From great to disaster

If you’re interested, please see my replies to other threads on this topic, particularly this one: https://www.ads-software.com/support/topic/beware-crashes-your-site/ .

200,000 sites did not go down. Did you explain to your customers that their live sites went down because you configured them with automatic and unattended plugin updates? That you’re performing zero testing as a result? That a faulty update to a plugin went straight to their live site that should have been caught on a pre-production environment? That this can happen again tomorrow either through human error, technical malfunction, or malice? Obviously I am sorry that a faulty version of WP Crontrol got deployed but you are the only person responsible for your customers’ live sites going down as a result.

I happily run fully automatic updates on my personal websites but I would never do it on a customer’s live website because it’s a risk.

there’s nothing that says this developer won’t do this again

On a more positive note, ironically the opposite is more likely. Since this faulty deployment I’ve put further safety nets in place to reduce the chance of the process breaking in the future, maybe moreso than many other plugins, and I’ve proposed subsequent improvements to the deployment tool that’s used by this plugin and many others in the WordPress ecosystem.

Moderator note: @webbmaffian You can edit your review if you wish.

@sterndata – Thanks, I’ll write a reply instead.

@johnbillion – The review was unfair, my apologies. I have nothing that confirms the numbers (roughly taken from the number of installations) nor the assumption about the future. I respect and appreciate the time and work you put into open source software.

However, it’s weird to me that you try to blame the users of your plugin, and in a rather condescending and disrespectful way. We did indeed believe that this plugin wouldn’t make a mess, and we are indeed responsible for our customers, but the cause of the incident is yours and yours alone. You can’t negate that.

I do believe that you will learn from this, but at least we are moving on.

As for automatic updates in general, I agree that there is room for improvement and appreciate your propositions.

Thank you.

I’m sorry if my reply came across as condescending or disrespectful, that was not my intention. However it’s important that a distinction is maintained between the decisions and processes of a plugin author and the decisions and responsibilities of its consumers, otherwise it’s all too easy to head down the route where authors of open source libraries are blamed for effects (“my live business website went down”) versus what they actually caused (a PHP fatal error in this case).

I would much preferred to have received a whole load of reviews and support threads from people telling me that their development environments were showing fatal errors.

Maybe the whole automatic update system in WordPress encourages too much risky behaviour when it comes to live websites without sufficient protection mechanisms in place. There are initiatives ongoing to increase the reliability of the automatic update system itself (in case of a fault during installation of an update), but as far as I know nothing to address problems that can arise after an update, whether that’s a high impact error like a fatal or any other bug really.