GNOME Bugzilla – Bug 608908
gvfsd-smb: file content messed when using low-level write utilities (echo)
Last modified: 2011-10-17 11:40:30 UTC
Hello, I have recently faced a VERY strange behaviour with GVFS. This behaviour was not present before I updated Ubuntu Jaunty (Gnome 2.26, GVFS 1.2.2) to Karmic (Gnome 2.28.1,GVFS 1.4.2) Steps to reproduce behaviour: 1. Mount a GVFS-SMB share (via smb://someserver/someshare) 2. Open terminal and cd to this share 3. Run commands: user@host:~/.gvfs/someshare on someserver$ echo 100 > ./test user@host:~/.gvfs/someshare on someserver$ echo 200 >> ./test user@host:~/.gvfs/someshare on someserver$ echo 300 >> ./test user@host:~/.gvfs/someshare on someserver$ echo 400 > ./test 3. Expected file contents: 400 4. Real file contents: user@host:~/.gvfs/someshare on someserver$ cat ./test 400 200 300 Similar behaviour is when using a text editor writing with fopen (e.g. Geany: http://www.geany.org). When using gedit (which seems to write via g_file_set_contents), all is OK. I'm using Ubuntu 9.10 Karmic with Gnome 2.28.1 and GVFS 1.4.2 Is it a bug or an expected behaviour (e.g. all apps dealing with GVFS should use g_file_set_contents)?
Can you please test the patch posted in bug 627567? Also, appending to a file seems to work with smb only, but not for ftp or sftp.
potter, can you please respond to comment#1 ?
(In reply to comment #2) > potter, can you please respond to comment#1 ? Will test the patch as soon as I am near my Linux PC (2-3 days). According to bug 627567 the patch has been integrated into GNOME since 2.32, is that correct?
(In reply to comment #3) > According to bug 627567 the patch has been integrated into GNOME since 2.32, > is that correct? Yes, 1.6.6 release for gnome 2.32.x carries the patch (not vanilla 2.32.0).
(In reply to comment #1) > Can you please test the patch posted in bug 627567? Also, appending to a file > seems to work with smb only, but not for ftp or sftp. I do confirm that the bug has vanished at least since GNOME 2.32.1 and GVFS 1.8.0. I performed the same test that was described in my original post.
Great, thank you for confirming this.