flexinode towards 4.7: update, and it needs additional maintainer(s)
Hello, After some busy weeks, I found some more time to devote on flexinode. I beleive flexinode is critical for 4.7. Also for the Drupal 4.7 release. It is a popular module, and rather a complex one. Without a 4.7 release we will have a big lot less testers and people who want to move to 4.7. Samc recently hopped in and is testing and making patches like a madman, thanks for that! I am doing at least five to seven patch tests and/or commits a day too, but we are still far from a release, that is by far not enough to ensure a safe upgrade to 4.7 for flexinode users soon. So: * No. I will not drop flexinode in favour of CCK. CCK is not tested in critical environments, there is no (tested, full-proof) upgrade path, etc. It is just not ready for prime time yet (IMO). I will maintain flexinode for at least one release cycle, depending on the speed of development in CCK. * I would love to have someone maintain the 4.6 branch of flexinode. I you have several or one sites running 4.6 flexinode, and some time left, mind doing this? I do not actively maintain any myself, and I would like to focus entirely on getting a solid 4.7 flexinode out. Just Do It, its nothing official, but please announce it somewhere so we know who to turn to. * I would love help in that road to 4.7. Both in testing patches, and in committing them. Anyone in for some flexinode joy? Same as above, its nothing official, but you'd need at least commit rights to CVS, and some experience with patches and flexinode (otherwise it will cost us only more time). Cheerios, and flexible greetings, Bèr
On 21 Apr 2006, at 12:35 PM, Bèr Kessels wrote:
* No. I will not drop flexinode in favour of CCK. CCK is not tested in critical environments, there is no (tested, full-proof) upgrade path, etc. It is just not ready for prime time yet (IMO). I will maintain flexinode for at least one release cycle, depending on the speed of development in CCK. Could you provide us a detailed list of what's missing, and what's not possible.
I remember you mentioning multi-selects, but there seems to be a recent commit that jonbob made that added this. http://cvs.drupal.org/viewcvs/drupal/contributions/modules/cck/ optionwidgets.module?rev=1.7&view=log -- Adrian Rossouw Drupal developer and Bryght Guy http://drupal.org | http://bryght.com
On 21 Apr 2006, at 12:35 PM, Bèr Kessels wrote:
CCK is not tested in critical environments, Umm. the difference is that flexinode is _known_ to not work in critical environments.
-- Adrian Rossouw Drupal developer and Bryght Guy http://drupal.org | http://bryght.com
CCK is not tested in critical environments, Umm. the difference is that flexinode is _known_ to not work in critical environments.
An attempt to cull a potential flame. Flexinode doesn't scale well. Due to it's exprimental design it has loads of performance issues, so I wouldn't recomend it for critical environments [1]. It is a workable solution for sites with modest traffic. It does offer functionality not yet available in cck, mainly, I suppose, due to time constraints, or natural unwillingness to commit to a particular implementation: what approach to use for file fields? (to mention one from the top of my head) By design, CCK scales better, provides better fundamentals, is a better long-term choice. Practicaly CCK is flexinode 2.0, so like all decent 2.0 releases it is still in beta. Having said that, I agree that 1.0 & 2.0 are going to coexist for some time. I suppose the fittest will survive. Both have valid applications. One is young(ish), the other one is aged. Please don't argue on the merrits of either - they are obvious, as the disadvantages of both. ------------------------------------------------- [1] critical environment depends on your definition of one, but for me the requerements for the software are circling around: * security * scalability - both how many websites can co-exist and how flexible it is. * stability under load ( intentionally ripped off from above) * performance - related to the above Flexinode has issues with most of the above. I'm using it on minor card sized websites, or applications - because iit is flexible. Nothing challenging.
