On Tue, 16 Nov 2004, Matthew Dillon wrote:
>
> :Okay, I downloaded the same cups sourcefile three times from different
> :mirrors and got a perfect result every time.
> :
> :I set SACK=1 again and downloaded a corrupt file from the same ftp server
> :which gave me a perfect file just minutes earlier.
> :
> :So far: five out of six files corrupt with SACK=1 and zero of four with
> :SACK=0.
> :
> :After years of reading scientific journals I can say without any hesitation
> :that SACK is responsible for the file corruption (p=0.0001).
> Could you post the commands you used to download the files (the exact
> same ones you used in your tests). This way Jeff and I won't have to
> go searching around for download locations to test with.
fetch http://ftp.easysw.com/pub/cups/1.1.22/cups-1.1.22-source.tar.bz2
or fetch ftp://ftp2.easysw.com/pub/cups/1.1.22/cups-1.1.22-source.tar.bz2
or fetch ftp://ftp3.easysw.com/pub/cups/1.1.22/cups-1.1.22-source.tar.bz2
See /usr/ports/print/cups/Makefile.common. My first problem was with
ftp.funet.fi, so the problem seems easy to reproduce.
I sent an email to the webmaster for easysw.com to ask him about the
SACK/linux question -- I'll let you know if he answers.