Shell Programming and Scripting

Hi All,
I am new to this unix stuff.I just have one doubt:suppose i have two tar files and sometimes it happens that when we just check these files from outside these two tar files look same

Code:

 
"Eg:
ls -lrt
drw-r--r--   1 oasis    logadmin   37067 Apr  3 05:48 file1.tar
drw-r--r--   1 oasis    logadmin   37067 Apr  3 05:48 file2.tar"

but when we check inside that we may find there are some jars which have been changed.Now i want to have one script which will compare these two tars and then provide as output those jars inside these two tars which have been changed.Pardon me for any grammatical mistakes.

Thanks in advance

user tar -vft tarfile to see the content and size of the file

you can use cksum command to see if the checksum of the tar files are same. But for identifying the changed files, you have to extract the tar(tar -xvf) and compare the files.

Hi all,
Thanks for your replies.
As i am a newbie to scripting i am not sure how to proceed.I am looking for a script that differentiates between two tar files without extracting(due to space issues) and output the files that have been changed.Please note that these tar files will contain jar files and inside those there are files.
The structure is :
tarfile :it is an external directory inside that there are jars which are also directories.and inside that there are files.and i have to differentiate between such two tarfiles that wich all files have been changed inside that.it is not always necessary that there is a difference between two tarfiles.It may happen that nothing has changed.
I work in java and javascript and am not so comfortable with unix.So requesting your help.
Thanks in advance.

In general, it is not possible to know which file has changed by looking at the size of the file or the time stamp.

In general, the command

Code:

cmp

will report the byte at which the first difference is found.

Code:

sum

or (better)

Code:

md5sum

are statistically highly likely to work. Well, the chances of md5sum reporting the same hash value for two different files is vanishingly low, but cmp is probably faster.

So you should begin by determining if the files are indeed different

Code:

cmp file1.tar file2.tar
file1.tar file2.tar differ: differ: char 6, line 1

If they are, you should next compare the table of contents:

Code:

tar tf file1.tar >/tmp/file1.toc
tar tf file2.tar >/tmp/file2.toc
cmp /tmp/file1.toc /tmp/file2.toc

If the TOCs differ, then use sdiff

Code:

sdiff /tmp/file1.toc /tmp/file2.toc

If the TOC does not differ, you will need to extract the files from the two .tar files, one by one, and cmp these extracted files.

Code:

[ ! -d /tmp/file1 ] && mkdir /tmp/file1
[ ! -d /tmp/file2 ] && mkdir /tmp/file2
tar tf file1.tar >/tmp/file1/file1.toc
tar tf file2.tar >/tmp/file2/file2.toc
for f in $(cat /tmp/file1/file1.toc)
do
   cd /tmp/file1
   tar xf /path/to/file1.tar $f
   cd /tmp/file2
   tar xf /path/to/file2.tar $f
   Echo comparing  /tmp/file1/$f  /tmp/file2/$f
   cmp /tmp/file1/$f /tmp/file2/$f
done

#d vfg3

@siri,

Just do diff b/w the two tar files.

diff -ibw file1.tar file2.tar
O/P: Binary files file1.tar and file2.tar differ.
But it will not show the difference.

cmp file1.tar file2.tar
The cmp result show differ in byte number, line number