When the userspace daemon mount.davfs dies before the file system is
unmounted, then the kernel file system fuse will hang forever and it is
not possible to normally unmount. I think this is a bug in fuse and
filed a report to the kernel bug tracker long ago.
https://bugzilla.kernel.org/show_bug.cg
I would prefer to report an error on unknown options to inform the user
about the error. But with fstab there are options that davfs2 has to
ignore because they may be required for other programs. To ignore some
options (but not all) the C-functions used by davfs2 to parse the
options would require
This looks like a problem in the communication with the server. You
will probably find the missing files in the lost+found directory.
I would be very surprised if this problem really depends on the version
of Debian.
Please first try to set option "if_match_bug 1" in your davfs2.conf
file (it wil
Thanks for the patch. I have applied it to the CVS-repository on
Savannah.
I suggest to apply this patch to the Debian package in Unstable too.
Am Wed, 17 Aug 2016 17:33:01 +0200
schrieb Harald Braumann :
> PS: I would have reported this upstream, but I'm not allowed to.
Because in my experien
Thanks for the additional information. It helped me to reproduce and
locate the bug.
I will release a new upstream version soon and then notify the Debian
maintainer.
Werner
This is a (very) late reply from upstream.
Sorry for the delay. I hoped somebody else would care because I had no
32-bit testing system.
Now I have upgraded an old Thinkpad from Jessie to Stretch installed
davfs2-1.5.2-1.2 from Sid and tested. Result:
Works as expected; no stack smashing error. I
Am Tue, 30 Sep 2014 14:24:18 -0400
schrieb Antoine Beaupré :
> How does davfs2 know which username to use to communicate with the
> WebDAV server then, especially considering the thing is ran as root?
Either from the secrets file or by asking you.
Questions in return:
Where do you take your inf
Am Mon, 29 Sep 2014 17:30:27 -0400
schrieb Antoine Beaupr? :
>
> Right now there are two ways of sending the password to the WebDAV
> server: the secrets file and the clunky username=foo mount option.
That's wrong.
Option "username= " is for use with pam_mount only. This is
documented in the mou
Package: libneon27-gnutls
Version: 0.29.6-3
Severity: normal
Dear Maintainer,
when libneon27-gnutls is used to connect to a HTTPS-sever it leaks about 4 KB
of memeory every time the connection is closed.
This bug was discovered by a user of davfs2 which - when running for a long
time - finally
0 +0200
+++ davfs2-1.4.6.new/ChangeLog 2013-09-15 11:05:42.0 +0200
@@ -1,6 +1,11 @@
ChangeLog for davfs2
+2013-09-08 Werner Baumann (werner.baum...@onlinehome.de)
+* kernel_interface.c, mount_davfs.c:
+ Don't create /dev/coda and /dev/fuse.
+
/ChangeLog 2013-09-15 11:05:42.0 +0200
@@ -1,6 +1,11 @@
ChangeLog for davfs2
+2013-09-08 Werner Baumann (werner.baum...@onlinehome.de)
+* kernel_interface.c, mount_davfs.c:
+ Don't create /dev/coda and /dev/fuse.
+ Remove insecure calls of system().
+
With Wheezy Debian replaced /etc/mtab with a symbolic link
to /proc/mounts. This causes trouble with davfs2.
Background:
davfs2 uses file system type 'davfs' in fstab and mtab. But it uses the
fuse kernel module (or coda). So in /proc/mounts the file system type is
'fuse' and the unmount helper um
This seems to be a duplicate of bug#693607.
The problem is caused by a bug in the LOCK implementation of box.com.
Please look at bug#693546 for details.
With box.com you should always set option 'use_locks 0' in davfs2.conf.
