Opened 2 months ago

Last modified 6 days ago

#19442 new enhancement

qt5-5.15.13 (Possible Archive/Partial removal)

Reported by: Bruce Dubbs Owned by: blfs-book
Priority: normal Milestone: 12.2
Component: BOOK Version: git
Severity: normal Keywords:
Cc:

Description (last modified by Rahul Chandra)

New point version.

Perhaps we should remove the full qt5 and just stay with qt5-alternate.

Note that qt5-components is only needed for audacious (qtmultimedia), network-manager (qtdoc for examples), and wireshark (qtmultimedia if qt6 is not installed).

The reason those two modules were split off is due to their dependencies.

(Reopened as lxqt-2.0 was just added)

Change History (12)

in reply to:  description comment:1 by ken@…, 2 months ago

Replying to Bruce Dubbs:

New point version.

Perhaps we should remove the full qt5 and just stay with qt5-alternate.

Note that qt5-components is only needed for audacious (qtmultimedia), network-manager (qtdoc for examples), and wireshark (qtmultimedia if qt6 is not installed).

The reason those two modules were split off is due to their dependencies.

For audacious, on my qt6 systems I'm using the following changes for both audacious and its plugins (and omitting -Dgtk3=true)

 -Dgtk=false \
 -Dqt6=true \

I will again mention that unless you specify -Dgtk=false it will look for gtk2 irrespective of the other choices.

With qt6 the controls have text labels instead of icons for pause, play, etc.

comment:2 by Bruce Dubbs, 8 weeks ago

Milestone: 12.2hold
Summary: qt5-5.15.13qt5-5.15.13 (Hold until lxqt-2.x)

After lxqt-2.x re-examine need and possibly archive qt5.

comment:3 by Rahul Chandra, 3 weeks ago

Description: modified (diff)
Milestone: hold12.2

comment:4 by Rahul Chandra, 3 weeks ago

Summary: qt5-5.15.13 (Hold until lxqt-2.x)qt5-5.15.13 (Possible Archive/Partial removal)

comment:5 by Rahul Chandra, 3 weeks ago

I think we can archive Qt5 at this point, we should probably leave alternate in but after 12.2 we should be able to remove it entirely.

comment:6 by Douglas R. Reno, 3 weeks ago

I'm currently not at home so typing this on my phone, but isn't Qt5 a requirement for VLC? VLC is needed for Phonon for Plasma still, so we really need that to be Qt6 ready

comment:7 by Bruce Dubbs, 3 weeks ago

For vlc we currently have:

Optional interface plugins

  • (Qt-5.15.12 or qt-alternate-5.15.12) (required for the graphical user interface).

Wouldn't mpv or mplayer be alternatives for a player with a UI?

comment:8 by thomas, 2 weeks ago

Just for the record:

I've to add

sed -e "/c++1z/s;^CONFIG;#CONFIG;" \
    -e "s;c++14;c++17;" \
    -i qtlocation/src/3rdparty/mapbox-gl-native/mapbox-gl-native.pro

and add

-c++std c++17

to the configure instruction to get our current qt5 to compile with icu-75.

Last edited 2 weeks ago by thomas (previous) (diff)

in reply to:  8 comment:9 by Xi Ruoyao, 6 days ago

Replying to thomas:

Just for the record:

I've to add

sed -e "/c++1z/s;^CONFIG;#CONFIG;" \
    -e "s;c++14;c++17;" \
    -i qtlocation/src/3rdparty/mapbox-gl-native/mapbox-gl-native.pro

and add

-c++std c++17

to the configure instruction to get our current qt5 to compile with icu-75.

Note that technically mapbox-gl-native is not compatible with C++17. The code is using some C++ features removed in C++17. So this hack is not sound.

I'm leaning forward to just archive Qt5.

comment:10 by Xi Ruoyao, 6 days ago

Another possibility is just keep qt5-alternate for VLC. Qtlocation is not built for qt5-alternate, so the bit rot mapbox-gl-native won't be a issue.

comment:11 by Bruce Dubbs, 6 days ago

From Uwe Düffert via the mailing list:

 To get it (5.15.13) built (on current system including icu-75.1) I also needed: 
mkdir -pv qtbase/.git

sed -e "/pragma once/a#include <cstdint>" -i \
   qtlocation/src/3rdparty/mapbox-gl-native/include/mbgl/util/geometry.hpp \
   qtlocation/src/3rdparty/mapbox-gl-native/include/mbgl/util/string.hpp \ 
   qtlocation/src/3rdparty/mapbox-gl-native/src/mbgl/gl/stencil_mode.hpp

sed -e 's@std=c++1y@std=c++17@g' -i qtbase/mkspecs/common/g++-base.conf

in reply to:  11 comment:12 by Xi Ruoyao, 6 days ago

Again 's@std=c++1y@std=c++17@g' is unsound and we should avoid it, unless we really have no option.

But to me the alternative is quite clear: just disable qtlocation or switch to qt-alternative completely.

Note: See TracTickets for help on using tickets.