I posted my analysis of top uses of deprecated code in Drupal contributed projects in May 2019 two months ago. There are some great news since then. First of all, Matt Glaman of Centarro implemented support for deprecation messages directly in reports, so we can analyse reports much better in terms of actionability. Second, Ryan Aslett at the Drupal Association implemented running all-contrib deprecation testing on drupalci.org (Drupal’s official CI environment). While showing that data on projects themselves is in the works, I took the data to crunch some numbers again and the top deprecated API uses are largely the same as two months ago. However, we have full analysis of all deprecation messages and their fixability which gets us two powerful conclusions.
Stop using drupal_set_message() now!
If there is one thing you do for Drupal 9 compatibility, make it getting rid of using
drupal_set_message(). As the API documentation for
drupal_set_message() explains you should use the messenger service and invoke the
addMessage() method on it.
Of the total of 23750 deprecated API use instances found in 7561 projects, 29% of them were instances of
drupal_set_message(). So statistically speaking if you stop using this single function, you are already 29% on your way to Drupal 9 compatibility (likely more for most projects).
Dezső Biczó already built automated deprecation fixers covering
drupal_set_message() and more, based on rector.
72% of deprecated API use can already be resolved
On top of the 29% of
drupal_set_message(), there is another 43% of various other API uses that can be fixed now. This means that those deprecated APIs were marked before Drupal 8.6.0 was released and are therefore in currently unsupported Drupal core versions. So stopping the use of them would still keep your code compatible with all currently supported Drupal core versions. In other words, as of today, drupal.org project maintainers can resolve almost three quarters of the outstanding code changes for Drupal 9 compatibility. Considering we are still ten months from before Drupal 9’s planned release date, this is quite good news!
Upgrade Status is a nice visual tool to explicitly list all the errors in the projects you scan. It provides instructions on how to fix them and immediately fixable issues are highlighted.
Work with project maintainers
You are not a drupal.org project maintainer but want to help? Yay! Based on the above it may be tempting to run to the issue queue and submit issues for fixing all the things. Good plan! One thing to keep in mind though is to work with the maintainers of projects so your help benefits the project most effectively. Drupal.org project owners may specify Drupal 9 plan information that should help you engage with them the best way (use the right meta issue, know of their timeline plans, and so on). Check the project page of projects you are interested to be involved with to make sure you contribute the best way.
https://t.co/hf2ENvlZSo projects can now specify Drupal 9 porting information, so *you* can direct *your* contributors to provide the most valuable help on the way to Drupal 9, fund the process or just step back (for now). Edit your project to help your contributors help you! pic.twitter.com/l1OWwOllBK
— The Drop is Always Moving (@DropIsMoving) May 21, 2019
More to come
I think the data is super-interesting and I plan to do more with it, for example cross-referencing with project usage. Stay tuned for more information in the future. In the meantime all the source data is available for your mining as well.
Disclaimer: The data is based on the state of contributed projects on July 29, 2019 based on Drupal core’s 8.8.x development branch on July 29, 2019. As contributed module maintainers commit fixes, the data will get better. As core introduces more deprecations (until Drupal 8.8.0-alpha1), the data could get worse. There may also be phpstan Drupal integration bugs or missing features, such as not finding uses of deprecated global constants yet. This is a snapshot as the tools and state of code on all sides evolve, the resulting data will be different.