16 Dec
2011
16 Dec
'11
12:58 p.m.
On 12/16/2011 09:07 AM, Liam Proven wrote:
accustomed to?
If that's why, be aware that gzip's user interface is
patterned after a tool that far predates pkzip/pkarc/arc/etc, and that
behavior has been the accepted one that has been considered logical and
good since the 1970s.
With this I agree 100%...
It unpacks to the name(s) of the file(s) [and folders]
in the archive
and leaves the original archive the fsck alone.
In my early Linux days, I was /very/ annoyed several times when Gunzip
cheerfully *removed the archive I told it to unpack*. This is
*absolutely* unacceptable to my mind.
Why? Because it differs from another tool that you may have been
Also be advised that you didn't tell it to unpack an archive (something
gzip
doesn't do), you told it to decompress a compressed file. Tar (say)
doesn't remove a tar file after extracting its contents, because...well,
cheerfully removing the archive you told it to unpack would be absolutely
unacceptable.
Decompressing a compressed file is a whole different animal; you can get
things
right back where they were (if you need to) with a single command.
The only possible difference is the original user's compression setting,
and I respectfully submit that the loss of that information isn't
frequently a big deal.
File compression is something that you "do" to a single file. Should
"undoing" it result in two copies of the file, one with it "done" and
the
other with it "undone"? THAT seems ridiculous to me.
To each their own. Decompression if a file is, to me, a process of reading
the compressed file& wiring an uncompressed one. That the UNIX tradition
appears to be "oh yeah, do that& feel free to tamper with the source as
well" does not wash.
I am no respecter or admirer of tradition. Tradition is often wrong. I am of course prepared to be educated& told why
this is wrong, but I will
only be persuaded if the reasons are really remarkably good ives.
...however, having first executed "compress" pushing thirty years
ago, and working exclusively in the UNIX world (except for some VMS)
since not too long after that, this is the first and only time I've ever
heard it suggested that this behavior is somehow a bad idea.
I'm not saying you're wrong, just saying you're in a vanishingly
small minority in that opinion. Not saying that *means* anything, just
pointing it out.
-Dave
--
Dave McGuire
New Kensington, PA