Werner
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.deb
This seems to be neither a Duplicity bug nor a davfs2 bug but one of
box.com's many bugs. This is what happens (as far as I can see):
- davfs2 gets an Open request for file#1
- davfs2 sends a LOCK request for this filename to box.com
- box.com returns "200 OK"
- davfs2 returns ok
- duplicity write
Did you try option "delay_upload 0" in davfs2.conf.
davfs2 will write the file to the server as soon as it is closed and
the file system call will only return when the the upload is complete.
But if the upload fails the file will stay in the cache and davfs2 will
retry later which may result in the
Another try to explain:
mount.davfs and umount.davfs are designed to be mount and umount
helpers for the system's mount and umount program. The intention is to
make mounting a davfs2 file system behave as much as possible like any
other file systems. You usually mount with 'mount ' and
unmount wit
This looks like a sudo problem not a davfs2 problem.
According to the sudo manpage a return value of 1 without any message
means that there is a configuration/permission problem with sudo.
Werner
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscr
Hello Joey,
the error message originates from umount(8) not from umount.davfs. It
usually happens when your system lacks /etc/mtab or when /etc/mtab is
replaced by /proc/mounts.
You can use mount option 'users' instead of 'user' to work around the
problem. But 'users' will allow any user to unmoun
In Debian this is a configuration option.
You should be asked at installation time whether to set the setuid bit
or not. You may change your decision at any time by invoking
'dpkg-reconfigure davfs2'.
Werner
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject o
Setting option '_netdev' in /etc/fstab should do the trick.
Please be aware that automatically unmounting a davfs2 file system at
shutdown is not always safe. If there are a lot of data in the cache
that have not yet been synchronized with the server davfs2 will try to
synchronize when beeing unmo
Hello Zbynek,
where did you take your information about option 'username' from?
This is what 'man mount.davfs' says:
username=WebDAV_user
Use this name to authenticate with the WebDAV server. This option
is intended for use with pam_mount only. When this option is set the
credentials in the
If you remove the trailing slash from the mount point in /etc/fstab it
should work (/home/charpent/alpha-dav instead
of /home/charpent/alpha-dav/).
It is a bug in davfs2 and I will fix it in the next release.
Directory ~/.davfs2 will only be created after some basic checks
succeeded. In this case
Am Sat, 26 Dec 2009 11:11:58 +0800
schrieb Jerome BENOIT :
> Nevertheless, the error message seems not related, and
That's what you usually get when there is a bug.
WebDAV-servers send a HTTP-status-code and clients like davfs2 have to
rely on it to check wether the request succeeded or what the
In upstream version 1.4.5 I added a log messages for this.
Please note:
In cases like this davfs2 only can make assumptions about the cause of
the error. There is always a chance that the message will give a hint
that is misleading.
Cheers
Werner
--
To UNSUBSCRIBE, email to debian-bugs-dist-r
Am Sat, 24 Oct 2009 12:29:17 +0100
schrieb Reuben Thomas :
> As I explained above, some FUSE file systems allow users in the fuse
> group to create mount points. Apparently, davfs2 does not, which makes
> it much less convenient to use.
>
> Would you agree that it is better for me to open a new b
Am Fri, 23 Oct 2009 22:06:15 -0300
schrieb Luciano Bello :
> Is this possible?
> If the error is a "File exists" you can append "or the server has a
> broken LOCK. Please try option 'use_locks 0' in your davfs2.conf
> file."
>
> There is no need to be literal. But a mention to use_locks looks
> n
Hello Luciano,
I added this to my TODO list but what I can do might not be what you
expect:
- in case of broken servers that send wrong information, davfs2 cannot
know whether it is a real problem or just wrong information from the
server. So the message could only be of the kind
"Either there is
Am Thu, 22 Oct 2009 00:37:04 +0100
schrieb Reuben Thomas :
> I think perhaps my original bug report was unclear. My problem was not
> strictly that I couldn't see how to mount a davfs2 filing system as a
> normal user, but that I couldn't see how to create a mount as a normal
> user (assuming of c
This is often caused by servers with a broken implementation of the
LOCK request (e.g. cPanel has this bug).
