Mailing List Archive

On Sun, Jul 26, 2009 at 3:25 PM, Ben de Groot<yngwin@gentoo.org> wrote:
> Dear fellow devs,
>
> We would like to draw your attention to the fact that the Gentoo Qt team
> now officially discourages further usage of Qt3. Version 3 is no longer
> being developed or supported upstream. All users are strongly encouraged
> to use Qt version 4 where applicable.
>
> This means that we request that no new Qt3-based packages will be
> introduced to the portage tree. Also, where there is a choice, Qt4
> should be preferred over Qt3 by default.
>
> We will still do some minimal maintenance and support on the existing
> Qt3 libs and applications for about another 6 months. This will give
> application maintainers the time to upgrade or look for alternatives.
> After this period we will move the Qt3 libs and apps along with KDE3
> into an overlay where users of those legacy packages can maintain them.
>
> If you have a package for which both Qt3 and Qt4 versions exist, please
> have the Qt4-based version marked stable (if it isn't already) within
> the next few months and remove the legacy Qt3-based version, if possible.
>
> Thanks,
>
> Ben de Groot
> Gentoo Qt team lead
>
>

MythTV still uses Qt3 and there is NO way that the Qt4 based MythTV
could even remotely be considered stable. It is still undergoing
constantly changing and there are many codepaths that are incomplete.

--
Doug Goldstein

Ben de Groot wrote:
> Dear fellow devs,
>
> We would like to draw your attention to the fact that the Gentoo Qt team
> now officially discourages further usage of Qt3. Version 3 is no longer
> being developed or supported upstream. All users are strongly encouraged
> to use Qt version 4 where applicable.

Wait a minute. Qt3 is deprecated, but people are still adding new
Qt3-based packages to the tree:

On the 26th, scarabeus added gerix, as seen on our front page p.g.o feed:

net-wireless/gerix-0.20 Qt3 Based aircrack GUI

. . . wtf?

Ben de Groot wrote:
> Dear fellow devs,
>
> We would like to draw your attention to the fact that the Gentoo Qt team
> now officially discourages further usage of Qt3. Version 3 is no longer
> being developed or supported upstream. All users are strongly encouraged
> to use Qt version 4 where applicable.
>

This is gentoo-dev-announce material.

Regards,
Petteri

While I fully understand that people want to deprecate "old cruft", I assume
this is far too early. (just think back how long it took us to deprecate
gtk+-1)

I'm still on kde 3 and my previous three attemps to switch to kde 4 all ended
up in the conclusion that kde 4 is far from being stable yet. It has tons of
regressions.

The amount of qt3 apps not having a sane or just no qt4 port yet is probably
enormous. I also maintain such packages.


Though we don't know what the situation is in 6 months, I'd assume that
there's no chance we can consider qt3 fully deprecated then.

--
Hanno Böck Blog: http://www.hboeck.de/
GPG: 3DBD3B20 Jabber/Mail: hanno@hboeck.de

http://schokokeks.org - professional webhosting

On Tue, July 28, 2009 09:29, Hanno Böck wrote:
> While I fully understand that people want to deprecate "old cruft", I
> assume this is far too early. (just think back how long it took us to
> deprecate gtk+-1)

I fully agree with this thought. The qt4 world is simply nowhere near
as complete and usable (nor stable either, which is worse) as the qt3
one is.

> I'm still on kde 3 and my previous three attemps to switch to kde 4 all
> ended up in the conclusion that kde 4 is far from being stable yet. It has
> tons of regressions.

kde4 is just unusable for serious world. Magnificent work, yes, and some
programs can be used without too much problem. But kde4 as a whole just
can't be still compared to kde3.x

> The amount of qt3 apps not having a sane or just no qt4 port yet is
> probably enormous. I also maintain such packages.

There's absolutely no match at all in the linux world for k3b, and
whomever has used the qt4 port will agree that's it's unstable and
very far in which regards features to the qt3 version.

There are lots of qt3 apps that have no equivalent for qt4. I happen
to like kaffeine, I am pretty sure that everyone else can name a couple
of other examples.

Being that said, of course, there's no way that we can impose devs and
maintainers which to support. It's their choice. All I say is that, to
me, this deserves a much dedicated and deep thought.

