Wrong filename for unix download

On the download page, the filename of the unix archive is "ideaIU-138.1696.tar.gz". However, when I download the file, I receive a file named "ideaIU-138.1696.tar", missing the "gz" suffix. This could confuse programs which try to open the file, since it is a gzip archive.

Comment actions Permalink

No disrespect, but could that just be an effect of your browser that you are seeing?

Some browsers automatically decompress .gz files, in which case the file you have will now be a larger 623Mb tar archive, rather than the 281Mb downloaded .gz file.

Trying to help

Comment actions Permalink

Hi Stephan,

Actually *.tar.gz is a tar archive compressed with gzip.
Gzip is a single-file compression method, and does not itself provide an archive format that can contain multiple files.
Tar provides the archive in this case, and gzip compressed the single archive file.

So as David suggests, likely your browser (or a helper app) decompressed the downloaded gzip file, but did not bother
to continue to unpack the tar archive.

Comment actions Permalink

Thank you both for the explanation, but I already knew the difference between tar and tar.gz. ;)

It seemed strange to me beacuse the downloaded file was saved by my browser (Chrome) as ".tar", but "file" proved that it is still still a compressed ".tar.gz".

However, I ran some tests and found that both wget and Firefox chose the correct filename. It seems that Chrome is the only client which parses the content type (which is returned incorreclty as "application/x-tar" by the download page).

So, could you please update the content type to "application/x-gzip"?

Comment actions Permalink

I'm not an expert, but I think this may be your problem.


Incorrect naming of .tar.gz downloaded files

Because I see this:

curl http://download-cf.jetbrains.com/idea/ideaIU-138.1696.2.tar.gz?_ga=1.64599149.335720998.1367407123 -v >/dev/null

* Hostname was NOT found in DNS cache

*   Trying

  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current

                                 Dload  Upload   Total   Spent    Left  Speed

  0     0    0     0    0     0      0      0 --:--:-- --:--:-- --:--:--     0* Connected to download-cf.jetbrains.com ( port 80 (#0)

> GET /idea/ideaIU-138.1696.2.tar.gz?_ga=1.64599149.335720998.1367407123 HTTP/1.1

> User-Agent: curl/7.37.0

> Host: download-cf.jetbrains.com

> Accept: */*


< HTTP/1.1 200 OK

< Content-Type: application/x-tar

< Content-Length: 281334521

< Connection: keep-alive

< Date: Wed, 20 Aug 2014 15:21:41 GMT

< x-amz-meta-s3cmd-attrs: uid:572/gname:cds/uname:cds/gid:574/mode:33188/mtime:1408485897/atime:1408485891/md5:f8b190e6a6fcbaf39431fa82d244ef4e/ctime:1408485897

< Content-Encoding: gzip

< Last-Modified: Tue, 19 Aug 2014 22:23:58 GMT

< ETag: "fe739bfaf9c7bed03f20e051b5a05ef2-18"

< Accept-Ranges: bytes

* Server AmazonS3 is not blacklisted

< Server: AmazonS3

< Age: 64184

< X-Cache: Hit from cloudfront

< Via: 1.1 2297ae65d06c070d63c49ed7259f5fbb.cloudfront.net (CloudFront)

< X-Amz-Cf-Id: rIcQTFPJnRQTTxuv77lMrgot0KZ-z_LsfiByZ9MHlKt6bn9HH-Npzg==


{ [data not shown]

Again, I'm not an expert in how to run content delivery for best overall effect.

Comment actions Permalink

Thanks for the test, you highlighted the error perfectly. The downloaded file is not a tar-Archive which is encoded with gzip. For comparison, try a random download from SourceForge, for example vim-latex. Choose a mirror and you will see that the content-type is set correctly to "application/x-gzip".


Please sign in to leave a comment.