please folks. Do not start this useless debate. CCK is NOT ready. There are NO working upgrade paths and people DO use flexinode a lot. I beleive CCK is the way to go, for sure. But if we drop flexinode now, we have NOTHING for a long time. meaning that we force thousands of sites to remain on 4.6. Hence, i will maintain flexinode, if only for my own sites. And I already run one CCK site, but so far, am rather dissapointed by its state. Its *concepts* are *far* better then flexinode, but still very new, far too new to be a proper replacement for flexinode. I do not want to make flexinode a competitor for CCK, not at all. Just want to provide a properly working solution for 4.7. Bèr
On Apr 21, 2006, at 8:11 AM, Bèr Kessels wrote:
CCK is NOT ready. There are NO working upgrade paths and people DO use flexinode a lot. I beleive CCK is the way to go, for sure. But if we drop flexinode now, we have NOTHING for a long time. meaning that we force thousands of sites to remain on 4.6.
Hence, i will maintain flexinode, if only for my own sites.
Here's the official stance from CCK's mainainer. :-) - CCK will be ready for 4.7's release. For some values of "ready." That is, I have been launching production sites using it for some time, and I feel that it's stable. - It will probably be a couple months yet before all of Flexinode's features are implemented in CCK, so people who need those should stay with Flexinode. - An upgrade path from Flexinode to CCK is critical. Owen is working on this. - I am very very very happy that Bèr is stepping up and maintaining Flexinode as I throw my efforts elsewhere. The two can coexist very nicely during the 4.7 cycle. -- Jonathan Chaffer Algorithm Alchemist, Structure Interactive (616) 364-7423 http://www.structureinteractive.com/
Holy cow. Sanity. On 4/21/06, Jonathan Chaffer <jchaffer@structureinteractive.com> wrote:
On Apr 21, 2006, at 8:11 AM, Bèr Kessels wrote:
CCK is NOT ready. There are NO working upgrade paths and people DO use flexinode a lot. I beleive CCK is the way to go, for sure. But if we drop flexinode now, we have NOTHING for a long time. meaning that we force thousands of sites to remain on 4.6.
Hence, i will maintain flexinode, if only for my own sites.
Here's the official stance from CCK's mainainer. :-)
- CCK will be ready for 4.7's release. For some values of "ready." That is, I have been launching production sites using it for some time, and I feel that it's stable. - It will probably be a couple months yet before all of Flexinode's features are implemented in CCK, so people who need those should stay with Flexinode. - An upgrade path from Flexinode to CCK is critical. Owen is working on this. - I am very very very happy that Bèr is stepping up and maintaining Flexinode as I throw my efforts elsewhere. The two can coexist very nicely during the 4.7 cycle.
-- Jonathan Chaffer Algorithm Alchemist, Structure Interactive (616) 364-7423 http://www.structureinteractive.com/
On 21 Apr 2006, at 13:09, Vladimir Zlatanov wrote:
[1] critical environment depends on your definition of one, but for me the requerements for the software are circling around: * security * scalability - both how many websites can co-exist and how flexible it is. * stability under load ( intentionally ripped off from above) * performance - related to the above
Flexinode has issues with most of the above. I'm using it on minor card sized websites, or applications - because iit is flexible. Nothing challenging.
I agree, but the complexity of the CCK scares me a bit. Simplicity is an important aspect too; it makes the code accessible and lowers the barrier to entry. -- Dries Buytaert :: http://www.buytaert.net/
I agree, but the complexity of the CCK scares me a bit. Simplicity is an important aspect too; it makes the code accessible and lowers the barrier to entry. To be honest the complexity of flexinode is what really scares me, but you are right CCK is a thing to get your head around, at least initially. But I find some real beauty and elegance in it's design.
Huge thanks to JonBob, John, Earl, Robert, Steven and the rest of the CCK crowd for working this.
Hello Op vrijdag 21 april 2006 15:23, schreef Dries Buytaert:
I agree, but the complexity of the CCK scares me a bit. Simplicity is an important aspect too; it makes the code accessible and lowers the barrier to entry.
