Firefox/Nautilus losing mtime stamp on file copy to folder

Bug #247980 reported by kpoole
4
Affects Status Importance Assigned to Milestone
gvfs
Fix Released
Medium
glib2.0 (Ubuntu)
Invalid
Undecided
Unassigned
gvfs (Ubuntu)
Fix Released
Low
Ubuntu Desktop Bugs

Bug Description

Under 7.10 and Firefox 2, there was a behavior available where an image could be dragged and dropped from a website, not all but some websites, to a folder displayed by Nautilus or the Desktop and the filename and last modified time stamp were preserved in the copy. Since upgrading to Ubuntu 8.04 and Firefox 3 the filename is still preserved but the last modified time stamp is now updated to the current system time.

If one uses wget to retrieve the file the last modified time stamp is preserved, which indicates that the remote mtime time stamp is still accessible, but this is a poor backup to what used to be a simple drag and drop operation.

As an example http://fc02.deviantart.com/fs9/i/2006/037/6/c/Fancy_Belt_by_baba49.jpg can be dragged from Firefox to a folder and gets the current system time as the last modified time stamp where a wget preserves the remote mtime stamp of 06 Feb 2006 07:22;21 PM.

You may not believe me but I swear these were preserved in the Ubuntu 7.10 with Firefox 2.

Addition: As a result of other discussions, I installed Ubuntu 7.10 on a spare machine and confirmed that the last modified time stamps were being preserved in this drag and drop operation. I also installed the Ubuntu 7.10 available Firefox 3.0 alpha and find the last modified time stamps are still preserved. I haven't found a Nautilus 2.22.3 backported to 7.10 to continue the test to see if the timestamps are lost when Nautilus 2.22.3 is introduced into the situation. I might have to go back to the Nautilus website and get the tar.gz and try to install it but that will be when I have some more time to devote to this.

Further addition: I have also installed FIrefox 2 on my Ubuntu 8.10 machine so I can be certain that under Ubuntu 7.10 with Nautilus 2.20.0 mtime stamps are preserved when copying a file from Firefox 2 or 3 to a Nautilus folder and that these same mtime stamps are lost when using Firefox 2 or 3 to copy a file to a Nautilus folder in Nautilus 2.22.3 under Ubuntu 8.04.

Last edit: Sorry to trouble everyone. We'll just call this a horrible mistake and I'll ask that someone cancel any further action on it. I can't find a cancelled status to give it.

Related branches

kpoole (ken-poole)
description: updated
kpoole (ken-poole)
description: updated
Revision history for this message
kpoole (ken-poole) wrote :

Sorry, Steve, don't know how this got here. I marked it invalid for you.

description: updated
Changed in glib2.0:
status: New → Invalid
kpoole (ken-poole)
Changed in firefox:
assignee: nobody → ken-poole
status: New → Invalid
kpoole (ken-poole)
description: updated
Revision history for this message
Tormod Volden (tormodvolden) wrote :

Why did you close the bug by marking it invalid? And do not assign it to yourself unless you intend to fix it yourself.

Changed in firefox:
assignee: ken-poole → nobody
status: Invalid → New
Revision history for this message
Tormod Volden (tormodvolden) wrote :

I can confirm the bug, the timestamp is lost when dragging and dropping from Firefox to the Desktop, but is preserved when I use wget. Not sure it is a Firefox bug, but the Firefox crowd might know who's to blame.

Changed in firefox:
status: New → Confirmed
Revision history for this message
kpoole (ken-poole) wrote : Re: [Bug 247980] Re: Firefox/Nautilus losing mtime stamp on file copy to folder

Tormod Volden wrote:
> Why did you close the bug by marking it invalid? And do not assign it to
> yourself unless you intend to fix it yourself.
>
> ** Changed in: firefox (Ubuntu)
> Assignee: kpoole (ken-poole) => (unassigned)
> Status: Invalid => New
>
>
Just for the record and because I really hate to leave questions
unanswered, having given up on the bug and being unfamiliar with the
system I hoped that I'd find an option to simply close a bug that was
assigned to me, likening the bug tracking system to the trouble tracking
systems I was used to. It's not.

Making it invalid was the last hope that the system would then clear it
out. I wouldn't expect the system to hang onto invalid bugs.

Thanks for confirming the bug.

