Recent searches
Search options
@carlton@fosstodon.orgI don’t think we could get rid of the Django LTS now — TOO MANY FOLKS use it — but it’s a pain, and I think it encourages bad habits…
> My Car is an LTS, so I don’t bother servicing it. 
If we didn’t have the LTS already, no way would I support introducing it now.
Times just aren’t what they were years ago: there’s no stability issue — JUST UPGRADE! 
From: @tonybaloney
https://fosstodon.org/@tonybaloney/113361683233996580 #Django
@carlton LTS lets us slowly upgrade, without relying on package developers upgrading immediately too. Sure, the supported versions overlap, but for large projects with other priorities, it's not always enough.
With that said, I've had so many conversations about people using LTSs as "Why do 1 upgrade a year when we can do 1 every 3 years", and how it really doesn't work quite like that.
@jake Yes, I understand the arguments... — I GET that people use the LTS, and are used to doing so.
@carlton 100% agree, you should be good at doing small upgrades then it's never a big thing that happens rarely and you don't have a lot of experience with.
@carlton @tonybaloney We should switch to a modified CalVer while we are at it. Which version are you using? Django 26 which would mark the major release in that year. Give even years two years of updates or something clever or less clever. (1/2)
I always thought it was weird to not mark the major.0 as the LTS version versus the major.2 release. That gives two releases to bake something new in and stabilize it with a shorter overall maintainance window before its time to hop to the next LTS. (2/2)
@webology But no-one updates until the .2 I thought?
(Just shitposting. Don’t think I disagree.)
@carlton I think our book author friends get hit the hardest by the LTS to next major version change falling in the same year or months later.
@webology yes, that’s an issue. Folks will still be demanding v28 coverage 2mins after v27 is released though, so not sure that’s the key point.
Switching to Covers 5.x seems a good move.
@carlton This is why I default to CalVer(ish) so there is some meaning to the "major release from 2026" which I'm happy that Python has a PEP to possibly make this change.
Which Python are we moving to? Oh yes, Python 26 (for 2026) sounds easy peasy. Maybe Django will move this way in the next decade or so.
@webology I’m on the record as thinking something like this would be helpful. Whether it happens or not, I can’t say.
For better or worse, no mad billionaire to push things through…
@carlton Well, I already subscribe to your newsletter, but yes, plz.
C(arlton's)EP has a nice ring to it 
@webology This month’s is Django related… but I’m unlikely to touch on the version numbering.
If I’m entirely straightforward, I think it’s the least of things we should be thinking of. There’s only so much give at any one time, only so much effort, and I’d really like us to think about where the ROI is highest on that. (That I likely will touch on 
) #TheStackReport
@carlton I don't blame you. Re: "it’s the least..." I disagree, but it's not worth your energy. I always look forward to the next Stack Report.
@webology I had to look it up recently, because that’s what I wanted at the time. My memory is a bit fuzzy again, but I think it made it harder to support the upgrade path from LTS to LTS, it would take that much longer to drop stuff that’s in an LTS.
https://groups.google.com/g/django-developers/c/MTvOPDNQXLI/m/kR02ZMXgsoMJ
@webology that’s cool. I was trying to respond to your wondering about the .0 not being the LTS. I’m not quite sure what you were showing me with the link, but I appreciate the resource.
@ryanhiebert I thought you were referencing the time between LTS dates.
@carlton I could get behind dropping LTS if support for the shorter releases were lengthened some. A lot of the projects out there lack sufficient test coverage to be able to support regular upgrades, so even if Django is stable, these projects would give some the impression that the framework is the problem.
@danielquinn You’ll note I didn’t actually suggest dropping the LTS