The qt3 stuff is stable, this is not another xmms-like thing, it works,
and it doesn't only work, but today it works *far better* than the qt4
newer alternatives, in most cases anyway.

--
Jesús Guerrero

Doug Goldstein wrote:
> MythTV still uses Qt3 and there is NO way that the Qt4 based MythTV
> could even remotely be considered stable. It is still undergoing
> constantly changing and there are many codepaths that are incomplete.

We can re-assess the situation when the time comes, early next year.
There is still time, so we'll see. Of course MythTV is quite a major
application, so we may hold on to Qt3 for a bit longer because of it.

Cheers,
Ben

Josh Saddler wrote:
> Wait a minute. Qt3 is deprecated, but people are still adding new
> Qt3-based packages to the tree:
>
> On the 26th, scarabeus added gerix, as seen on our front page p.g.o feed:
>
> net-wireless/gerix-0.20 Qt3 Based aircrack GUI
>
> . . . wtf?

Indeed, the developer in question had forgotten we had agreed to this
new policy... This prompted me to announce this on list.

Cheers,
Ben

Jesús Guerrero wrote:
> On Tue, July 28, 2009 09:29, Hanno Böck wrote:
>> While I fully understand that people want to deprecate "old cruft", I
>> assume this is far too early. (just think back how long it took us to
>> deprecate gtk+-1)
>
> I fully agree with this thought. The qt4 world is simply nowhere near
> as complete and usable (nor stable either, which is worse) as the qt3
> one is.
>
>> I'm still on kde 3 and my previous three attemps to switch to kde 4 all
>> ended up in the conclusion that kde 4 is far from being stable yet. It has
>> tons of regressions.

Please note that you need to make a distinction between Qt4 and KDE4!
The KDE team had already decided to move KDE3 to an overlay early next
year, with the release of 4.4. We as Qt team are just following suit.
It's just that the KDE team hasn't made as public an announcement yet.

A KDE 4.3 version is supposed to be marked stable within the next few
months, and this then clears the way for KDE3 deprecation.

When KDE3 will be removed early 2010, we as Qt team see no need to keep
Qt3 around any longer either. We currently have no developers who are
interested in maintaining it, and upstream doesn't either. Qt4 really is
a major improvement over Qt3, especially where it concerns non-KDE apps.


>> The amount of qt3 apps not having a sane or just no qt4 port yet is
>> probably enormous. I also maintain such packages.
>
> [...] k3b, and [...] kaffeine

These are KDE apps, so you need to address the KDE team about that, not
the Qt team. But I'm assuming that if these have no working KDE4-based
version by early next year, that's just too bad. And on the other hand,
for those people who still want to use these applications, there will be
the overlay.

So, there are very few "pure" Qt3 apps, and even fewer that have no Qt4
port yet. The only major project I'm aware of is MythTV. If there are
any others please let me know.


> The qt3 stuff is stable, this is not another xmms-like thing, it works,
> and it doesn't only work, but today it works *far better* than the qt4
> newer alternatives, in most cases anyway.

No, Qt4 works far better than Qt3, and has loads more features. Of
course, it is up to the application developers how to use this. Do not
confuse KDE4 instability with Qt4 instability. If you compare non-KDE
Qt3 apps with non-KDE Qt4 apps, the huge improvements should be
self-evident.

Even so, our main argument is the question of maintenance. We want to
stop wasting time on Qt3 related issues, especially as (1) it is no
longer maintained upstream, and (2) its main consumer, KDE3, is
scheduled for removal from portage.

If you feel passionate about Qt3 and want to make the case for keeping
it in portage, then step up and become its maintainer.

Cheers,
Ben

On Mon, August 3, 2009 14:48, Ben de Groot wrote:
> Jesús Guerrero wrote:
>
>> On Tue, July 28, 2009 09:29, Hanno Böck wrote:
>>
>>> While I fully understand that people want to deprecate "old cruft", I
>>> assume this is far too early. (just think back how long it took us
>>> to deprecate gtk+-1)
>>
>> I fully agree with this thought. The qt4 world is simply nowhere near
>> as complete and usable (nor stable either, which is worse) as the qt3 one
>> is.
>>
>>> I'm still on kde 3 and my previous three attemps to switch to kde 4
>>> all ended up in the conclusion that kde 4 is far from being stable
>>> yet. It has tons of regressions.
>
> Please note that you need to make a distinction between Qt4 and KDE4!
> The KDE team had already decided to move KDE3 to an overlay early next
> year, with the release of 4.4. We as Qt team are just following suit. It's
> just that the KDE team hasn't made as public an announcement yet.

