Rails 3 - what's our long term strategy here?

[Ed W]

Hi all. Rails 3 is looking pretty stable now. Where are we with the
idea of having a supported Rails 3 branch?

VHockstein's branch now seems to have very heavily diverged from master
and has rather a lot of interesting development in it? Is there any
reason that his branch might not become part of the official project,
and perhaps even a lot of those interesting looking commits rebased onto
master and perhaps cherry-picked into the older 2.3 branch even?

From the point of view of looking at github, vhocksteins branch *is*
activescaffold now? Is a friendly fork the only option here or can he
become part of the formal development team? Like many smaller OSS
projects we seem to be suffering from fragmentation of the few core
developers right now - please lets rescue things? Sergio in particular
is offering excellent support and I'm hugely grateful for all his
development over the last few years, however, how can we pull in new
development from other contributors, ie increase the core development team?

I personally love AS - how can we kick start things and get development
coordinated and moving again?

Regards

Ed W

[ser...@entrecables.com]

On Mon, 28 Feb 2011 17:51:25 +0000, Ed W <li...@wildgooses.com> wrote:
> Hi all. Rails 3 is looking pretty stable now. Where are we with the
> idea of having a supported Rails 3 branch?
>
> VHockstein's branch now seems to have very heavily diverged from
> master and has rather a lot of interesting development in it?

Vhochstein never told about merging their rails3 changes in official
repository, he didn't send pull requests as he did with rails2.3 fixes.
When I knew his rails3 branch I was very busy to test it. I have been
very busy in my free time last months because I was moving, and I'm very
busy in work time and I only have time to fix little bugs.

> Is there any reason that his branch might not become part of the official
> project, and perhaps even a lot of those interesting looking commits
> rebased onto master and perhaps cherry-picked into the older 2.3
> branch even?

I agree with pulling vhochstein fork into activescaffold repository and
adding him to official repository, but I would like to have some missing
changes in his fork, as helper overrides naming schema. I don't know
whether there are more missing changes.

That change was a reported issue, agreed by previous developers, and
it's useful for complex forms with subforms, where you have a field with
the same name in different models. For example, you want to override
name form field in a form for team with players, but you want to
override it only for players. It's hard without changing the helper
override name, so in master branch name is <model>_<field>_form_column

>
> From the point of view of looking at github, vhocksteins branch *is*
> activescaffold now? Is a friendly fork the only option here or can he
> become part of the formal development team? Like many smaller OSS
> projects we seem to be suffering from fragmentation of the few core
> developers right now - please lets rescue things? Sergio in particular
> is offering excellent support and I'm hugely grateful for all his
> development over the last few years, however, how can we pull in new
> development from other contributors, ie increase the core development
> team?

I would like to support rails3, and use rails3 in my new projects, but
I'm failing in that because I don't have enough time for it.

>
> I personally love AS - how can we kick start things and get
> development coordinated and moving again?

Thank you for trying to restore AS development in main repository. I
should have contacted with him some time ago viewing I didn't have time
to support rails3.

[Ed W]

On 01/03/2011 09:43, ser...@entrecables.com wrote:
>> Is there any reason that his branch might not become part of the official
>> project, and perhaps even a lot of those interesting looking commits
>> rebased onto master and perhaps cherry-picked into the older 2.3
>> branch even?
> I agree with pulling vhochstein fork into activescaffold repository and
> adding him to official repository, but I would like to have some missing
> changes in his fork, as helper overrides naming schema. I don't know
> whether there are more missing changes.
>

Hi Volker

Why don't you consider joining the "core development team" and pulling
your changes in the AS main repo?

Hope we can get some dialog going here ? Thanks to both of you for
maintaining AS - great work!!

As an aside, this project obviously still has scope for advocacy,
documentation management, web site, mailing list spam management and
more. Anyone looking to step up and assist?

All the best

Ed W

[Hernan Astudillo]

Hi Ed,

I've been using AS for years in many projects, and is very favorable to me to support this great project, so whatever area i could give a hand please let me know.

regards

--
You received this message because you are subscribed to the Google Groups "ActiveScaffold : Ruby on Rails plugin" group.
To post to this group, send email to actives...@googlegroups.com.
To unsubscribe from this group, send email to activescaffol...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/activescaffold?hl=en.

