[Bug-tar] tar ignores files called "http:*"

[matthew green]

hi folks.


i've found that when i have a file pathname that looks like a http URL,
GNU tar incorrectly tries to open a http connection instead of reading
the file on my disk.  eg:

   myhost /var/tmp> tar xzpf http://hostname.here/file/path/here.tgz
   tar (child): Cannot connect to http: resolve failed

   gzip: stdin: unexpected end of file
   tar: Child returned status 128
   tar: Error is not recoverable: exiting now
   myhost /var/tmp> ls -l ./http://hostname.here/file/path/here.tgz
   -rw-r--r-- 1 nobody nogroup 2962981999 May 28 13:29 
./http://hostname.here/file/path/here.tgz

it of course works if i use "./http:..." as the pathname.  last i looked,
the opengroup/posix tar does not specify http: as a special path so i find
this behaviour incorrect (and annoying.)

perhaps tar should try to open() the pathname first, before trying to
guess what it really is.  or perhaps, it should try to open() it if the
guess made turns out to be not so good.

thanks,


.mrg.