Eclipse Community Forums
Forum Search:

Search      Help    Register    Login    Home
Home » Eclipse Projects » NatTable » next release
next release [message #1221706] Wed, 18 December 2013 04:12 Go to next message
Arjan Kok is currently offline Arjan Kok
Messages: 29
Registered: July 2009
Junior Member
hi Dirk,

Do you know already when if and when there will be a next release of the NatTable? I remember that you mentioned this in another thread some time ago.

Kind Regards, Arjan Kok.
Re: next release [message #1221725 is a reply to message #1221706] Wed, 18 December 2013 04:54 Go to previous messageGo to next message
Dirk Fauth is currently offline Dirk Fauth
Messages: 1213
Registered: July 2012
Senior Member
Hi,

on my list are 7 topics that I want to finish prior releasing the next and potentially final release of the current NatTable architecture. Some of them are minor changes, some of them need some bigger implementation and testing. Being the only one implementing and testing stuff in NatTable, I'm not able to tell when I'm done. I want the next release to be really stable, so any help will be appreciated and will decrease the time until the next release can be rolled out.

Greez,
Dirk
Re: next release [message #1229803 is a reply to message #1221725] Fri, 10 January 2014 03:46 Go to previous messageGo to next message
Paweł Doleciński is currently offline Paweł Doleciński
Messages: 36
Registered: January 2014
Member
Hi Dirk,

I would like you ask you actually the same question.

Can I find somewhere a progress status or whatever? Do you consider a next release in months or rather weeks?
Is your 7 topics list available somewhere?

I am asking you because right now with my team we are using NatTable quite a lot as a main component. It would be even possible from our side to support you with the implementation.
Moreover, there is one topic in which we are interested the most, that is the "group by" feature which has some performance issues according to our tests.

I've got one additional question. What about further implementation connected with changing architecture and support for instance JavaFX? Is it that you are the only committer too or not? And again, is it planed to be done in years or rather months?

Cheers,
Paweł.

Re: next release [message #1229924 is a reply to message #1229803] Fri, 10 January 2014 09:58 Go to previous messageGo to next message
Dirk Fauth is currently offline Dirk Fauth
Messages: 1213
Registered: July 2012
Senior Member
Hi,

to be more precise, ATM I am the only ACTIVE committer in the project. Hopefully that changes soon again.

As we are tracking every task in Bugzilla you can find the tasks I want to take care of via this Bugzilla search:

https://bugs.eclipse.org/bugs/buglist.cgi?list_id=7907332&classification=Technology&emailtype1=substring&query_format=advanced&emailassigned_to1=1&bug_status=UNCONFIRMED&bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&email1=dirk.fauth&product=NatTable

Some of them are almost done, I just need verification that it is really done. So that would be helpful too.

When does the performance issues in groupby appear? As this is currently something I was working on with the help of others, it might be something to consider also for the upcoming release.

Regarding JavaFX support, that is considered for the new architecture. First we need to finish 1.1 to have a stable version we don't need to take care of much anymore. We have some plans, designs and so on, but we are not far enough. Without much support, there is no progress of course.

So to answer your questions in short:
I hope to release NatTable 1.1 in the next weeks.
Plans for NatTable 2.x can be done when 1.1 is released.

Greez,
Dirk
Re: next release [message #1229940 is a reply to message #1229924] Fri, 10 January 2014 10:54 Go to previous messageGo to next message
Dirk Fauth is currently offline Dirk Fauth
Messages: 1213
Registered: July 2012
Senior Member
OK, I can see a performance issue for huge data sets when using GroupBy. It is related to GroupByDataLayer.updateTree()

I added that method a while ago to update the tree structure when the TreeList.Format has changed. This happens e.g. on applying a grouping, filter or sorting.

See the following links why I added that:
http://java.net/jira/browse/GLAZEDLISTS-521
http://glazedlists.1045722.n5.nabble.com/sorting-a-treelist-td4704550.html

If you know a better solution to get the update working correctly, let me know.
Re: next release [message #1230042 is a reply to message #1229940] Fri, 10 January 2014 16:21 Go to previous messageGo to next message
Paweł Doleciński is currently offline Paweł Doleciński
Messages: 36
Registered: January 2014
Member
Thanks for the comprehensive answer, Dirk. This is exactly what I expected.

I've done a quick glance on the issues list and have to say I am interested in most of them to be completed more or less asap.
So if you don't mind, I would like to somehow be involved, at least by verification.

About the new architecture, so far that info is enough for me to make decisions. This is because, in a not so far future, we are interested in switching renderers in our e4 based app.

What do you mean by huge data sets? I see problems when having even 10k rows. It is not much I guess.
Everything is fine when using GroupBy demo. Problems come up when GroupBy is added to the "Everything" demo (not saying about blinking and group by together).
Maybe I did a wrong configuration, do not know.