[Atastor]

Let me play the cheering crowd here: YIIIIHAAAA!!

:-)
Michael

[Ed W]

Hi Sergio

As there are so few AS "core" listening on this list now, in fact it seems only you?  Perhaps you could help project manage adding more folks to responsible positions?

I don't see that Volker replied onlist?  Please Volker, consider joining the core team and help drive this project forward?

We could still benefit from folks to maintain the website(s), help manage spam in the mailing list and obviously interested committers is also nice?

Thanks and good luck

Ed W

[Sergio Cambra .:: entreCables S.L. ::.]

On Jueves, 3 de Marzo de 2011 11:01:46 Ed W escribió:
> Hi Sergio
>
> As there are so few AS "core" listening on this list now, in fact it
> seems only you? Perhaps you could help project manage adding more folks
> to responsible positions?

There are no other developers some months (or few years) ago, nobody apart
from me commits since long time. Volker was the only person who sent several
pull requests, so he would be a great candidate to join ActiveScaffold team.

>
> I don't see that Volker replied onlist? Please Volker, consider joining
> the core team and help drive this project forward?

Volker has answered some people with problems with his fork, so he continues
reading and replying to this list. I hope he will reply sooner or later. I
will try to get time this weekend to look at rails3 fork and merge in master
anyway, but I can't promise anything.

>
> We could still benefit from folks to maintain the website(s), help
> manage spam in the mailing list and obviously interested committers is
> also nice?

I can give access to website and activescaffold issues in google code, but not
to mailing list, I'm not the owner and I don't know who is.

--
Sergio Cambra .:: entreCables S.L. ::.
Mariana Pineda 23, 50.018 Zaragoza
T) 902 021 404 F) 976 52 98 07 E) ser...@entrecables.com

[Lindsay Morris]

I heartily support this idea!  I am working on a project that will use AS, and need the config_list column picker feature.
I have not yet managed to get my stuff running on Rails 3, but am working on it.  it would be great is it were easier.

[vhochstein]

It s good to see that there are still some guys out there who are
interested in activescaffold project.

However, the past has shown that just to be interested is not enough
to keep that project up and running.
Core Team has degraded to one Person (Sergio thanks a lot for all your
work and answering all those questions),
commits to activescaffold are more a less just bugfixes
postings to this group are less than for example a year ago
Other projects like rails_admin getting more attention

To be honest I do not see a bright future for activescaffold.

You may ask why I ve decided to do a rails 3 version anyway... That s
a good question and if I had known the effort and work to get it done
it advance, I would nt have done it.
Core reason was and still is that I ve got several projects up and
running with activescaffold and I needed an option to use rails 3.

If anybody else is using my fork great, if anybody else wants to
contribute also great, if somebody wants to fork or merge, yeah sure
go ahead, that s the reason I love git and which drives a project
forward.

However, if you ask me how I see the future of activescaffold.. I
think I ve already given that answer. There are really too few
developers involved in this project, no marketing, etc.

Bottomline, if you would like to revamp core activescaffold, I wish
you all the best.
However, I do not think that I will join at this point, because I just
do not believe that it will work.

Sorry guys but just my two cents.

[Kenneth Ortmann]

Sorry I got moved off of my RubyOnRails project at work and have had to deal with Liferay/Java portlets for the past year.  I just got put back on the RubyOnRails project and might start paying more attention.  

I agree with vhochstein though.  The effort to move active scaffold to rails 3 and keep it up and stable is far greater than what it is worth.  I knew the code was going to be hard to move to rails 3 and that is why I didn't try to do it right away.  

Basically all of the inner guys of active scaffold should be re-written and if you have to do that then why not just start from scratch.  Every time something in rails changed that broke active scaffold, active scaffold was hacked to work with rails instead of being fixed the right way(or at least that's what it seemed like to me the 2 years I was working with it).

Sergio/vhochstein thanks for all the work the past couple years.

