[Tickets #11999] Re: Prevent inserting empty share "objects" into $all

Discussion:

[Tickets #11999] Re: Prevent inserting empty share "objects" into $all_shares array

(too old to reply)

n***@bugs.horde.org

2013-01-28 10:57:12 UTC

DO NOT REPLY TO THIS MESSAGE. THIS EMAIL ADDRESS IS NOT MONITORED.

Ticket URL: http://bugs.horde.org/ticket/11999
------------------------------------------------------------------------------
Ticket | 11999
Updated By | Thomas Jarosch <***@intra2net.com>
Summary | Prevent inserting empty share "objects" into
| $all_shares array
Queue | Horde Framework Packages
Version | Git master
Type | Bug
State | Unconfirmed
Priority | 1. Low
Milestone |
Patch | 1
Owners |
+New Attachment |
0001-Fix-11999-Prevent-inserting-empty-share-objects-into.patch
------------------------------------------------------------------------------

Thomas Jarosch <***@intra2net.com> (2013-01-28 10:57)
uploaded:
0001-Fix-11999-Prevent-inserting-empty-share-objects-into.patch

http://bugs.horde.org/h/services/download/?module=whups&actionID=download_file&file=0001-Fix-11999-Prevent-inserting-empty-share-objects-into.patch&ticket=11999&fn=%2F0001-Fix-11999-Prevent-inserting-empty-share-objects-into.patch

--
bugs mailing list
Frequently Asked Questions: http://wiki.horde.org/FAQ
To unsubscribe, mail: bugs-***@lists.horde.org

n***@bugs.horde.org

2013-02-01 20:19:12 UTC

Permalink

DO NOT REPLY TO THIS MESSAGE. THIS EMAIL ADDRESS IS NOT MONITORED.

Ticket URL: http://bugs.horde.org/ticket/11999
------------------------------------------------------------------------------
Ticket | 11999
Updated By | Michael Rubinsky <***@horde.org>
Summary | Prevent inserting empty share "objects" into
| $all_shares array
Queue | Horde Framework Packages
Version | Git master
Type | Bug
-State | Unconfirmed
+State | Feedback
Priority | 1. Low
Milestone |
Patch | 1
Owners |
------------------------------------------------------------------------------

Michael Rubinsky <***@horde.org> (2013-02-01 20:19) wrote:

I'm not so sure we should be silently ignoring any case where we can't
load an explicitly requested share.

Do you know *why* at least one of the shares requested in $cids was
not found by _getShares()?

--
bugs mailing list
Frequently Asked Questions: http://wiki.horde.org/FAQ
To unsubscribe, mail: bugs-***@lists.horde.org

n***@bugs.horde.org

2013-02-04 08:22:15 UTC

Permalink

Post by n***@bugs.horde.org
I'm not so sure we should be silently ignoring any case where we
can't load an explicitly requested share.

Should we throw an error instead?

Post by n***@bugs.horde.org
Do you know *why* at least one of the shares requested in $cids was
not found by _getShares()?

Unfortunately I couldn't properly trace it as it was a busy productive system
and I didn't want to add too many debug printfs() in there. Every time
I forget an ";", a few users end up with a blank page :o)

--
bugs mailing list
Frequently Asked Questions: http://wiki.horde.org/FAQ
To unsubscribe, mail: bugs-***@lists.horde.org

n***@bugs.horde.org

2013-02-04 14:53:18 UTC

Permalink

Post by n***@bugs.horde.org

Post by n***@bugs.horde.org
I'm not so sure we should be silently ignoring any case where we
can't load an explicitly requested share.

Should we throw an error instead?

Horde_Exception_NotFound should be thrown from the concrete share
class' _getShares() method.

--
bugs mailing list
Frequently Asked Questions: http://wiki.horde.org/FAQ
To unsubscribe, mail: bugs-***@lists.horde.org

n***@bugs.horde.org

2013-02-08 09:34:55 UTC

Permalink

Post by n***@bugs.horde.org

Post by n***@bugs.horde.org
I'm not so sure we should be silently ignoring any case where we
can't load an explicitly requested share.

Should we throw an error instead?

Horde_Exception_NotFound should be thrown from the concrete share
class' _getShares() method.

I've checked the Kolab Share code and it does thrown an exception if
one requests a share id that cannot be found.

While debugging this, I remember from my debug output that a rather
long and a short share id was requested. Wild guess: The short id was
just the username.

I think this might be a Kolab specific problem, call stack is roughly
like this:

Base->getShares($ids) -> Kolab->_getShares($ids) -> Kolab->_getShareById($id).

The found share is stored with "$share->getName()" as the key.

Wild guess: For the Kolab case, "$share->getName()" might return
something different than the request id. For example, you just
requested the username as id and get an IMAP folder name back.

The Base->getShares() function assumes the returned array of share
objects strictly uses the id as array key.

@Jan: Is it by design that "Kolab->_getShares($ids)" uses
"$share->getName()" as key?
The SQL driver seems to work this way.

The "getName()" function in Share/Object/Kolab.php uses getId() as
fallback if "share_name" is not available. So this some kind of
fallback automagic...

--
bugs mailing list
Frequently Asked Questions: http://wiki.horde.org/FAQ
To unsubscribe, mail: bugs-***@lists.horde.org

n***@bugs.horde.org

2013-03-19 14:04:33 UTC

Permalink

--
bugs mailing list
Frequently Asked Questions: http://wiki.horde.org/FAQ
To unsubscribe, mail: bugs-***@lists.horde.org

n***@bugs.horde.org

2013-08-27 10:41:44 UTC

Permalink

n***@bugs.horde.org

2013-12-18 10:49:48 UTC

Permalink

ping?

ok, I found something out about this:

The Kolab backend crashes a horrible death if you turn the INBOX of a
user into a groupware folder. This does not make any sense at all, but
one user managed to set his INBOX to the kolab folder type "notes".

-> We crash badly with this.

I think that was the same issue here. Too bad I don't have the "id"
anymore that caused it as it's a base64 encoded serialized php array
-> we could easily reverse it and take a look.
That would also explain the "short" IDs I've been seeing.

--
bugs mailing list
Frequently Asked Questions: http://wiki.horde.org/FAQ
To unsubscribe, mail: bugs-***@lists.horde.org