My experience over there is that the ego-driven staff there do not appreciate it when we can solve problems which they cannot solve.
In the past, I was banned for simply pointing out some basic technical errors, as I recall. But it was long ago, and I "let go" of any feelings toward them. After all, their ego issues are not my "cross to bear".
In other words, I don't help out there any more because "No good deed goes unpunished." (or banned) from simply speaking the truth.
If you look at the root cause of this issue, it's a mistake by their co-founder, who released a new version of the rbtrace gem:
as seen here on github:
The above rtrace gem release happened the day after a new discourse release which has references and artifacts to an earlier version of rbtrace (v 0.5.0)
So, in a nutshell, discourse publicly releases a new version of discourse which pulls a discourse docker base image with rbtrace v0.5.0 artifacts and the very next day they released the gem rbtrace v0.5.1.
However, the Gemfile of the public release did not specify the required version; and so the final build pulls rbtrace v0.5.1; and this conflicts with the required discourse base docker image with rbtrace v0.5.0, whick breakes the build (final bundle) process.
Those "expert support" guys over there at meta will never admit this error by their co-founder and will never say "thank you" or any sincere appreciation; and this kind of detailed root cause analysis on my part got me banned in the past.
Their core goal has always been centered around money - to convince people they are the "smartest people in the room" and that everyone should feel blessed to pay for their (expensive) discourse hosting service.
Therefore, I think it's OK to reply to any direct or private messages over there; but you will not see me posting any solutions over there.
I mean, who releases a new version of rbtrace like that without testing against the currently specified discourse docker base image, released just the day before?
Not many admins are old enough and advanced enough to track this root cause problem down; and especially when the only replies from the "expert" support teams are totally off target (OS and platform questions) and irrelevant to the root cause (bundle)?
And not to mention, this core problem was caused by a discourse co-founder who has a huge ego and banned me at least twice years ago for posting the truth (the facts, you know, the correct technical details)!
The current "solution / advice" over there tells people to build a new app from scratch; but that will not work (at this time) until discourse releases a new version with a newer version of the discourse base docker image which has the current version of rbtrace. I proved that early on by trying to build from scratch and it failed for the same reason.
AI is not the thing we should be afraid of - it's the people who create AI and the people who use AI we should fear.
Happy New Year