[Ed W]

Hi Volker and Kenneth

Volker, I guess I'm sad that you feel this way... The point of inviting
you to join the development team is that actually it appears that YOU
are the lead developer at present? I personally see AS as being largely
in maintenance mode, it's basically feature complete and new features
should largely become plugins. I think it merits plenty of development
for those with needs though, in particular enhancing the basic template
to support more types of plugins perhaps. However, largely speaking I
think it's looking for folks to simply maintain the status quo?

I would urge you to reconsider and join the core team? For at least the
time being that seems to mean little difference than which repo you
commit to..?

I think the state of AS is not quite as sad as you say. I think the
decline of AS input has largely been:

1) Largely dead promotional website
2) Lack of general chatter about releases in the Rails blogs, which
itself generates substantial interest
3) Mailing list became closed to defeat spammers, however, I have
noticed a massive decline in postings from users since then? Perhaps
consider other ways to defeat spam?
4) Difficulty in maintenance of the core project resources, ie no one
seems to have control over the mailing lists, commit access, etc?

I would expect that if we can fix the infrastructure (we can fork to a
new name if necessary). Arrange for "releases" to be made periodically
which include some buzz about them to create interest in the rails
community. Attract 2-4 core committers who can run development and
feeding bug fixes. Then with this I think the project is back on the rails?

With regards to Kenneth's thoughts:

> Basically all of the inner guys of active scaffold should be
> re-written and if you have to do that then why not just start from
> scratch. Every time something in rails changed that broke active
> scaffold, active scaffold was hacked to work with rails instead of
> being fixed the right way(or at least that's what it seemed like to me
> the 2 years I was working with it).

I would disagree with this. I only get time to hack on AS
intermittently, but breakages have been reasonably rare actually and
largely concern features that AS was creating that didn't natively exist
in Rails. In particular there is a view "overlay" feature which creates
a search path for a view and allows something higher up the path to
override things lower down the path. This wasn't a core rails feature
and several releases broke what was then a monkey patch on the core (not
unreasonable way to achieve a very useful feature). This should now be
resolved with new features that have come in I think rails2.3? Other
breakages have been all the normal changes that would need to be made to
any project on a rails upgrade, eg feature deprecation, function
renaming, etc. I recall the changes to 2.3 were something like changing
".RAILS_ENV" to ".rails_env" to avoid deprecation warnings and
similar..? (sorry, working from memory)

The core of AS looks scary initially, but it's actually fairly
straightforward. Largely it uses some dynamic magic to figure out the
list of columns in a given view, then iterates over those columns
generating some html for each column. At each stage it breaks out the
creation of each item into a fairly granular process that allows easy
over-riding of the created HTML. There is additionally some magic to
help support routes for nested stuff, but basically that's your whole
algorithm above...

I don't see a great re-write needed, although I would personally like to
shuffle some of the internals around for performance reasons such that
we generate many of the column lists once at boot time rather than per
view. However, that's a nice to have

Personally I think AS could have a very bright future, but it does need
interested developers to progress. I am available from time to time,
but I have many roles and I only have time to work on Rails things in
spurts (like many folks). I would dearly love to see Volker join and
add some grunt for as long as he is able, and it would be a shame if
this email thread puts him off continuing working on AS now...

Please folks, consider giving a bit of time and I think we can push
things forward here.

Regards

Ed W

[Sergio Cambra .:: entreCables S.L. ::.]

On Viernes, 4 de Marzo de 2011 11:28:28 Ed W escribió:
> are the lead developer at present? I personally see AS as being largely
> in maintenance mode, it's basically feature complete and new features
> should largely become plugins. I think it merits plenty of development
> for those with needs though, in particular enhancing the basic template
> to support more types of plugins perhaps. However, largely speaking I
> think it's looking for folks to simply maintain the status quo?

IMHO AS needs more testing and simplify code using new rails features (as
accepts_nested_attributes_for)

> 4) Difficulty in maintenance of the core project resources, ie no one
> seems to have control over the mailing lists, commit access, etc?

Only resource I don't have access is mailing list. I have access to issues in
google code, website and github (to give commit access too).

