Hi, Peter!
You are right. There is no way to transactionally create a directory :(
Oliver
2007/7/5, Peter Coppens [EMAIL PROTECTED]:
Hello,
I am just starting with the Commons Transaction API...so apologize my
ignorance.
Now, is there anyway I can create (transactionally) a directory? I
Folks!
Does anyone really *need* transactional file operations?
Doesn't the existing commons transaction implementation add way too
much complexity to your applications?
Couldn't you just write to the files and programmatically keep the
previous version for recovery from fatal errors?
I'd
Dear community!
Jakarta Commons Transaction 1.2 has been released. This is a bug
fix/maintenance/feature release of Commons Transaction. A lot of bug
fixing work from a number of people has gone into this release. This
makes Commons Transaction more stable and reliable. Additionally, some
nice
Hi Paul!
The short answer is: no.
The long answer is: That is not the original idea of a FileSequence.
But you could easily write something like that using the ordinary
FileResourceManager: Initially create a file using the manager, write
0 as it's content. When incrementing just open the file,
2006/5/24, Holger Hoffstaette [EMAIL PROTECTED]:
On Tue, 23 May 2006 23:23:13 +0200, Oliver Zeigermann wrote:
However I think performance could be considerably improved if there were
to be a movement over to one of Doug Lea's kinds of Concurrent Hashmaps,
probably from his older oswego
Hi Antranig!
2006/5/23, Antranig Basman [EMAIL PROTECTED]:
The most straightforward of them is in GenericLockManager, where
the lock owner is not being released properly - I have
the following:
Index:
Hi, again!
2006/5/23, Antranig Basman [EMAIL PROTECTED]:
Oliver Zeigermann wrote:
Hi Antranig!
Hi!
2006/5/23, Antranig Basman [EMAIL PROTECTED]:
The most straightforward of them is in GenericLockManager, where
the lock owner is not being released properly - I have
the following
I look after it and unfortunately seem to be the only remaining one
who does so :(
Shall I include the maven release in the release?
Oliver
On 8/19/05, Brett Porter [EMAIL PROTECTED] wrote:
I'm not sure who looks after commons-transaction, but it should be ok.
This doesn't require anything
object so both classes could obtain
the same reference to the manager object. Correct?
LeRoy
Oliver Zeigermann
oliver.zeigerman
[EMAIL PROTECTED]
To
Jakarta
Most likely your problem ist that new File(..) creates different
objects in each thread. I would try using something like the path to
the file as String.
Oliver
On 7/5/05, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote:
I'm trying to use the ReadWriteLock class to acquire a write lock on a
] [EMAIL PROTECTED] wrote:
Oliver,
I tried your suggestion by changing my ReadWriteLock statement to
ReadWriteLock fileLock = new ReadWriteLock(c:/logRec.txt,loggerFacade);
However, I received the same results as when I used new File(..).
LeRoy
Oliver
a singleton map of locks. Surely only the
'resourceId' must be the same, and not the actual ReadWriteLock reference?
LockManager: Encapsulates creation, removal, and retrieval of locks. Each
resource can have at most a single lock.
Aaron
Oliver Zeigermann wrote:
Ooops, sorry, you are right
By the way, when giving me first advice I stumbled over exactly this
difference between the lock manager and the lock itself...
Oliver
On 7/6/05, Oliver Zeigermann [EMAIL PROTECTED] wrote:
You are right, but only for the lock manager. The lock manager takes
care of uniquely mapping a resource
across threads. That's very
interesting.
Thanks Oliver,
Aaron
Oliver Zeigermann wrote:
On 6/27/05, Aaron Hamid [EMAIL PROTECTED] wrote:
I will use Thread.currentThread() for the owner. I'm not clear as to the
utility of non-thread owners...what is the semantics of synchronization
Hi Aaron!
On 6/24/05, Aaron Hamid [EMAIL PROTECTED] wrote:
Hi all,
I suspect I just reimplemented FileResourceManager, because I was not aware
of it until recently (despite using the
org.apache.commons.transaction.locking package). Basically I created a
hierarchical Node interface,
On 6/27/05, Aaron Hamid [EMAIL PROTECTED] wrote:
I will use Thread.currentThread() for the owner. I'm not clear as to the
utility of non-thread owners...what is the semantics of synchronization if
the owners are not threads? Or is the intention that owners ultimately must
be associated
:
Great stuff, guys! Congratulations!
Any plans to implement a JTA resource adapter for files, to allow the
participation in distributed transactions, for the near future?
Regards,
Catalin
Oliver Zeigermann wrote:
The Commons Community is proud to announce the final release of
Commons
The Commons Community is proud to announce the final release of
Commons Transaction 1.1.
Commons Transaction provides utility classes commonly used in transactional
Java programming.
Version 1.1 aims at polishing (interface) oddities, improving
locking and making the file store more flexible.
