Problem with bz2 file solved (was Re: Python 2.0b1 is released!)

Piet van Oostrum piet at cs.uu.nl
Tue Sep 12 15:49:57 EDT 2000


>>>>> Piet van Oostrum <piet at cs.uu.nl> (PvO) writes:

>>>>> "Fred L. Drake, Jr." <fdrake at beopen.com> (FLD) writes:
FLD> Piet van Oostrum wrote:
>>> bunzip: Data integrity error when decompressing.
>>> Input file = pdf-a4-2.0b1.tar.bz2, output file = pdf-a4-2.0b1.tar

FLD> Tim Peters writes:
>>> Try again?  I just downloaded it and Cygwin's bunzip2 had no problem.  No
>>> problem un-tarring or viewing in Acrobat either.

FLD> "bunzip" may not be able to handle it; "bunzip2" is a different
FLD> program using a (slightly?) different algorithm.

PvO> I forgot to tell it's on Windows 95. I tried both the cygwin version and
PvO> the (much smaller) non-cygwin version from the official bzip2 home page
PvO> (http://sources.redhat.com/bzip2) and they give the same error. Must be a
PvO> problem on Windows then.

I finally found the problem. After Guido's advice to try another download
(which, by the way, I had already done before), I downloaded it again but
this time with Internet Explorer rather than Netscape. And now the file was
correct. It appears that netscape translates LF to CR,LF on Windows. This
is caused by the fact that the file is served with MIME type text/plain,
which of course is incorrect. Internet Explorer is know to ignore the MIME
types and only look at the file extension, which generally is a bad idea
but saves the file in this case.
I always thought that Netscape would save files downloaded (with
SHIFT/click or RIGHT MOUSE/Save link as...) without translation.
Unfortunately I don't know a way to specify that a file should not be
mangled. 
-- 
Piet van Oostrum <piet at cs.uu.nl>
URL: http://www.cs.uu.nl/~piet [PGP]
Private email: P.van.Oostrum at hccnet.nl



More information about the Python-list mailing list