[calligra] /: Qt3Support is also needed in Calligra.

[Jaroslaw Staniek]

On 2 December 2011 16:49, Markus Slopianka <markus.s at kdemail.net> wrote:
> On Freitag 02 Dezember 2011 15:45:42 Jaroslaw Staniek wrote:
>
>> I'd like to see plans for backporting KDE 5 modules to currently used
>> distros, let's say, all lines of Ubuntu 10.x.
>
> Already released distributions are in package / version freeze. No new packages will be
> introduced into standard repositories.
> Ubuntu 10.x doesn't contain any version of Calligra at all anyway (only KOffice).
>
>
>> > So yeah, many users will get Qt5/KF5 pretty soon. I don't think it would
>> > be wise to do any Calligra release after summer 2012 that requires Qt 4.
>>
>> When will this end? With this deployment strategy, again and again we
>> lack the stability - constantly (practically) requiring to reinstall
>> whole OSes to get minor upgrade for apps.
>
> New major versions of applications do not get included into already frozen standard
> repositories anyway (Firefox and Chromium due their development cycle are usually two
> exceptions).
>
> If you want to discuss deployment of packages by distributors, contact them.
> I merely explained the situation.

There is life outside of standard repos. Finishing Calligra/Qt 4
release cycle after just one release will no doubt block possibility
to ship fresh Calligra software (and frameworks) to (I believe)
majority of users. So this decision needs to be discussed within the
team, with pros vs cons analyzed, also from each app's POV. I am sure
we'll have schedule soon. And I am openly providing some factors. Our
goal is first to have software easily available, as frequently updated
as possible, then having quality that comes from interaction with the
right user base (integrators, power users for now, specialized cases),
and only then to be test suite for new frameworks. This is my opinion.

For example on Windows we are have contributors (in this case Calligra
developers) that deploy Calligra software without depending on
bleeding edge Windows 7 or 8 but supporting most popular versions.

I see no reason to so quickly decide on going with bleeding edge
distros only for KDE software when the newest competitors in
frameworks now (Firefox and Chromium, you mentioned them) enjoy of
being treat in special way. Their HTML5-based 'toolkits' and alike
have time from release to upgrade measured is days instead of weeks,
and updates taking hours, not days. It's not yet on the desktop but I
can resend you this very note in two years...

>> 10.04 LTS  support ends in 04/2013.
> No Calligra included.
>
>> 12.04 LTS support ends in 04/2017 - it'll be very very new by the time
>> of our release and it will include Qt 4 of course. It's not guessing
>> if I say not every apps that use Qt4 will drop Qt4 by the time. Also
>> KDE apps. So the distro will include KDE apps v4, right?
>
> Yes, they may include them and never upgrade them within a distribution version to a new
> major app release. Ubuntu 12.04 will have Calligra 2.4 in its repos and never upgrade it
> for 12.04 to a new Calligra major version.
>
>
>> Why Linux users cannot have access to installable apps as on other
>> OSes and have only to deal with what was available within the time of
>> first 1 or 1.5 years?
>
> This mailing list is not the place to discuss how distributors should behave.
>
>
>> If you tell me that's not your business, I am OK with this. But the
>> explanation that users generally constantly upgrade OSes is not the
>> answer for me.
>
> Again: This mailing list is not the place to discuss how distributors should behave.

As said above, all the target distros so far allow for custom
repositories, so no need expect changes in how distros release the
software.
It's not that anyone among Calligra contributors have to declare
investments into Calligra packages maintenance for the current most
popular stable distros. But better to have this option available.

> Fact is: Once Qt 5 is released, Qt 4 reaches End Of Life (maybe Digia will continue to
> support the closed source version of Qt 4 longer). Still depending on Qt4 for new Calligra
> releases after summer 2012 would mean to force users to install Qt 4 and potentially
> expose them to bugs in Qt 4 that may never get fixed.

Even assuming porting is easy as planned and there are no regressions
for QWidget world, there is big module not ported even to Qt 4 in
Kexi.
Kexi has technologies more important than Qt 5 where it should be
ported, like Predicate or new mdbtools. It's good to note that for
certain apps users won't notice change from Qt 4 to 5 and to KF5.
Another reason why maintainers of these apps may want to assign this
task 3rd priority.

Please understand me: this discussion is about availability of
Calligra for most of the adopters from the release #1 (i.e. current
stable/established not just bleeding edge).

Finally, one may except that we should first know the usage statistics
per distro/version. So I plan to collect this anonymous data
specifically for Calligra apps that want to join the effort. The above
reason is one (but important) to do that.

>> I don't propose to have Calligra/Qt4 maintenance until 2017 but to
>> consider another Qt4-based release after 2.4. We plan to release more
>> frequently in Calligra.
>
> Again: New Calligra major versions won't get included in existing standard repositories
> anyway.
> New major version would be required to be obtained via an unofficial 3rd party repository.
> That repo may just as well hold Qt 5 / KF5.

This is always an option. Is co-installability of v4 and v5 a requirement?

Thanks for your input, as always, it helps a lot!

-- 
regards / pozdrawiam, Jaroslaw Staniek
 http://www.linkedin.com/in/jstaniek
 Kexi & Calligra (kexi-project.org, identi.ca/kexi, calligra-suite.org)
 KDE Software Development Platform on MS Windows (windows.kde.org)