[Development] Commercial-only LTS phase starts: Closing the 5.15 branch(es) on 5th January

[Volker Hilsheimer]

> On 5 Jan 2021, at 16:23, Scott Bloom <scott at towel42.com> wrote:
> 
> On Monday, 4 January 2021 20:32:27 -03 Scott Bloom wrote:
>>> The funny thing, I remember at a Qt Dev Days when Qt 5 was about a 
>>> year away.  The "we will never do a Qt 3-4 type major version change 
>>> again" was said time and time again.
>>> 
>>> Where functionality was missing, and no one was happy with Qt 4.0 
>>> except the people who got to say "4.0" released
>> 
>> Both statements are true but their combination is misleading.
>> 
>> Yes, Qt developers promised never to do a Qt 3 to 4 transition again. Yes, Qt
>> 4.0 was missing some functionality from Qt 3, which made 4.0 an impossible target, as well as having brand new functionality that was hadn't yet matured (itemviews).
>> 
>> But the first statement is taken out of context and thus makes it sound that the second statement is the reason for the first. It is not. The reason why we didn't want a 4.0 transition is because of the massive amount of code changes that were required for code bases to transition to 4.0. That is not the case here, as most codebases can compile with 6.x with limited porting effort.
>> 
>> It's also true that some of them can't be ported to 6.0, but once 6.1 and 6.2 come along, they will be.
>> 
>> The concern raised by people who can't move forward and aren't getting bug fixes either is a valid one, though.
>> 
>> --
> I remember the discussion in terms of "never again having a 3-4" type transition, to be two fold.  First it was definitely the level of changes to the API, no doubt.  But it was also, releasing a x.0 with missing functionality that would prevent migration.
> 
> Scott


Independent of what was or wasn’t said in some Qt Dev Days event in the past [1], the question is: would it really have been a good idea to not release the essential modules of Qt with Qt 6.0 until every single module has been ported over? We can argue whether another week or another month would have helped in any way (and how bad is it anyway?), but waiting for every module would probably have meant another year or more, I’d assume.

The way we have it now, we can get feedback from early adopters and make improvements to the essential modules while porting the add-ons, and in a year we should have a pretty solid platform. How is that not preferable (assuming that the quality of the Qt 6.0 modules released is not complete rubbish)?


Apart from that: is Qt 5.15.2 really so broken that people can’t use it without getting more patches?

Volker


[1] My memory might fail me, but I don’t even recall Qt 3 being multiple libraries or modules, and can’t recall at all that we removed any substantial Qt 3.3 features in Qt 4.0. There wasn’t that much in Qt 3 that wouldn’t have been essential functionality anyway.