To decide whether this is rather a server or client side problem, the same test done on a current Hardy client (the server is Jaunty in both cases):
#> ls -la
total 12
drwxr-xr-x 2 stefan stefan 4096 2009-06-19 10:37 .
drwxr-xr-x 99 stefan stefan 4096 2009-06-19 10:37 ..
-rw-r--r-- 1 stefan stefan 6 2009-06-19 10:37 a
-rw-r--r-- 1 stefan stefan 0 2009-06-19 10:37 b
#> cp -a * /mnt/Test/
#> ls -la /mnt/Test/
total 1024
drwxr-xr-x 2 stefan stefan 0 2009-06-19 10:38 .
drwxr-xr-x 9 stefan users 0 2009-06-19 10:26 ..
-rw-r--r-- 1 stefan stefan 6 2009-06-19 10:38 a
-rw-r--r-- 1 stefan stefan 0 2009-06-19 10:37 b
So at least the bad access rights and the error on cp are a regression since Hardy. The fact that the timestamp is only preserved for empty files seems to be consistent between both.
To decide whether this is rather a server or client side problem, the same test done on a current Hardy client (the server is Jaunty in both cases):
#> ls -la
total 12
drwxr-xr-x 2 stefan stefan 4096 2009-06-19 10:37 .
drwxr-xr-x 99 stefan stefan 4096 2009-06-19 10:37 ..
-rw-r--r-- 1 stefan stefan 6 2009-06-19 10:37 a
-rw-r--r-- 1 stefan stefan 0 2009-06-19 10:37 b
#> cp -a * /mnt/Test/
#> ls -la /mnt/Test/
total 1024
drwxr-xr-x 2 stefan stefan 0 2009-06-19 10:38 .
drwxr-xr-x 9 stefan users 0 2009-06-19 10:26 ..
-rw-r--r-- 1 stefan stefan 6 2009-06-19 10:38 a
-rw-r--r-- 1 stefan stefan 0 2009-06-19 10:37 b
So at least the bad access rights and the error on cp are a regression since Hardy. The fact that the timestamp is only preserved for empty files seems to be consistent between both.