WordPress core committers, core contributors, and former release leads made strong, last-minute appeals on Monday for the 5.0 release to be deferred to January. RC was expected Monday but those urging its delay cited the large number of open issues on the milestone and the fact that many confirmed bugs are being aggressively punted to followup releases.
“I do not see how we can seriously ship a release candidate today,” Joe McGill said. “In doing so, we are either saying we’re ok with shipping a major version of WordPress with this many known issues, or that the term ‘release candidate’ does not actually have meaning. I would suggest that we revise the schedule to push back RC for at least 4 weeks so we have a reasonable deadline and, in the mean time, continue releasing betas.”
Nearly every contributor involved in the discussion was enthusiastic about Gutenberg but urged release lead Matt Mullenweg to allow for four weeks of RC and code freeze to give the community to prepare.
– Building site with Gutenberg
– Find bug, go to GH to file an issue
– Find the ticket already exists.
– Bug has already been punted to "5.0.x Follow Ups"
– Find all the other *known* bugs planned for 5.0 launch— Bill Erickson (@BillErickson) November 19, 2018
Contributors said they don’t understand the rush to get 5.0. Several noted that Gutenberg seems to be measured by a different rod of success than previous releases where headline features were held to a different standard in regards to shipping known bugs.
“We’re fast approaching a million (Jetpack tracked) posts made through the editor, with the non-tracked number probably a multiple of that,” Mullenweg said in response to contributors’ concerns. “There’s been an explosion of plugins building on top of Gutenberg and some things like the work ACF and Block Lab have done that seem really transformational for WordPress. For those whom the editor is not a good fit they can opt in at any point, including post-5.0, to Classic and continue using WP exactly as they had before until at least 2022 and likely beyond.”
Mullenweg identified a few questions he sees as “good measures of success for Gutenberg:”
- Are people, when given the choice, choosing to use it over the old editor?
- Can they create things they weren’t able to create before?
- Are new-to-WP users more successful (active, happy with what they create) than pre-Gutenberg?
- Are interesting things being built on top of it?
Interesting plugins are being built on top of Gutenberg but they are breaking with every release of the plugin. Gutenberg 4.5 was released yesterday, matching the first 5.0 RC feature set. It includes a large number of changes and bug fixes that have gone relatively untested by the community at large. Most notably, 4.5 introduced a regression that caused a white screen of death when trying to load custom post types in the classic editor, forcing a 4.5.1 release earlier in the day. Every release introduces changes that cause plugins to break, requiring immediate updates from plugin developers.
Our new #gutenberg development cycle
1) New version of GB is released
2) We cross our fingers
3) Find out GB breaks stuff
4) Fix GB issues
5) Issue release
6) New version of GB released
repeat process 😓— Jamie Pootlepress – YouTuber and WP Plugin builder (@pootlepress) November 21, 2018
Gutenberg technical lead Matias Ventura posted an update today, confirming that WordPress 5.0 will miss the planned November 27 release date but did not offer a secondary date.
“The date for 5.0 release is under consideration, given it’s not plausible for it to be the on 27th,” Ventura said.
WordPress 5.0 Will Ship “When It’s Ready,” Contributors are Focusing on Getting Release Candidate out ASAP
When the second set of November dates for release were missed, many assumed WordPress 5.0 would fall back to the secondary dates in January, but that has not yet been confirmed. The previous scope and schedule Gary Pendergast outlined said the November dates could slip by up to eight days if necessary and that if additional time was required, they would aim for the January dates:
Secondary RC 1: January 8, 2019
Secondary Release: January 22, 2019
During the regularly scheduled core developers’ chat today, the discussion regarding WordPress 5.0’s release date became heated, as contributors continued to push for a January release. Pendergast suggested that December might have a viable date, to which Yoast CEO Joost de Valk responded, “I’m going to raise hell if we do December.”
WordPress plugin developers and agencies are trying to plan for upcoming holidays and want to have staff available when the release lands. Many of those who attended the meeting were hoping to receive confirmation on the release being pushed back to January.