Please try option 'use_locks 0' inyour davfs2.conf file.
If this does not help please set option 'debug most' and send the log
messages from syslog or debug. As there will be a lot of debug
Some hints from me, the upstream maintainer.
This isn't a bug. You should not only read README.Debian, but also
README and the manpage ('man mount.davfs').
This is about how mounting is handled on Unix-like operating systems:
Normal (unpriveleged) users can not mount, only root can. But an entry
davfs2 is ignorant about the difference between libneon27 and
libneon27-gnutls. It takes whatever the libneon developer package offers
(using /usr/bin/neon-config to detect the neon library, no configure
option needed).
I don't know of any problems with using the gnutls version and to avoid
license
Thanks for the additional information. It helped to locate and fix one
more stupid bug.
This patch will fix it:
diff -Naur davfs2-1.3.3/src/mount_davfs.c davfs2-1.3.3.new/src/mount_davfs.c
--- davfs2-1.3.3/src/mount_davfs.c 2008-06-30 20:29:42.0 +0200
+++ davfs2-1.3.3.new/src/mount
The same error was reported on
http://sourceforge.net/tracker/index.php?func=detail&aid=2351083&group_id=26275&atid=386747
Unfortunately I can't reproduce this bug and the bug submitter stopped
responding.
If you can reproduce the problem, this might help to track down the problem:
1) Set optio
Package: davfs2
Version: 1.3.3-1
Severity: important
Function dav_head miscalculates the Last-Modified-Date. Under certain
circumstances this can prevent davfs2 from uploading a changed version
of a file to the server, or from downloading a new version.
The following patch fixes the problem.
Hello Luciano,
Luciano Bello wrote:
Can you send the info asked in http://bugs.debian.org/484300#10 ? Have you got
any problem to get it?
Unfortunately the supplier of the bug report did not respond to my
questions, so I can only guess:
There is another bug report concerning davfs2 and zope
It would be useful to have same debug information from davfs2. Please do
- add option "debug most" to your davfs2.conf-file
- mount the file system
- copy one (small) file into your davfs2 file system
- wait 20 seconds
- unmount the file system
The debug messages will be in /var/log/syslog. If th
CSights wrote:
Out of curiosity, has this been fixed in v 1.3 of program?
Yes.
It is fixed since version 1.2.2
Cheers
Werner
(upstream maintainer)
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Thanks for the report and the patch.
Unfortunately the patch does not work on Debian/Etch, as ps without the
-C option does not show command line arguments. But the checking I do is
paranoid anyway, so I removed the check for the mount point (still
checking PID and program name).
Can you tes
Hello Matieu,
mount.davfs does not *send* login/password before verifying the server
certificate.
The reason, why it asks the user in advance is: it can only ask the
user, as long as it is not in daemon mode. But request from the server
for credentials may come at any time.
For example: so
P.S.:
The real propable reason:
index.html is just a lot of java script (one of that famous web 2, java
script php applications). These java script programs may do almost
anything. When you edit something they usually try to to upload it using
HTTP POST.
What they will not do (and should no
Hello Loic,
I have tested.
My system: Etch, some Athlon board
davfs2 1.1.2-3
What I have done:
1. Mount the the resource
Create files testwb.html and testwb.xml (using gedit)
unmount the resource
2. Test with cadaver:
The files are on the server and are readable
3. Delete
Hello Christian,
this is a bug of the GMX-server. For details please see
http://sourceforge.net/tracker/index.php?func=detail&aid=1594808&group_id=26275&atid=386747
Please send a bug report to GMX. As I am not a customer of GMX they are
not willing to read my bug report; they offer their telefo
Comment by upstream maintainer:
The complete message ist:
Connection failed, mounting anyway. File system will only be usable
when connection comes up.
Question: Did the connection never come up?
In this case it would be useful to log HTTP-traffic, to see why the
server refuses the connectio
Hello Ferenczi,
this seems to be a problem with Apache content negotiation. Apache can
serve different files for the same URL, e.g. according to the prefered
language.
