Backport CiviCRM 2.1 to Drupal 5 ?

[Matt2000]

This is just a preliminary inquiry to find out if there is sufficient demand for a Drupal 5 compatible branch of CiviCRM 2.1.

I'm pretty sure that the core developers have no interest in doing this (and I'm not sure I even want them spending their valuable time on it.)

But I wonder if there is enough desire & skill in the community to make this happen? I know I have several clients who would love an upgrade to civiCRM 2.1, but can't leave behind Drupal 5 yet due to other module requirements. (Personally, I haven't decided yet if it might be easier to just upgrade those other modules to Drupal 6, but that's not the point of this thread.)

So, are people willing to step up and support a new branch with contributions of:

- Coding work
- Documentation
- Testing
- Cash to entice developers

If a team can be formed, I can provide a server to host a code repository, and will contribute what code I can, but I'm not sure I'm up to the task by myself.

If this takes off, advice is also welcome on how best to make sure future bug fixes & enhancements can be ported to the D5 branch with the least possible difficulty.

Please speak up if this would be valuable to you.

All the Best,

Matt

[Donald Lobo]

thanx for taking the initiative matt

from a civicrm core team perspective, we'd be happy to support the initiative and would be willing to host the branch on the civicrm svn servers etc and do periodic merges from the 2.1 branch (to keep it up2date with the latest bug fixes etc). Note that while coding will be a fair part of the project, the larger component will be support (IMO), so please do budget time/money/resources for that. We'd be happy to create a top level topic for D5 - Civi 2.1 support

from a code perspective, we've abstracted things away from drupal even more (which i suspect matt will not be too happy about), but this does make the port significantly easier. Some drupal hooks have changed in D6 so you'll will need to roll back to the code in the 2.0.x version of the module. If i had to guess, doing the backport is a 50-100 hour project (very conservative estimate). Support however would be 500 hours+ in the long term

lobo

[excell@drupal.org]

Oh yes - we are most definately interested in this...

Establishing our first large Drupal website on 5x for a non profit assoc.

With the modules we are using we cannot see upgrading to 6 for some time but would dearly love to upgrade CiviCRM to the latest features.

As far as what we can contribute? Not much! We are newbies here, (with little resources) so about the best we could do would be to be involved in some testing and reporting etc.

Note: we have actually held back from trying to set up much with CiviCRM so far because the new version looks so much more attractive...so we don't have much experience in either Drupal or CiviCRM.

I know I am not much help - but thought I would pipe up to express our interest, thanks!

[Matt2000]

OK, follow-up question for interested parties:

Would you be interested in a paid-support plan for Drupal5+CiviCRM2.1 ? If so, what structure would best suit your organization: pay per ticket, pay per hour, pay per week/month/year?

Through my various businesses and partnerships, myself  and a couple close associates currently support nearly 40 Drupal 4 or 5 based CiviCRM installations. If there were sufficient demand, I think I could systematize our offerings sufficiently to offer services on a wider scale, but I need to know if it would be sustainable over the long term, as my available capital would not make payroll for very long without a steady stream of needy clients...

[Matt2000]

Add to the list of potential contributors:

-Grant writers

I've worked with NPO's long enough, that I should know more about the subject than I do, but I'm really at a loss.

If this project were to gain such interest as to take it seriously enough to seek grants, the I would propose the following on-going objectives:

- Provide & support CiviCRM branches compatible with all currently supported versions of Drupal. (Hence, when Drupal 5 is obsoleted, and CiviCRM moves to Drupal 7, we would take over support of the Drupal 6 branch. Long term support is key for adoption of open-source projects by large organizations, cf. Ubuntu)
- Improve integration between CiviCRM & Drupal by developing add-on modules and enhancements (e.g., Views2 integration, FormsAPI-based, themeable templates?)
- Update existing Drupal integration modules to keep pace with CiviCRM as needed (e.g., CiviNode)
- Liason with Drupal module developers to avoid conflicts with CiviCRM (e.g., namespace issues). Test new CiviCRM releases running together with the most popular Drupal modules.