On 5/23/05, David J. M. Karlsen [EMAIL PROTECTED] wrote:
Oliver Zeigermann wrote:
On 5/21/05, David J. M. Karlsen [EMAIL PROTECTED] wrote:
Hmm, yeah, but still - I cannot see why txId should be unique, the code is:
public String generatedUniqueTxId() throws ResourceManagerSystemException
On 5/20/05, David J. M. Karlsen [EMAIL PROTECTED] wrote:
Hi!
I keep getting ResourceManagerExceptions with Duplicate transaction
id. I took a look at the implementation of FileResourceManager - and
generatedUniqueID only returns system.getCurrentMillis() - which isn't
random enough in a
On 5/11/05, Niklas Gustavsson [EMAIL PROTECTED] wrote:
when using FileResourceManager, you need to supply a storeDir. Now, if I
need to write files to different locations within the same transaction.
For example c:\somedir\somesubdir\file.txt and
c:\someotherdir\someotherfile.txt, could I
On Apr 6, 2005 2:13 AM, Mélanie Bats [EMAIL PROTECTED] wrote:
Hello,
Oliver Zeigermann a écrit :
All this would be in the lock manager and would not require changes to
the lock classes.
In order to understand correctly how to implement the solution to my
problem, I looked
PROTECTED] [EMAIL PROTECTED] wrote:
Hi,
Selon Oliver Zeigermann [EMAIL PROTECTED]:
I suppose what you want is some sort of hierarchical lock? Like if a
parent is locked the children are locked as well? Is that true?
Yes it is something that looks like this.
Do you have some ideas
Hi Mel,
I suppose you pass the interval as the resourceId, right? The problem
is that normally a single locks gets assigned to a resourceid. I do
not think there is an easy solution if your intervals are not
discrete. But if they are you could just enumerate all values as
resourceIds and try to
On Apr 5, 2005 11:24 AM, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote:
Hi,
Selon Oliver Zeigermann [EMAIL PROTECTED]:
I suppose you pass the interval as the resourceId, right?
Yes exactly.
The problem is that normally a single locks gets assigned to a resourceid. I
do not think
1, 2005 9:31 AM, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote:
Hi,
Selon Oliver Zeigermann [EMAIL PROTECTED]:
Hmmm, not quite sure that I get this all right, but to me it seems you
describe the simple read/write locks (without the upgrade step). Is
that possible? If so you could simply use
If so that's exactly what the read/write lock and then the read/write
lock manager does.
Oliver
On Apr 1, 2005 6:30 PM, Oliver Zeigermann [EMAIL PROTECTED] wrote:
Still not quite clear, most likely because I am a bit dumb ;) What is
the use of your kind of upgrade lock? Or is it that you want
Hmmm, not quite sure that I get this all right, but to me it seems you
describe the simple read/write locks (without the upgrade step). Is
that possible? If so you could simply use the ReadWriteLockManager.
Oliver
On Thu, 31 Mar 2005 14:34:24 +0200, [EMAIL PROTECTED]
[EMAIL PROTECTED] wrote:
Everything you describe seems to be right to me.
Oliver
On Thu, 24 Mar 2005 09:30:59 +0100, [EMAIL PROTECTED]
[EMAIL PROTECTED] wrote:
Hi,
I'm a newbie with the transaction.locking package but I'm very interrested in
using the ReadWriteUpgradeLockManager class.
In my project, I need to
Not quite sure what you are planning. Are you talking about a JCA
connector for files?
Oliver
On Sun, 20 Mar 2005 10:12:50 +1100, Mason, Ross [EMAIL PROTECTED] wrote:
Hi,
I have been looking at using commons-transaction for the Mule project
(http://mule.codehaus.org) for managing file and
] wrote:
In the same way that you have XA support for VM transactions you could do
something similar for File-based transactions. These transactions could then
be managed by a Jca container or programmatically.
Ross
-Original Message-
From: Oliver Zeigermann [mailto:[EMAIL PROTECTED
On Sat, 05 Feb 2005 17:25:13 +0900 (JST), shirasu hiroyuki
[EMAIL PROTECTED] wrote:
From: shirasu hiroyuki [EMAIL PROTECTED]
Subject: [xmlio] On method signature of startElement
Date: Sat, 05 Feb 2005 17:15:29 +0900 (JST)
snip/
Why is it org.xml.sax.Attributes?
Sorry.
Why isn't it
The Commons Community is proud to announce the second beta release of
Commons Transaction 1.1!
Commons Transaction provides utility classes commonly used in transactional
Java programming.
The second beta does not add new features, but mainly fixes known
bugs, adds some tests, and provides more
AFAIK nothing like that scheduled. I was wondering how would this look
like for something ordered like a list?
What would happen if transaction a added a an enttry at position 0 and
transaction b another one at position 1? Would this even be allowed in
parallel? How would the resulting list look
Folks,
for anyone interested this is the current status report of commons transaction:
1.1 heading to beta1
--
Concerning my work the API extensions for 1.1 have been finished. Same
thing for new features like
- deadlock detection (hopefully, this was a major pain in
. Saving locks themselves should really be protected by 1.4
file locks as well...
