libdsk-1.3.3.tar.gz doesn't unpack

[Ivan Shmakov]

>>>>> s dubrovich <s_dub...@yahoo.com> writes:
>>>>> On Tuesday, August 21, 2012 12:57:38 PM UTC-5, Ivan Shmakov wrote:
>>>>> s dubrovich <s_dub...@yahoo.com> writes:
>>>>> On Sunday, August 19, 2012 12:02:38 AM UTC-5, Ivan Shmakov wrote:

[Cross-posting to news:comp.os.linux.misc and
news:comp.infosystems.www.misc, while dropping news:comp.os.cpm,
as this discussion has nothing to do with CP/M.]

[...]

>>>> $ sha1sum -- \
>>>> seasip.info/Unix/LibDsk/libdsk-1.2.1.tar.gz \
>>>> seasip.info/Unix/LibDsk/libdsk-1.3.3.tar.gz

>>>> 8d02a8ed0021a34a13c061858ca026c385c11c5b seasip.info/Unix/LibDsk/libdsk-1.2.1.tar.gz
>>>> 5ec36eb90cc55ba74b68b1529a15c51d60d382fe seasip.info/Unix/LibDsk/libdsk-1.3.3.tar.gz

[...]

>> Did you check the file(1) and sha1sum(1) output against the one I've
>> provided?

> Just did, and they are different..

> $ file -- libdsk-1.2.1.tar.gz libdsk-1.2.1.tar.gz: tar archive
> $ sha1sum -- libdsk-1.2.1.tar.gz
> bff5b7f52b0efde43babc1ad1f9ce4287e6226bd libdsk-1.2.1.tar.gz
> $ file -- libdsk-1.3.3.tar.gz
> libdsk-1.3.3.tar.gz: tar archive
> $ sha1sum -- libdsk-1.3.3.tar.gz
> 8868a4c066ac6d16f442f8595fabbb7b85961339 libdsk-1.3.3.tar.gz

> I'm not sure where to turn to next, so your advice is appreciated.

Fortunately, this has little to do with Debian: the Apache HTTP
server at seasip.info correctly encodes the "type" of the "file"
being served as:

$ wget -O /dev/full -- seasip.info/Unix/LibDsk/libdsk-1.3.3.tar.gz
...
Content-Type: application/x-tar
Content-Encoding: x-gzip
...
$

Then, the HTTP client used decides that the user isn't
interested in the "compressed" version of the file, and unpacks
it by itself. The net result is that the file saved on disk is
no longer .tar.gz, but simply .tar. Indeed, its SHA-1 above
matches the one of the uncompressed one:

$ zcat < seasip.info/Unix/LibDsk/libdsk-1.3.3.tar.gz | sha1sum
8868a4c066ac6d16f442f8595fabbb7b85961339 -
$

The tar(1) command line invocation for the available .tar file
would be:

$ tar -x < FILE

(Instead of $ tar -zx < FILE for a "real" .tar.gz file.)

My personal preference is to use GNU Wget to obtain "files" from
HTTP (HTTPS) servers, but Aria2, Curl and lwp-download(1p) (from
libwww-perl) will also fit.

--
FSF associate member #7257 http://sf-day.org/