Fair enough :)

> A KDE 4.3 version is supposed to be marked stable within the next few
> months, and this then clears the way for KDE3 deprecation.

Not until all the kde 3 stuff has a port. I mentioned k3b and kaffeine
in the previous post. K3b/qt4 is *unusable* for a lot of people.
Of course, as you say, there's still a lot of time for it to improve.
So the whole issue will have to be re-evaluated when the time comes.

> When KDE3 will be removed early 2010, we as Qt team see no need to keep
> Qt3 around any longer either. We currently have no developers who are
> interested in maintaining it, and upstream doesn't either. Qt4 really is a
> major improvement over Qt3, especially where it concerns non-KDE apps.

I fully agree about that. And as said on the other post, there's no way
that I or anyone else can tell a developer what should s/he concentrate
on. If qt3 must die, then die it shall. I am just stating my view about
some aspects that might not have been considered by everyone. All in all,
I am a big supporter of the "do it yourself" philosophy: if something
doesn't suit you, fix it yourself, which means that if I truly need a
qt3 program I will take care myself. I am not going to scream around the
lists asking for support from anyone if qt3 becomes unsupported, and I
fully understand the aim behind this move, and agree with it because I
think that we must continue moving forward, and not fixing ancient stuff
which can only get worse and messier.

However, one can agree with the general picture while having some doubts
with specifics aspects (or just mere curiosity).

>>> The amount of qt3 apps not having a sane or just no qt4 port yet is
>>> probably enormous. I also maintain such packages.
>>
>> [...] k3b, and [...] kaffeine
>>
>
> These are KDE apps, so you need to address the KDE team about that, not
> the Qt team. But I'm assuming that if these have no working KDE4-based
> version by early next year, that's just too bad. And on the other hand,
> for those people who still want to use these applications, there will be
> the overlay.
>
> So, there are very few "pure" Qt3 apps, and even fewer that have no Qt4
> port yet. The only major project I'm aware of is MythTV. If there are any
> others please let me know.
>
>
>> The qt3 stuff is stable, this is not another xmms-like thing, it works,
>> and it doesn't only work, but today it works *far better* than the qt4
>> newer alternatives, in most cases anyway.
>
> No, Qt4 works far better than Qt3, and has loads more features. Of
> course, it is up to the application developers how to use this. Do not
> confuse KDE4 instability with Qt4 instability. If you compare non-KDE Qt3
> apps with non-KDE Qt4 apps, the huge improvements should be self-evident.

Leaving the semantics apart, the fact is that there's no kde3 without qt3,
and users might miss some *important* apps like kaffeine or k3b. A user
doesn't care if the application links with kdelibs or not. The fact is that
even if kde is not qt, and qt is not kde, it comes without a doubt that
there must be a synchrony on the move between the two teams. Otherwise the
final user is which suffers. If both are to be removed from portage at
the same time, then it's fine, some users might miss some applications
but well, they'll eventually find their way after screaming for a while
on the forums.

I am sure that whatever the decision and the timings are they will be
for the best nonetheless. :)

If we leave that semantic stuff apart as I said, I think that very few
persons will agree that k3b/qt4 is better than the qt3 version. Actually
it's the other way around. The qt4 port is too unstable. Yes, it's not qt4
fault, and I am not even saying that k3b is a showstoper: that's open
for discussion. I am just signaling it as one of the offended apps :)

> If you feel passionate about Qt3 and want to make the case for keeping
> it in portage, then step up and become its maintainer.

Not really. I use just a couple of qt3/kde3 programs, and not that much.
All I just wanted is to shake this topic a bit to hear what people and
developers think about this.

An interesting thread. Thanks for all the insight.
--
Jesús Guerrero