace line 32

1. cat "$FNAME"|awk -v B64="$BASE64CODE" '{while(length()){split(substr($0,1,4),a,"");$0=substr($0,5);t=0;for(i=3;i>=0;i--) t=t+2^(6*i)*((a[4-i]=="=")?0:index(B64,a[4-i])-1);for (i=2;i>=0;i--) if (a[4-i]!="="){printf("\\\\\\x%X",t/2^(8*i));t=t%2^(8*i)}}printf"\n"}'|xargs -i printf {}

復(fù)制代碼

with below line

1. for raw_data in $(cat "$FNAME"|awk -v B64="$BASE64CODE" '{while(length()){split(substr($0,1,4),a,"");$0=substr($0,5);t=0;for(i=3;i>=0;i--) t=t+2^(6*i)*((a[4-i]=="=")?0:index(B64,a[4-i])-1);for (i=2;i>=0;i--) if (a[4-i]!="="){printf("\\\\\\x%X",t/2^(8*i));t=t%2^(8*i)}}printf"\n"}'); do printf "$raw_data";done

復(fù)制代碼

bikkuri

After I replaced the line with the new one.
The decode output is double as it should be.

1. root@unknown:/tmp/test# echo "Hello" > a
   
2. root@unknown:/tmp/test# gzip a
   
3. root@unknown:/tmp/test# ./base64_encode_decode.sh a.gz > a.64
   
4. root@unknown:/tmp/test# ./base64_encode_decode.sh -d a.64 > b.gz
   
5. root@unknown:/tmp/test# ls -l
   
6. -rw-r--r--    1 root     root            36 Jan  1 11:12 a.64
   
7. -rw-r--r--    1 root     root            26 Jan  1 11:12 a.gz
   
8. -rw-r--r--    1 root     root            52 Jan  1 11:12 b.gz
   
9. -rwxr-xr-x    1 root     root          2197 Jan  1 11:10 base64_encode_decode.sh
   
10. root@unknown:/tmp/test#

復(fù)制代碼

damcool 發(fā)表于 2014-01-01 11:00
Replace line 32with below line

damcool

That's what I could do now, if you can't manage to make it right I can't neither.

bikkuri

I tested the script on Linux also.
With the xargs line, the decode output is correct, but with the replaced line, the decode output is incorrect.
So the problem is with the replaced line...

damcool 發(fā)表于 2014-01-01 11:32
That's what I could do now, if you can't manage to make it right I can't neither.

damcool

嗨~！現(xiàn)在的年輕人一點(diǎn)點(diǎn)代碼調(diào)試能力都沒有！剛剛找到電腦，看了一下原來的代碼只要去掉xargs后面的-i和{}就可以正常運(yùn)行了。

1. cat "$FNAME"|awk -v B64="$BASE64CODE" '{while(length()){split(substr($0,1,4),a,"");$0=substr($0,5);t=0;for(i=3;i>=0;i--) t=t+2^(6*i)*((a[4-i]=="=")?0:index(B64,a[4-i])-1);for (i=2;i>=0;i--) if (a[4-i]!="="){printf("\\\\\\x%X",t/2^(8*i));t=t%2^(8*i)}}printf"\n"}'|xargs printf

復(fù)制代碼

或者

1. for raw_data in $(cat "$FNAME"|awk -v B64="$BASE64CODE" '{while(length()){split(substr($0,1,4),a,"");$0=substr($0,5);t=0;for(i=3;i>=0;i--) t=t+2^(6*i)*((a[4-i]=="=")?0:index(B64,a[4-i])-1);for (i=2;i>=0;i--) if (a[4-i]!="="){printf("\\\\x%X",t/2^(8*i));t=t%2^(8*i)}}printf"\n"}'); do printf "$raw_data";done

復(fù)制代碼

bikkuri

剛才測試了一下，確實(shí)可以正常輸出了！
謝謝您的指點(diǎn)！