IMO this is one of the things that has improved a lot between flexinode and CCK. However, making a new field for fexinode requires less then 1/3rd of the code needed for a field in CCK. I think we should really see that this gets "solved". A single field for CCK needs more code then a simple node type containing such a field. This can be bad, because developing a single field is often harder then doing the same for flexinode. A single field for CCK requires you to build a whole new module. For most of us this is easier then some arbitrary concept like the flexinode .inc. So this can be an improvement. But for this to be usable, we REALLY need some dependency and an overhaul of the module-managing interface. A module containing a lot of fields is much smaller then the same amount of fields in flexinode. This is good thing too. I think that the complexity of CCK is mostly there, because it was not documented fully. Also because its concepts (widgets, fields etc) need finetuning and explanation. All in all, I have a lot of confidence in CCK. Ber
On 22 Apr 2006, at 9:55 PM, Bèr Kessels wrote:
A module containing a lot of fields is much smaller then the same amount of fields in flexinode. This is good thing too.
No reason they can't be they are just being developed separately for now. I would like to see all the default fields (whatever they may be) collapsed into a single module (preferably content.module) -- Adrian Rossouw Drupal developer and Bryght Guy http://drupal.org | http://bryght.com
On Apr 22, 2006, at 3:55 PM, Bèr Kessels wrote:
However, making a new field for fexinode requires less then 1/3rd of the code needed for a field in CCK.
I'm pretty sure this is not the case anymore. The defaults in content.module handle loads, inserts, updates and deletes in nearly all cases now, so a field really just has to declare its DB columns, how to present the edit form to the user (i.e. a widget), any validation necessary, and display. That's about half the hooks that were previously necessary. -- Jonathan Chaffer Algorithm Alchemist, Structure Interactive (616) 364-7423 http://www.structureinteractive.com/
Hi, On 4/21/06, Bèr Kessels <ber@webschuur.com> wrote:
Hello,
* I would love to have someone maintain the 4.6 branch of flexinode. I you have several or one sites running 4.6 flexinode, and some time left, mind doing this? I do not actively maintain any myself, and I would like to focus entirely on getting a solid 4.7 flexinode out. Just Do It, its nothing official, but please announce it somewhere so we know who to turn to.
You probably saw some of my patchs and ideas in the bugtracker. Those mean that I use and develop things in flexinode for my clients. I'm glad to make some more patchs (I know I have to refactor some of them, that were sent) to fix and to add some features. But to do this, I must have some "garantie" that the work I would do in flexinode 4.6 would have a port to 4.7; so I can then upgrade and not loose work. What do you think? Regards, Fernando Silva
Op vrijdag 21 april 2006 14:14, schreef Fernando Silva:
But to do this, I must have some "garantie" that the work I would do in flexinode 4.6 would have a port to 4.7; so I can then upgrade and not loose work.
Unfortunately non of us can guarantee this. In fact, I think that we can safely assume that work done for 4.6 flexinode will not be ported to 4.7 without lots of additional effort. On top of that, my #1 priority now, is to get 4.7 stable. After that, we can see what to do with new features for 4.6 and 4.7. and with non-critical bugfixes for 4.6. Unless we have a dedicated 4.6 flexinode maintainer, I also think it is good idea to state that we do not allow ANYTHING but critical bugfixes into the 4.6 flexinode branch. Bèr
On Apr 21, 2006, at 3:35 AM, Bèr Kessels wrote:
I beleive flexinode is critical for 4.7. Also for the Drupal 4.7 release. It is a popular module, and rather a complex one. Without a 4.7 release we will have a big lot less testers and people who want to move to 4.7.
Flexinode has consistently been the most searched term on Drupal.org for anonymous users. Interestingly enough, CCK is now the most commonly searched term for authenticated users with Flexinode in second with half the searches so far this week. Cheers, Kieran
participants (8)
-
Adrian Rossouw -
Bèr Kessels -
Dries Buytaert -
Earl Dunovant -
Fernando Silva -
Jonathan Chaffer -
Kieran Lal -
Vladimir Zlatanov