[geilhufe]

The grant route is a tough one... however, if among those 40 organizations you are supporting for CiviCRM installs, a bunch share the same institutional funder (a foundation, corporation, really rich person, etc.), then there might be a good angle.

You get the grantees to go to the funder as a unified group and ask for a capacity building grant. Something where the grant pays the up front costs and then the organizations chip in to pay for monthly support to finance the long term. This can work pretty well.

That grant then pays for your bullet points.

The funder (and probably even your clients) will not care/ understand your bullet points.... instead it should be things like modernize technology, increase individual fundraising capacity through technology, etc.... the CiviCRM users will know how to frame the story for their funder.

[Chris Burgess]

I'd definitely be interested in participating, although I think we'll move most sites quickly to D6 anyway.

[mfb]

I was able to easily get CiviCRM 2.1 working on Drupal 5 by just downgrading civicrm_menu() back to Drupal-5 style menu hook.  The CiviCRM 2.0 to 2.1 upgrade seemed to run ok on Drupal 5.  I haven't yet tested heavily or dug in to what else needs to be done.

[Kurund Jalmi]

Bunch of things that ( I remember ) changed in CiviCRM v2.1

1.  url() api changes. In civicrm.module.
2. i think there were some changes to drupal execute function, (can't remember clearly), during CMS user creation.

HTH

Kurund

[Donald Lobo]

helps to abstract more things away from the cms

some other changes include: changes to user hook and block hook in v2.1 for D6

lobo

[mfb]

Thanks for the tips.  Maybe I'm stating the obvious but I think the CiviCRM team could massively accelerate 2.1 adoption by offering official D5 compatibility

[Donald Lobo]

maybe we are not stating the stuff clearly enough: but we do think that its good for the community to group together and support each other by taking charge of the drupal5 backport (along with the related support issues)

lobo

[Matt2000]

@mfb,

Thanks for getting this started! I was just about to start looking into myself when I realized that coder module doesn't have back-porting reviews. :-(

Should we go ahead and have the Dev team make a branch where you can commit you initial adaptations? That would make it easy enough for others the checkout the code and test further.

My question is if it's possible for the Dev team to continue to commit their fixes to beta to our branch also? Or should we wait until 2.1 is stable?

Once the backport is complete, can the core team see to the rolling of tar-balls for sourceforge?

Also, it would be nice if JIRA could be configured to designate issues which are particular to the D5 branch, so we can track them separately.

Finally, though I must admit I am very weak in the area of unit testing and such, I wonder if the Core team has any testing tools which would be useful to us? Directions to any documentation which would help in the application of these tests would be appreciated.

Between those of us who've spoken up here already, I think we could make a it a realistic goal to have a D5 compatible version available shortly after the official release of the D6 version. I'm willing to monitor the issues queue if mfb, xurizaemon, and excell are willing to assist me with fixes and support.

[rkdesantos]

Hi Matt,

As you might recall, we've e-mailed about my site installing CiviCRM at some stage.  We're still on Drupal 5.10 and we'll be on 5.x for at least another few months, perhaps up to a year.  The reality is that we use a number of modules that are not yet available in Drupal 6 versions and until they are, or equivalent solutions can be found for D6, we won't move forward.  Many, many module developers for Drupal have not moved forward with D6 versions, let alone D7.  An inventory of the modules we use showed that as of this morning, about 35% are D6 ready, 40% are in development, and 25% haven't even started.  So, no matter how much the core developers want to move on, the users and module developers move at their own pace. 

It would therefore be a great advantage to us to have a backport of CiviCRM 2.1 to be D5 compatible as it would facilitate our use of CiviCRM and our move forward. 

[jakedeo]

Here here for the idea! Not sure if I could help coding much but would be willing to do any nominal tasks involved to help as I can.