Codecov Report

Merging #2005 into 2.x will not change coverage.
The diff coverage is n/a.

@@            Coverage Diff            @@
##                2.x    #2005   +/-   ##
=========================================
  Coverage     93.64%   93.64%           
  Complexity     1545     1545           
=========================================
  Files            50       50           
  Lines          4030     4030           
=========================================
  Hits           3774     3774           
  Misses          256      256

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update e58da7e...0e46548. Read the comment docs.

Codecov Report

Merging #2005 into 2.x will not change coverage.
The diff coverage is n/a.

@@            Coverage Diff            @@
##                2.x    #2005   +/-   ##
=========================================
  Coverage     93.64%   93.64%           
  Complexity     1545     1545           
=========================================
  Files            50       50           
  Lines          4030     4030           
=========================================
  Hits           3774     3774           
  Misses          256      256

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update e58da7e...4aa1542. Read the comment docs.

@gchtr looks good! Just one small tweak and then we can say "hello" to a much simpler Composer-only future

@jarednova @gchtr
If I can say something here I would like to vote against this change.
I can explain it from a different view than yours.

In last few years I've used Timber in many websites using mostly plugin version as it was just quicker. I introduced Timber to a few 'non-techy' people and they fell in love with it because of simplicity. You just install plugin, download starter theme and it works just like that. You just edit the .twig and .php files.

If the only reason is a problem with informing users about 2.0 update and changes that will brake site - maybe consider creating another plugin like "Timber 2". Current one would stay as it is and people could just install the second version for new websites that would be compatible with new features and not using deprecated stuff. If it's not technically possible for some reason just ignore this idea.

I personally can handle this change, I've used composer in a few projects with Timber but for those who wanted simplicity it's gonna be another obstacle - installing composer, using terminal etc. It's another thing to learn.

Right now Timber has 30 000+ active installs from WP repository. I don't know this for sure, but I assume a lot of those installs are just websites made by people attracted by Timber's low entry-level knowledge and that it works as I said before "out of the box". It is powerful but to start you don't need much. In fact that was the reason that attracted me 5 years ago.

Of course I don't know what is going on behind the scene with the current plugin. Maybe it would consume too much time to stay up-to-date with composer version? If yes, I understand it totally and I would also support this plugin removal change. But if the reason is not work and time but just this update issue, maybe consider leaving plugin as optional install method.

@tomlecki I respect a lot of the appeal of the plugin version of Timber for fast development. But the truth is as said above: Timber is not a plugin. Shipping it as such is misleading and can cause tons of issues for site maintainers. Especially as technical debt piles up, I'm concerned about vulnerabilities. For instance, the plugin version of Timber never adopted full Twig 2.0 support.

I think there should be a sunsetting plan. Can we think about a period of time for maintenance updates? And of course we'll need to work on migration documentation.

@tomlecki @gsalkin thanks for following-up on this. I really appreciate the passion and enthusiasm. It means that what we're doing is important and reaches people (this is good!)

The plan is to continue development into the 2.x branch as Composer-only. That doesn't mean that Timber's presence in WP.org will be going away though.

We want to maintain the best possible developer experience, so we'll leave the existing plugin on WP.org for the indefinite future. If we find any critical bugs, we'll update them too. But what we can't do is maintain two versions into the distant future. Syncing the versions is an error-prone process and I'm very wary of (and have made mistakes with, despite scripts and other tools to automate the process).

All that said, these are plans and things could change (especially if someone volunteers for a role reaching out to new devs just getting into WordPress and templating!??!??!?!) — but for now we're planning on making 2.0 Composer-only once it's released.

... and of course as @gsalkin notes, we'll be sure to document a migration path and on-boarding materials. I hope @tomlecki you can keep us honest on this front to ensure the great experience you have (and have introduced others too) can continue!

Some of the points raised by @tomlecki are pretty good and I agree and I have a bit of mixed reviews about this as for now, it will definitely take away the usability for not-so-tech users of timber but considering the long term impacts and how it should be treated, I appreciate the decision. It absolutely puts it out as a more mature solution for developers and clears the confusion about where it stands as a plugin/tool/framework.

Here is an interesting thing, a few weeks back I installed Twig using Composer directly into wordpress theme which I was developing because timber as plugin didn't feel like the right choice, mostly because it does feel like part of a theme more than a plugin and it worked well (minus all the features that timber has but twig can not for obvious reasons). For that particular instance Timber would have been a better choice if it was available with Composer, although for most of developments I install Timber as a plugin.

I've been using Timber for about 3 years and initally I hated 😡 it but as I came to know more, I now prefer to make a whole theme on top of it because of the clean separation it provides which gives ability to a bigger team (designing & development) to work on things simultaneously. ♥😍

Composer does have a bit of learning curve but all serious developers have to learn it at some point anyway, because any complex development will also have additional libraries such as emogrifier, emailer libraries, data validation libraries, pdf generation and such which have to be installed with Composer.

I don't know all ins and outs of Timber yet, but I think this steps makes it adapt the newer twig versions easier than before. (or may be I'm incorrect and someone should correct me 😛)

Keep up the good work 👍👍