[Ed W]

On 04/03/2011 11:37, Sergio Cambra .:: entreCables S.L. ::. wrote:
> IMHO AS needs more testing and simplify code using new rails features (as
> accepts_nested_attributes_for)

From memory this is hard to make use of because AS works around the
rails attribute masking?

We could use the same syntax, but I'm not sure what it really buys us in
terms of reduced code anyway?

> Only resource I don't have access is mailing list. I have access to issues in
> google code, website and github (to give commit access too).

Do you have access to http://activescaffold.com ?

Cheers

Ed W

[Sergio Cambra .:: entreCables S.L. ::.]

On Lunes, 7 de Marzo de 2011 20:23:29 Ed W escribió:
> On 04/03/2011 11:37, Sergio Cambra .:: entreCables S.L. ::. wrote:
> > IMHO AS needs more testing and simplify code using new rails features (as
> > accepts_nested_attributes_for)
>
> From memory this is hard to make use of because AS works around the
> rails attribute masking?

We can reject columns excluded from config.action.columns or forbidden by
security methods with a before filter.

>
> We could use the same syntax, but I'm not sure what it really buys us in
> terms of reduced code anyway?

AttributeParams module, unsaved_associated and unsaved_record could be removed
probably. Views could be simplified and multilevel subforms would be easier.

>
> > Only resource I don't have access is mailing list. I have access to
> > issues in google code, website and github (to give commit access too).
>
> Do you have access to http://activescaffold.com ?

Yes, I got access finally. I put some articles some months ago, and fixed some
links to github doc.

--

Sergio Cambra .:: entreCables S.L. ::.

[nicolae.claudius]

If Volker doesn't push his fork to official we could always
"designate" his repository as "the official one" (write this in the
readme and wiki) as (currently) he is actively solving issues, helping
people and blogging about it. The main issue is all of us to combine
work and take the decisions of development together.

If he ever stops working we can fork his work back in main.

[Atastor]

My two cents, in case it cheers anybody up: I do actively use AS in
several projects, private (3) and at work (3, soon 4) (telco
provider). At work the AS projects are just now starting to get bigger
(2 active RoR developers and counting). Given the nature of the
projects, AS really helped tackling the core problems. It does this by
mostly removing the need to handle the usual controller/view part in
detail.

Call it lack of vision, call it content, but we mostly use AS as it
is, happily embracing every new feature we never missed but somebody
else thought of. You might see this as an argument pro feature-
completeness.

So far Volkers branch is great for my single private RoR3 project, but
we still shy away from migrating our RoR2 projects, due to the too
different code base (actually: I dont really now HOW different it
is...*THINK*...I just use AS as always until something goes wrong,
then I read the group or Volkers Blog or ask Volker directly (always
getting insta-answers, just as from Sergio)).

Regards
Michael

[Sergio Cambra .:: entreCables S.L. ::.]

On Miércoles, 9 de Marzo de 2011 09:48:38 Atastor escribió:
> So far Volkers branch is great for my single private RoR3 project, but
> we still shy away from migrating our RoR2 projects, due to the too
> different code base (actually: I dont really now HOW different it
> is...*THINK*...I just use AS as always until something goes wrong,
> then I read the group or Volkers Blog or ask Volker directly (always
> getting insta-answers, just as from Sergio)).

You shouldn't shy away from migrating them, I will take Volker's branch soon
and developing will be based in his branch.

[Michael Latta]

Something to keep in mind. In RoR 3.1 the options method of doing find with conditions and order, and such, is being deprecated. It will be removed in 3.2. This means all the API around customizing association options, and list conditions will need to adapt to using arel relations.

Michael

[nicolae.claudius]

@Michael
Volker already kik-started this by use of beginning_of_chain

[vhochstein]

Hi,

that s actually already one step further.

Activescaffold converts your condition_for_collection, etc into arel.

You should nt see any deprecated messages in your logs.

In case you do, please open an issue.

--
Volker


On Mar 10, 11:09 am, "nicolae.claudius" <claudius.nico...@gmail.com>
wrote:

[Michael Latta]

I was poking around after posting that and found the use of the conditions for where rather than hash inputs. There was not an override for select options like beginning of chain where I could add includes. I guess that would be a good pull request.

Michael

Sent from my iPad