On Mon, 1 May 2000, Matthew Hawkins wrote: > Well the odd funky filesystem is ext2, the odd funky kernel is 2.2.15pre17 and > 2.3.99pre* (* currently == 6). I'm not aware of anything in these kernels > which would cause such problems.
Hmmmmmmmmmmmmmmmm The stat information is fine. The atime shows the last access and the ctime shows the last write to the file. The mtime should *exactly* corrispond to the file on the remote site. However, it appears to be the time at the start of the download! AFAIK their are only two ways that can happen.. The first is if the server does not send a Last-Modified header (or one could not be found) or their is some really big cock up in glibc's timegm function (unlikely!) There is also a slim chance that there is something funny with the gzip process. That is easy to check, can you stat the Release files and compare timestamps with the original? Can you use strace to capture a session? Use something like strace -o /tmp/apt -ff -f -s 1024 apt-get update And send me the trace for the http method doing downloading of the files that fail. > Note how the Change and Modify dates differ. I don't think this is normal. It is normal if the program uses mtime to transfer the file timestamp. Jason