Revision history for this message
Sebastien Bacher (seb128) wrote :

the issue is a gvfs one, see http://bugzilla.gnome.org/show_bug.cgi?id=547133 which has some details

Changed in firefox:
assignee: nobody → desktop-bugs
importance: Undecided → Low
status: Confirmed → Triaged
Changed in gvfs:
status: Unknown → Confirmed
Changed in gvfs:
status: Confirmed → Fix Released
Revision history for this message
Sebastien Bacher (seb128) wrote :

the bug has been fixed upstream now

Changed in gvfs:
status: Triaged → Fix Committed
Revision history for this message
Launchpad Janitor (janitor) wrote :

This bug was fixed in the package gvfs - 0.99.6-0ubuntu1

---------------
gvfs (0.99.6-0ubuntu1) intrepid; urgency=low

  * New upstream version:
   - Better cross-backend copy/move logic. Now will perform an actual file
     system move if possible, even when the source and target are handled
     by different backends
    - Now requires libsoup >= 2.23.91.
    - Bug fixes
    548841 - Incorrect usage of plural forms in gphoto backend
    547133 - Getting mtime over http backend is broken (lp: #247980)
    538573 - lastmodified uses ISO 8601 date/time where it should use http-date
    549253 - error path leaks
    549553 - gvfs mangles uri for unhandled schemes
    550100 - gio uses 10^3 base (SI) for volume names
    529971 - Restore from trash appears to do a file copy (lp: #222150)
    RH 460223 - gnome-mount no longer automatically opens LUKS-encrypted
    partitions

 -- Sebastien Bacher <email address hidden> Tue, 02 Sep 2008 11:49:45 +0200

Changed in gvfs:
status: Fix Committed → Fix Released
Revision history for this message
kpoole (ken-poole) wrote :

Launchpad Bug Tracker wrote:
> This bug was fixed in the package gvfs - 0.99.6-0ubuntu1
>
> ---------------
> gvfs (0.99.6-0ubuntu1) intrepid; urgency=low
>
> * New upstream version:
> - Better cross-backend copy/move logic. Now will perform an actual file
> system move if possible, even when the source and target are handled
> by different backends
> - Now requires libsoup >= 2.23.91.
> - Bug fixes
> 548841 - Incorrect usage of plural forms in gphoto backend
> 547133 - Getting mtime over http backend is broken (lp: #247980)
> 538573 - lastmodified uses ISO 8601 date/time where it should use http-date
> 549253 - error path leaks
> 549553 - gvfs mangles uri for unhandled schemes
> 550100 - gio uses 10^3 base (SI) for volume names
> 529971 - Restore from trash appears to do a file copy (lp: #222150)
> RH 460223 - gnome-mount no longer automatically opens LUKS-encrypted
> partitions
>
> -- Sebastien Bacher <email address hidden> Tue, 02 Sep 2008 11:49:45
> +0200
>
> ** Changed in: gvfs (Ubuntu)
> Status: Fix Committed => Fix Released
>
>
I notice this is fixed in Intrepid. I'm running Heron and had planned
to for a while yet. Is there any hope that this will be backported to
Heron?

I'm going to download the Intrepid beta to see if this is really fixed
as expected and decide from there if this is enough to switch from heron
earlier than planned. I have found an inelegant workaround in a Firefox
add-on while I'm in Heron.

Revision history for this message
Halifax (ckwcheng) wrote :

I've had problems with Beta 2, similar to this.

When I copy files from my windows machine via SCP or SFTP files to my server, or drag and drop via windows explorer, most of the time the file names have preserved timestamps. The odd occasion, the time stamps change to the current time.

What is really bothersome is that the timestamp on my file folders change to the current time, even if they were created years ago. The creation date and the modification date become the same (i.e. the current date).

This is quite the bug, and is really bad for trying to archive things or find things by date.

Changed in gvfs:
importance: Unknown → Medium
Revision history for this message
kpoole (ken-poole) wrote :

Okay, this has gotten silly. The problem was reported, a fix released for Intrepid but not backported to Hardy two years ago. The last previous update before it was just upgraded to medium was over a year ago and I think all of us LTS types have probably upgraded from Hardy to Lucid by now. I actually just did last weekend after waiting for the point release. I've tested the behaviour I was reporting as a bug and find that it works as expected in Lucid.

Let's just call this fixed and close it in such a way that it can't be updated again.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.