long due TO-DO item: removal of Qt3 from Debian

A couple of weeks ago was the first anniversary of orphaning Qt3 in Debian, see bug 625502.

In this year, Qt3 has got a few QA uploads with the most relevant change being support to multiarch. And, more importantly, nobody seemed to care enough to step into maintaining it.

In the last days, I have taken a look into how much needed to be done to remove Qt3 and there were slightly more than 50 packages depending directly or indirectly from Qt3. A removal from Wheezy seemed doable given that removing packages is never a problem during the Debian freeze ;-)
All the packages affected have a bug opened since more than one year and half ago and I have pinged all the bugs with some maintainers responding quick (thanks!). I also filed some removals for packages that were clearly unmaintained and didn’t seem worth keeping, with ftp-masters responding quick too (Thanks!). And finally, a couple of QA upload for orphaned software that were still useful without Qt3.

There is a wiki page tracking the status of the removal if you are curious:
http://wiki.debian.org/qt3-x11-freeRemoval

If in the future, you are reading this and you need Qt3 in Wheezy, you can fetch it from Debian snapshots.

13 thoughts on “long due TO-DO item: removal of Qt3 from Debian

  1. Z_God

    I hope the Trinity Desktop project effort is also considered regarding this. See http://www.trinitydesktop.org/

    Second, as far as I am aware, Qt3 is part of the LSB. I suggest caution with making Qt3 unavailable other than regarding the packaged software.

  2. ana Post author

    @Z_God: The Trinity Desktop project is understaffed so you hardly can rely on them to give support.

    About LSB, well… a small collateral damages.

  3. Kevin Kofler

    Meanwhile, Fedora is still shipping both qt3 and kdelibs3. We found those packages to require very little effort to keep them going, so IMHO dropping them is a pointless disservice to your users.

  4. Malvin

    @Kevin Kofler:
    People wanting the packages it in Debian had a year to take over them and nobody seemed to care.

  5. ana Post author

    @Kevin Kofler: Nobody in Debian is willing to invest that ‘very little effort’ that you say. I disagree greatly that is little effort…
    Time is very scarce and it is better invested in getting KDE 4.8 or Qt5 pre-releases packaged. We are talking here about software that was EOL 5 years ago :-)

  6. Sven

    Maybe there is a slight misunderstanding about what is meant by “keep going”. To keep a package building is certainly easy, yet maintaining is something completely different and requires activity as well as knowledge of the code. Which is very hard to cope with if all experts on that code moved on.

    I think it is a good decision to drop packages that build but are de-facto unmaintained.

  7. ana Post author

    @Sven: thank sven, good point. When I talk of maintenance I talk of both: Debian maintenance (that needs some time but it is feasible) and ‘upstream maintenance’ (that requires a lot of knowledge and can quite time consuming)

  8. Kevin Kofler

    @Sven: Those libraries have very few bugs that need fixing (mainly FTBFS issues with current g++, a couple aliasing issues with newer g++, and the occasional security issue). Given that those are compatibility libraries, I don’t see what other “upstream” maintenance you’d even need other than build fixes, build-induced regression fixes (e.g. stricter GCC/g++ strict aliasing) and (of course) security fixes. Nobody is (nor should be) reporting functionality bugs (and definitely not feature requests) against those libraries.

  9. poroto

    Removing qt3 is a BAD IDEA, too many software non-Debian packaged still depends on it. Will make life harder for many people.

  10. ana Post author

    @poroto: could you please share the information you use to say ‘too many’ software out of Debian do use Qt3?

Comments are closed.