Shell Programming and Scripting

Hi Experts,

I have a situation where I need to extract the code from a code base using curl which is returning it as zipped file.

1. I want the contents of this tar to be untared in the current directory i.e. the dir from where I am running this curl and tar command.

Below is the command I used -

Code:

$ curl -LsS "http://xyz.com/nexus/service/local/" | tar xz

But this is creating a directory under which has all the files untarred. I am going to put this command in a script, and in the current directory there will be other dirs so I wont know by script which dir to look inside for the code. So I want the code to be extracted to the current dir.

2. I have tried -C option of tar but looks like the tar is created using absolute path and its not untaring in current dir.

3. I tried couple of options using wget but its downloading it as .aar file and tar az errors out

Code:

$ wget --quiet "http://xyz.com/nexus/service/local/" | tar xz
gzip: stdin: unexpected end of file
tar: Child returned status 1
tar: Error exit delayed from previous errors

Please help on how to extract the file in the current directory either using curl/wget and which options of tar

Is there possibly a problem that tar is expecting names of files to archive, but is instead getting the output from curl (or wget)?

Usually, for tar like this, need f - for stdin as tar file and ( cd ... ; tar ) to lay it down in the right subtree:

Code:

curl-gets-remote-tgz | ( cd head_dir ; tar xzf - )

Thanks both for the response but the below command is still extracting the files into a new directory

curl-gets-remote-tgz | ( cd head_dir ; tar xzf - )
Is it because of how its been tared by the repository? Guys please help, I am stuck on this.

Otherwise is there a way to find the tar file returned by the curl command, for ex say i run wget or curl command it returns a tar file, how do i know the name of the file?
On the same lines when the tar is creating a new dir and extracting the files, is there a way to know the name of the directory the tar command created to extract the files.If i get to know the name of the directory then I can pull the files from it to the current directory.

What is "curl-gets-remote-tgz" command? Is that a standard UNIX command? I don't find it on a search.

Do wget or curl return a tar file? I'm not aware that they do.

Hanson, sorry there is no such command, it meant the curl command with the url...

Yes looks like curl command is returning the tar file.

Any help please.

If what you're trying to do is strip a leading directory component from every path in a gzip'd tar archive, the following untested pipeline should do it:

Code:

curl ... | gzip -dc | pax -r -s '#^[^/]*/##'

Needless to say, this can cause collisions if there is more than one archive member at the top level.

If this is not what you're trying to accomplish, then you should properly document the situation with a sample listing from the archive as it exists, the exact command used to unpack it, a listing of the resulting filesystem hierarchy, and a simulated listing of the desired filesystem hierarchy.

Regards,
Alister

---------- Post updated at 12:06 PM ---------- Previous update was at 12:02 PM ----------

Create an empty temp directory to work with. Anything that's put there is what you're looking for. You don't need to know its name; just use sh pattern matching wildcards, e.g. for f in *; do ... done.

Regards,
Alister