[BlueOnyx:25429] Re: RedHat 9 Beta released - BlueOnyx 5211R development status

[Michael Stauber]

Hi Ernie,

> FYI,
> Red Hat Enterprise Linux 9.0 is now officially available to Red Hat customers as
> stable.
> 
> Not sure how long the downstream will take to go stable.
Yeah. AlmaLinux published a Beta 9 not too long ago, so they should have 
their stable release out soonish as well.

I don't know what the fuck RedHat is doing or thinking by declaring EL9 
ready at the state it currently is in. Instead I wonder what their 
target audience is supposed to be. /shrug

I'm now five to six months into the 5211R development and I am still 
chasing dependencies. This isn't my first rodeo, but it has *never* been 
this bad. Back in December EPEL 9 has been launched and it started to 
get populated with EL9 RPMs and SRPMs. Which is a godsend.

Still: The amount of barebone basic libraries that are still completely 
absent in EL9 is bordering on ridiculous. On the other hand they have 
gone ridiculously overboard with dependencies in Fedora to a point where 
they're on the absolute other side of the spectrum from where EL9 is. 
That makes porting stuff from any recent Fedora to EL9 a nightmarish 
undertaking.

On that end I'm making some progress to fill the gaps, but only with 
great effort and at a snails pace.

While I was running into multiple dead ends on layered dependencies for 
some of our stuff ("Mailman 3" being one of them) I just said "Fuck it!" 
and set these open issues aside and continued with the parts that I 
could get working. The deeper I got into adapting the GUI to 5211R the 
more I realized that I eventually would have no choice but to tackle 
upgrading CodeIgniter to the latest version.

If I don't do it before the release of 5211R, then we're stuck with it 
until the EOL. Which would be unwise. But upgrading CodeIgniter is (due 
to many, many changes) like burning the house down once and starting 
over. It should be worth it due to many fundamental improvements, but 
it's a *really* time consuming process. Especially when this is on a new 
OS, where you don't know if something isn't working because of the OS 
side of things being different than expected, or CodeIgniter changes 
throwing a wrench into how you thought it should be done.

Means: I'm overhauling or redoing every bloody library, UIFC component 
and GUI page. And there are plenty of them, which means I can't even 
make a prediction of how long it'll take, as I'm currently chasing my 
own tail in stuff that is broken left and right.

I've also tentatively started to overhaul the theme of the GUI as well 
to give it a more modern appearance, although I'm not yet happy with the 
mock-up of it that I currently have. Still: Some parts of the old GUI 
are hard-coded into the old GUI and while I redo those parts, I make 
provisions to externalize the template stuff properly.

The ongoing priorities are currently these:

- Porting the GUI to CodeIgniter 4
- Deprecation and removal of legacy libraries and functions
- Finish Handler/Constructor overhaul to adapt to EL9 specific quirks
- New GUI template provisions for easier template changes.

While that is underway I'm also frequently diving into the dependency 
pool, trying to scoop up, rebuild or build-from-scratch whatever 
dependencies we're still missing to regain full functionality.

With that I mean: Feature wise 5211R must at least be on par with 
5209R/5210R or better.

That being said: As there is *nothing* in EL9 (as far as the OS goes) 
that is spectacularly better or drastically more modern than EL8. Nor 
does EL9 have any killer feature that EL8 doesn't have. So even if we 
had a 5211R now that was feature equal to 5210R: There wouldn't be much 
of a reason to upgrade or migrate to it. Except EL 9 being a thad newer 
and "Oh, Shiny!" But that "Shiny!" is at best Pyrite ("fools gold").

Bottom line: I don't see the need to rush the release of 5211R just for 
the sake of having it out as soon as possible. That would do us a 
disservice. It'll be out when I'm happy with it and then it ought to 
provide enough incentives to eventually make the switch from 5210R.

-- 
With best regards

Michael Stauber