In your case Apache thinks tiny_mce.js is a specialized version of
tiny_mce, and your request for tiny_mce is served by send
Package: davfs2
Version: 0.2.7-1
Severity: wishlist
Hello Luciano,
we have got a feature request from Hubert on our davs2-help-forum
(http://sourceforge.net/forum/forum.php?thread_id=1436976&forum_id=82589)
concerning the Debian package:
"When being installed, the package should ask wether t
Hello,
there seems to be some confusion about the problem Martin Samuelsson
reported. So here is my (upstream) point of view:
It is no longer the IIS-bug. This is fixed in davfs2 0.2.8.
The problem only occurs with a custom WebDAV module used by Martins
provider. To see what is the 'failsafe
Hello Luciano, hello Yann,
now I am convinced it is FHS compliant, I have added code in davfs2:
davfs2 will now at start up test for the existence of
'var/run/mount.davfs' and create it if nonexistent.
I think doing this in mount.davfs is easier and more straight forward
then doing it in a scri
The problem reported by Martin on Thu, 12 Jan 2006, has be tracked down
by Martin and me. It showed to be a bug not in davfs2, but in the server
(not IIS but a custom WebDAV module).
So I think the interoperability issues with IIS are resolved by davfs2
0.2.7-1. For details please see
http://so
davfs2 could do that, but I am not clear whether it should. I think it
is a question of policy.
So I ask for advice from people who are more familiar with Debian policy.
Werner
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
I think I fixed the bug now. The bugfix is included in the sources in
our cvs repository (http://cvs.sourceforge.net/viewcvs.py/dav/davfs2/).
'#define _GNU-SOURCE' is part of configure.in and also config.h. But
'#include config.h' was not the first include directive in util.c, and
therefore '#
Hello Daniel,
thanks for your prompt reply.
You are right concerning the missing prototype. But there is still a
problem with the documentation. That's what I get:
Compiling example from " 25.3.13 Parsing of Suboptions Example" as is:
ginster:/home/werner/csource# gcc -Wall -o testprg tes
Package: glibc-doc
Version: 2.3.2.ds1-22
Severity: minor
-- System Information:
Debian Release: 3.1
Architecture: i386 (i686)
Kernel: Linux 2.6.11.6
Locale: [EMAIL PROTECTED], [EMAIL PROTECTED] (charmap=ISO-8859-15)
-- no debconf information
When compiling with #define _GNU_SOURCE function gets
Hello,
thanks for the patch.
I added it to the release-0-2-4 branch of our cvs-repository.
Greetings
Werner
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Hello,
although my explanation was not only plausible but fundamentally
correct, it missed one (unimportant) detail: the real bug in davfs!
davfs2 does not escape the path in its first request (OPTIONS) (in all
following requests it does). So the connection fails.
I append a patch for this.
I have got a similar problem when upgrading from samba-3.0.10 to
samba-3.0.14a.
Cups and samba are both started with S20-links. And I get the error
message in samba's log:
smbd version 3.0.14a-Debian started.
Copyright Andrew Tridgell and the Samba Team 1992-2004
[2005/07/04 08:00:30,
Hello David,
it is a shell problem. When typing the mount command you must escape
spaces in arguments and options with '\' to prevent the shell from
(mis)interpreting them. But you must not do any encoding with respect to
http, as this will be done by davfs2.
So you should write your url lik
-0.2.3.orig/ChangeLog davfs2-0.2.3/ChangeLog
--- davfs2-0.2.3.orig/ChangeLog 2005-06-03 21:03:13.0 +0200
+++ davfs2-0.2.3/ChangeLog 2005-06-04 13:53:20.0 +0200
@@ -1,5 +1,14 @@
ChangeLog for Davfs2
+2005-06-03 Werner Baumann
+security fix (quick and brutal) concerning
57 matches
Mail list logo
