Re: wput: bug-report

[Hagen Fritsch <hagen@itooktheredpill.dyndns.org>]

Am Mittwoch, 14. Juli 2004 16:53 schrieben Sie:
> I'm actually on the same server for src and dest, so there shouldn't(??) be
> any time difference.
So in this case set time-deviation to 0.

> In the debug file the timestamp comparisons are as follows:
>
> Correct behavior of not transferring unchanged files:
> timestamping: local: 1089809710 seconds; remote 1089809635 seconds; diff:
> 75 timestamping: local: 1089809710 seconds; remote 1089809635 seconds;
> diff: 75 timestamping: local: 1089809710 seconds; remote 1089809635
> seconds; diff: 75 timestamping: local: 1089809628 seconds; remote
> 1089809635 seconds; diff: 18446744073709551609 timestamping: local:
> 1089809628 seconds; remote 1089809635 seconds; diff: 18446744073709551609
These high numbers are actually -7, but i was to lazy to implement a signed 
debug output. -7 means that the remote is 7 seconds older than the local one 
and therefore wput doesn't transfer it. Actually the file is not 7 seconds 
older but time-deviation + 7 seconds, in your case 37 seconds older.
But in case there is a file that is less than 30 seconds newer, the difference 
will be positive.

> Have no idea why the diff on the two files that didn't get transferred is
> so large. I would have expected it to be 0 or close to it.
>
> Incorrect behavior where unchanged files were transferred:
> timestamping: local: 1089810085 seconds; remote 1089809995 seconds; diff:
> 90 timestamping: local: 1089810085 seconds; remote 1089809995 seconds;
> diff: 90 timestamping: local: 1089810085 seconds; remote 1089809995
> seconds; diff: 90 timestamping: local: 1089810003 seconds; remote
> 1089809995 seconds; diff: 8 timestamping: local: 1089810003 seconds; remote
> 1089809995 seconds; diff: 8
> Here the two unchanged files have a diff of 8, so I don't understand why
> they were transferred.
So the remote file is regarded as older and therefore wput transfers these 
files.

I'm quite sure that setting timedeviation to 0 should fix the problem. But it 
seems as this system is a little too complicated for end-users, so I should 
take a look at other implementation and maybe find a better solution.

-- 
Mit besten Grüßen
 Hagen                          mailto:itooktheredpill@gmx.de
                      http://itooktheredpill.dyndns.org