By ncameron on

Hi there,

I was after some input on a small dilemma. For the past 6 weeks I have been testing and developing a prototype website on my local machine. In the next few days I will be ready to start working on my production site. My dilemma is this:

How should I deal with the upcoming release of 4.8/5.0? From what I hear (lullabot podcast 19) there will be quite a lot of differences in the new version. Would it better to go ahead and develop on 4.7 and then in a few months transfer over to 4.8/5.0 or would it better to wait for the next version and develop from scratch on that? My thinking is that if I wait for 4.8/5.0 many of the modules I am dependent on may not be available anyway. Another possible might be to use a CVS version is this worth thinking about?

Thanks for your thoughts on this, I'm sure I'm not the only one facing this problem.

Regards,

Neil

⋅ Categories: Drupal 4.7.x

Comments

webchick’s picture

webchick
she/they
Primary language English
Location Vancouver 🇨🇦
commented

If you're a developer, aren't planning to use many contrib modules (or are comfortable updating those that you want to use yourself), are okay with developing against moving targets (APIs shifting, functions being renamed, etc.), your site is likely to require a bunch of custom coding, and you're not planning to launch until the new year, then developing against HEAD may be right for you.

For most people though, you should probably stick with 4.7 and update later.

The reasons why developing against HEAD are not a good idea:

  • Currently, development is active in the HEAD branch, so huge changes are still possible and core and especially contributed modules may break. Often.
  • Even after the "code freeze" on September 1, API changes can still be made, although they tend to be fewer and further between, and focused mostly on peformance/bug fixing/usability/etc. 4.7 had changes right up until the 4th Release Candidate, AFAIR.
  • Because of this, many contrib module developers wil hold off updating their modules until 4.8/5.0 reaches "release candidate" status, so they don't dump a bunch of hours chasing a moving target.

Developing against HEAD now means a lot of "DIY" stuff... fix your own bugs, update contrib modules yourself, etc. That said, the earlier you get involved in this process, the earlier you can ensure that the next version of Drupal will do what you need it to do. ;)

webchick’s picture

webchick
she/they
Primary language English
Location Vancouver 🇨🇦
commented

You may realize this already, but just in case...

HEAD == CVS version == 4.8/5.0

Those terms all mean the same thing right now.

A "code freeze" happens on Sept. 1. That means that after that point, no new features are added, and the focus then becomes bug fixing, performance improvements, usability improvements, testing the upgrade path, etc.

After the "code freeze" there will be a few weeks of this process and once HEAD becomes relatively "stable," then a "beta" of either 4.8 or 5.0 (depending on what Dries chooses to call it, which is dependent on how many changes make it in before the code freeze date). This is when the larger Drupal developer community will generally start to play with the next version, report additional bugs, etc. These bugs get fixed and new "beta" versions are released as needed.

Once there are no more "critical" bugs, then a 4.8/5.0 "release candidate" is released. Here's where the bulk of the Drupal community will try out the new release. More bugs are probably found and fixed, there may be another release candidate or two, until finally it's deemed "stable" and...

Drupal 4.8/5.0 is released! This means the code is deemed good enough for use on production sites, and HEAD opens back up for new features up until the next "code freeze" date. Rinse, repeat. ;)

And before someone asks, no there is no known release date for 4.8/5.0. :) It's all totally dependent on how much help the developers get testing and bug fixing. This is why even if you're developing on 4.7, it would be great to have a HEAD environment setup as well to help track down (and fix, if you're feeling plucky!) bugs so the next version gets released sooner.

Robardi56’s picture

Robardi56 commented

Hi,
I had the exact same dilemma than you and took my decision yesterday. (btw, i'm NOT a programmer).

I decided to go with 4.7.x latest release and latest compatible stable modules.

I will wait until 4.8/5.0 is fully released before upgrading. This is months away. You have all the time to setup your new site, and wait that all the modules you are interested are ported to 4.8/5.0. Only then will you upgrade your site, with the best chance for the upgrade to be without problem.

My best advice if you are not a programmer: never be in a hurry to upgrade. Always wait for the stable releases.

Brakkar

chiggsy’s picture

chiggsy commented

I jumped the gun with 4.7 on a project and got a face full of pee.. was too much of a moving target, and the modules that could have saved me were not there. This time i'll wait.. i'm in no rush to be a beta tester on a clients expense again. My fault totally, but a most unpleasant experience

ncameron’s picture

ncameron commented

I am certainly not a programmer and like to keep my hands clean as much as I can, although I am partial to a little quick and dirty hacking. It seems far better to stay with 4.7.x for now and I will upgrade later on when 4.8/5.0 is stable and bug free.

Cheers guys!

--
Neil Cameron

v1nce’s picture

v1nce commented

Good idea... especially if developing on a live site!
-------------------------------------
CommunityTraveler