GNOME Bugzilla – Bug 578892
file-roller only creates empty folder when extracting to network-share
Last modified: 2010-05-18 13:42:13 UTC
Please describe the problem: "In Gnome File Browser (Nautilus 2.26.1), you can right-click on an archive file (.zip, .tar.gz, .tgz) and select "Extract Here". This works as expected in a local folder. When trying to use "Extract Here" on an archive file over a Network Folder (using Connect To Server...), this only creates an empty folder. I can reproduce this using a .zip, .tgz, .tar.gz files using a network share either over SMB or SSH. I haven't tried other archive types or mount types. I have tried with several different archive packages." Same happens when opening the archive with file-roller and extracting from there. Steps to reproduce: 1. Open network (smb/ssh) shared folder 2. right-click "extract here" on archive Actual results: Empty Folder is created Expected results: The folder should contain archive contents. Does this happen every time? Yes Other information: Reported on Launchpad: https://bugs.launchpad.net/bugs/360598
Ok, got some more Information. When opening the archive on the remote I get: Gtk-WARNING **: Operation not supported by backend While extracting, there is a (file-roller:18563): GLib-GIO-CRITICAL **: g_file_new_for_uri: assertion `uri != NULL' failed Could this be a GVFS-issue?
Got the same with an FTP if you click on extract there's actually no extraction and the same message is throw on the console.
Here is a backtrace for this with G_DEBUG=fatal_criticals
+ Trace 214821
I believe this is fixed in master by the following commit: commit 33c21fc3647af32de5a9c9d49e33765a526910f5 Author: Tomas Bzatek <tbzatek@redhat.com> Date: Wed May 12 13:54:29 2010 +0200 Bring back support for file caching when gvfs-fuse-daemon is not available The commit 73fa5d77a60861cd35c6f3425d27c88061af081c introduced trouble when user was trying to work with archives on remote GIO mounts and not having the fuse daemon running. This led to segfaults and error messages. Please see bug #617769 for details. I'm going to close this bugreport now, feel free to reopen it if the problem persists.