“Please also consider the plugin shops that are rearranging their priorities to have blocks ready for 5.0, only to have had to fix them several times in the last few weeks,” Kevin Hoffman said. “The success of 5.0 depends just as much on third-party support as it does core.”
“There’s agreement on that from all sides, that the amount of code churn and missed earlier deadlines means that the 27th is untenable,” Mullenweg said. “RC is still possible soon, but please don’t assume that implies a final release date until we see how that goes and pick one. I hope that it shows that we are willing to change decisions based on new information, it’s not about being ‘right’ or sticking to previous plans blindly.”
This statement indicates Mullenweg may be considering dates that were not included in the original schedule, as he later said,”If y’all can take the data without freaking out about what it means for the release date, there have been 8 major releases in December, it’s actually been 34% of our last 23 major releases.”
Several contributors agreed that getting an RC out ASAP would finally force a longer code freeze for Gutenberg’s UI, API, documentation, and features. This would give the community more time to prepare.
“As part of the development team for almost two years now, I’d love for us to draw the RC line soon for the sake of everyone’s fatigue,” Matias Ventura said. “And think it’s ready to be drawn. I am concerned with letting us do ‘one more little thing’ and pushing the stability line further down, in an almost endless process.”
Contributors are now wrapping up the last few tickets and the plan is to get the release candidate out tomorrow during the U.S. Thanksgiving holiday. Given WordPress’ global contributor base, releasing on the holiday shouldn’t be an issue. The team is also still investigating the possibility of bundling the Classic Editor plugin with updates for existing WordPress sites.
“Our focus right now is on a great RC,” Mullenweg said. Throughout Gutenberg’s development Mullenweg has said WordPress 5.0 would ship “when it’s ready.” No release date will be announced until the team has had time to evaluate the release candidate.
“It is true that the primary thing is whether it’s ready, and it’s not currently ready,” Mullenweg said.
In 1928, John A. Shedd published a little book called “Salt from My Attic.” It included a saying that U.S. Navy Rear Admiral Grace Hopper said was influential in her life: “A ship in harbor is safe, but that is not what ships are built for.”
Shipping a major overhaul of WordPress’ editor has brought a fair share of uncertainty and frustration to contributors and the community that depends on the software. After mission-critical issues have been resolved, it seems to become a cycle of fixing and breaking things that could continue indefinitely. Although the holiday timing isn’t ideal, if Gutenberg stalls much longer it’s going to be burning daylight. At some point the ship just needs to push away from the port and see how it sails.
I followed the #core slack chat. I was glad to see that the release would be deferred pas the Nov 27 date. I must admit that I assumed, when the decision was conveyed to defer release, that the release would be pushed to January.
Yes, there was some chatter that it could come in December, but…
1) That would depart from the schedule set by the Gutenberg team itself, which had January as the backup.
2) The profound stability concerns around Gutenberg (1000s of lines of new untested code, easily discovered bugs, breaking API changes, theme/plugin compatibility with a moving target codebase, etc) can’t be resolved by bumping the release a week or two – say to early December. That kind of bump wouldn’t alleviate any of the concerns people are expressing about a risky and unstable release.
3) Meanwhile, launching later in December – a week or two before Christmas – would be an act of unfathomable insanity and meanness.
We’d be launching known buggy software that could break millions of websites, disrupt businesses during the peak holiday season, and create massive work load for tens of thousands of WordPress developers – all just before the biggest holiday season of the year. It’s a plan to destroy Christmas for everyone using or involved with WordPress.
If any of the project leads are reading this, please, I implore you not to do this. Causing this much chaos and backlash is simply not worth saving 3-4 weeks of release time.
PS. I want to emphasize I’m not a Gutenberg hater. I like it a lot, and I think it’s the future of WordPress. It’s the timing. It just defies sense to rush a buggy release to 30% of the web just before the holidays.