Oliver
On Thu, 6 Jan 2005 23:34:09 +0100, Oliver Zeigermann
[EMAIL PROTECTED] wrote:
On Thu, 06 Jan 2005 17:21:44 -0500, Brant Boehmann
[EMAIL PROTECTED] wrote:
Single node is probably not an option because it would
Hi Brant,
1) This is currently not supported, but there was an approach to
achieve this with java.nio.channels.FileChannel#lock from JDK 1.4. It
was, however, abandoned as commons tx should remain compatible to
JDK1.2. On the other hand such an extension would be pretty simple and
could be
of the common tx package...
Oliver
On Thu, 6 Jan 2005 21:51:26 +0100, Oliver Zeigermann
[EMAIL PROTECTED] wrote:
Hi Brant,
1) This is currently not supported, but there was an approach to
achieve this with java.nio.channels.FileChannel#lock from JDK 1.4. It
was, however, abandoned as commons
And, finally, thinking this to an end, you may want to consider using
Jakarta Slide which is such a server, uses the FileResourceManager,
speaks WebDAV as a standard and already has this JCA implementation I
was talking about...
Oliver
On Thu, 6 Jan 2005 22:37:26 +0100, Oliver Zeigermann
[EMAIL
On Thu, 06 Jan 2005 17:21:44 -0500, Brant Boehmann
[EMAIL PROTECTED] wrote:
Single node is probably not an option because it would be a single point
of failure. (I realize a shared disk is also a single point of failure,
but in a somewhat different way as the shared disk I utilize resides on
a
On Mon, 3 Jan 2005 12:55:13 -0700, David Erickson [EMAIL PROTECTED] wrote:
Couple Questions regarding the new Transaction package, specifically the
FileResourceManager.
It was a bit vague in the FRM docs, but it talked about threaded access to
the FRM object, and left me still wondering. I
On Tue, 4 Jan 2005 14:10:45 -0700, David Erickson [EMAIL PROTECTED] wrote:
Thanks for the reply Oliver. I was also wondering is it possible to create
directories with this package? IE using my below example if I needed to
create /4 for a total of /storage/content/4 (4 being a directory)?
:
In fact at the most opportunistic level I'd like to integrate this with JTA,
since I have some database updates that depend on files being written
correctly etc. Has anyone attempted to plug this in with JTA?
-David
-Original Message-
From: Oliver Zeigermann [mailto:[EMAIL PROTECTED
Not quite sure what you mean with using locking?!
Oliver
On Tue, 4 Jan 2005 15:35:34 -0700, David Erickson [EMAIL PROTECTED] wrote:
Excellent thanks for the information Oliver. Are there any implementations
of this using both the FRM and locking on the files online? Specifically
I'd like
Hi Antranig,
I have now added the interface (plus initial implementations) you proposed:
http://cvs.apache.org/viewcvs.cgi/jakarta-commons/transaction/src/java/org/apache/commons/transaction/file/ResourceIdToPathMapper.java?rev=1.1view=log
I have not taken over getAllPaths, as it did not occur
The Commons Community is proud to announce the the final release of
Commons Transaction 1.0.
Commons Transaction has recently been promoted from the Commons
Sandbox and provides utility classes commonly used in transactional
Java programming.
Download page:
The Commons Transaction Cummunity is prood to announce the immediate
availability of the first release candidate of Commons Transaction
1.0.
Commons Transaction has recently been promoted from the Commons
Sandbox and provides utility classes commonly used in transactional
Java programming.
You mean a lock manager? Hmmm, I'd say locking in transaction is much
more low level. Let's see what Mario comes up with...
Oliver
On Wed, 24 Nov 2004 17:30:36 +0100, Eric Pugh [EMAIL PROTECTED] wrote:
Isn't this what [transaction] is supposed to do?
-Original Message-
From:
On Wed, 24 Nov 2004 18:17:29 +0100, Mario Ivankovits [EMAIL PROTECTED] wrote:
Daniel Florey wrote:
I'd recommend to clearify the scope of [vfs].
Is it supposed to be a very simple api to access different file systems?
Or is the scope larger so that we want to have abilities to manage
On Thu, 11 Nov 2004 18:16:53 +, Antranig Basman
[EMAIL PROTECTED] wrote:
Oliver Zeigermann wrote:
Another question - I'm looking over the locking strategy for FRM
and am keen that I can ensure it blocks rather than throwing
Yes, serializability (complicated word, hope the spelling
On Fri, 12 Nov 2004 17:44:50 +, Antranig Basman
[EMAIL PROTECTED] wrote:
Oliver Zeigermann wrote:
Sorry for causing confusion, these lock levels are just internal ones.
For you the
public boolean lockResource(Object resourceId, Object txId, boolean
shared) throws
Hi Antranig,
thanks again for your feedback!
On Thu, 11 Nov 2004 14:08:59 +, Antranig Basman
[EMAIL PROTECTED] wrote:
Thanks - I will proceed to use FRM in all confidence :)
I guess you can...
Yes, I could provide a patch for this, although expanding the
filesystem limit will not fall
52 matches
Mail list logo