Would you mind if I created an issue on bugzilla with further discussion and investigation?
If the problem lays in update (and my observation somehow confirms yours), then a better solution or at least working workaround seems to be extremely important to me.

Anyway, I have to really thank you for your and the others job!

Cheers,
Paweł.
Re: next release [message #1230842 is a reply to message #1230042] Mon, 13 January 2014 02:56 Go to previous messageGo to next message
Dirk Fauth is currently offline Dirk Fauth
Messages: 1213
Registered: July 2012
Senior Member
Quote:
So if you don't mind, I would like to somehow be involved, at least by verification.


Why should I? Your welcome to support in any way you like! Smile

Using this link you will find all tickets that have been resolved since 1.0.1.

https://bugs.eclipse.org/bugs/buglist.cgi?list_id=7918183&classification=Technology&query_format=advanced&bug_status=RESOLVED&product=NatTable&target_milestone=1.1.0

It should be possible to test and verify them using the latest SNAPSHOT. Note that the examples that are accesible via our website has not been updated since then. So they are no help on verification.
And if you have some code to contribute, that is welcome too. We have enabled Gerrit for NatTable to make it easier to contribute.

Quote:
What do you mean by huge data sets? I see problems when having even 10k rows. It is not much I guess.

Yes I can see it too with 10k rows. Of course for NatTable this shouldn't be much. But with the current implementation the whole 10k rows will be iterated, and added again to the list to solve the update issue.

Quote:
Would you mind if I created an issue on bugzilla with further discussion and investigation?

Nope, if it is an issue for you, create a ticket. Smile

Quote:
in a not so far future, we are interested in switching renderers in our e4 based app

If NatTable is a core component of your application, I have to tell you that using the 1.x architecture it won't be possible. There are hard dependencies to SWT. With the new architecture we are removing those hard dependencies, making it more flexible, but you will need to modify your code then. But it is yet to soon to talk about details.

If you are interested I can ping you when we start working on the new architecture again, so you are able to get involved. Any help and support is appreciated by the whole team!
Re: next release [message #1231316 is a reply to message #1230842] Tue, 14 January 2014 05:33 Go to previous messageGo to next message
Dirk Fauth is currently offline Dirk Fauth
Messages: 1213
Registered: July 2012
Senior Member
For the performance on grouping have a look here: https://bugs.eclipse.org/bugs/show_bug.cgi?id=425641

The only possible way to increase performance was to change the iteration and reapplying of list entries to a clear()-addAll(). The performance gets slightly better, but IMHO it will never be running as the tree structure gets calculated dynamically on grouping.
Re: next release [message #1231716 is a reply to message #1231316] Wed, 15 January 2014 03:43 Go to previous message
Paweł Doleciński is currently offline Paweł Doleciński
Messages: 36
Registered: January 2014
Member
Quote:
Why should I? Your welcome to support in any way you like! Smile

Using this link you will find all tickets that have been resolved since 1.0.1.

https://bugs.eclipse.org/bugs/buglist.cgi?list_id=7918183&classification=Technology&query_format=advanced&bug_status=RESOLVED&product=NatTable&target_milestone=1.1.0

It should be possible to test and verify them using the latest SNAPSHOT. Note that the examples that are accesible via our website has not been updated since then. So they are no help on verification.
And if you have some code to contribute, that is welcome too. We have enabled Gerrit for NatTable to make it easier to contribute.


OK then, thank you. I am going to switch my project today to the latest snapshot.

Quote:
If NatTable is a core component of your application, I have to tell you that using the 1.x architecture it won't be possible. There are hard dependencies to SWT. With the new architecture we are removing those hard dependencies, making it more flexible, but you will need to modify your code then. But it is yet to soon to talk about details.

If you are interested I can ping you when we start working on the new architecture again, so you are able to get involved. Any help and support is appreciated by the whole team!

Do not worry, for us it too soon as well. Simply wanted to know if this topic is still valid. And of course I am very interested and would be perfect for me to be involved.

Quote:
For the performance on grouping have a look here: https://bugs.eclipse.org/bugs/show_bug.cgi?id=425641

The only possible way to increase performance was to change the iteration and reapplying of list entries to a clear()-addAll(). The performance gets slightly better, but IMHO it will never be running as the tree structure gets calculated dynamically on grouping.

I've already seen the bug, tested it on examples and I am going to check this in my project today.
I see your point and I guess there is no other way to do it, for instance do not use glazed lists at this point?
I hope, for the time being, this will enough for us. Thanks!

Cheers,
Paweł.
Previous Topic:Command for setting ComboBoxCellEditor-Values?
Next Topic:ComboBoxCellEditor List (Selection)
Goto Forum:
  


Current Time: Fri Jul 25 07:34:01 EDT 2014

Powered by FUDForum. Page generated in 0.02127 seconds