Erik Huelsmann wrote:
> Lieven,
>
> I'm on a conference, so I cannot go into detail here, but the problem
> described below is encountered half the time on FAT (because it has a
> 2 second time resolution). [We should flat out state we don't support
> it.]
>
> Maybe that's what the user is experiencing?
>
> Bye.,
>
> Erik.
>
>
> On Tue, May 20, 2008 at 4:11 PM, <lgo_at_tigris.org> wrote:
>> http://subversion.tigris.org/issues/show_bug.cgi?id=3190
>>
>>
>>
>>
>>
>>
>> ------- Additional comments from lgo_at_tigris.org Tue May 20 07:11:55 -0700 2008 -------
>> I've checked your script, seems you're talking about commit instead of checkout.
>>
>> Subversion commit - and other commands - will wait till the next second before
>> exiting. This ensures that the first modification you make to a file will
>> atleast be one second younger as the base file. This shows when I run your
>> script on my machine (Ubuntu linux, ext3 fs), as it doesn't show the problems
>> you're encountering.
>>
>> Can you add these lines in your script right after the 'svn commit' line and see
>> what timestamps you get?
>>
>> "
>> echo '*** Current time'
>> date --iso-8601=seconds -u
>> echo '*** Timestamp of last file written during post-commit'
>> echo '*** Should be atleast one second earlier than current time!'
>> date --iso-8601=seconds -u -r trunk/dir1/file1.txt
>> "
>>
Erik,
This is ReiserFS on Linux, not FAT. I don't know a lot about ReiserFS,
but I wonder if this is an effect of the chosen journaling option.
It would be nice if we can find out the reason and possible workaround
for this issue though, if only for future reference.
Lieven
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe_at_subversion.tigris.org
For additional commands, e-mail: dev-help_at_subversion.tigris.org
Received on 2008-05-20 16:45:53 CEST