| Git Clone URL: | https://aur.archlinux.org/qt4.git (read-only, click to copy) |
|---|---|
| Package Base: | qt4 |
| Description: | A cross-platform application and UI framework |
| Upstream URL: | https://www.qt.io |
| Licenses: | custom, GPL3, LGPL, FDL |
| Conflicts: | qt |
| Replaces: | qt |
| Submitter: | arojas |
| Maintainer: | dviktor |
| Last Packager: | dviktor |
| Votes: | 74 |
| Popularity: | 1.58 |
| First Submitted: | 2019-05-01 11:51 (UTC) |
| Last Updated: | 2022-05-13 12:03 (UTC) |
To fix breakage with GCC 9 and the foreach macro we should backport this commit from Qt 5 (it doesn't apply cleanly and needs a bit of trivial adapting): https://github.com/qt/qtbase/commit/c35a3f519007af44c3b364b9af86f6a336f6411b
GCC developers closed https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90617
as a duplicate of https://gcc.gnu.org/bugzilla/show_bug.cgi?id=44715
which shows that the change was intentional.
Just wanted to let everyone know that there is a bug in GCC 9 (now in testing), which makes it miscompile the foreach macro in Qt4:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90617
Since that macro is used extensively throughout the Qt source code I can only imagine that compiling qt4 with the current testing gcc will cause severe breakage.
We can fix this in two ways:
1) Wait for gcc fix 2) Apply a patch to qt4 to change the way Q_FOREACH is coded.
Missing dependency gstreamer0.10-base?
cd WebCore/ && $/aur-qt4/src/qt-everywhere-opensource-src-4.8.7/bin/qmake $/aur-qt4/src/qt-everywhere-opensource-src-4.8.7/src/3rdparty/webkit/Source/WebCore/WebCore.pro -spec $/aur-qt4/src/qt-everywhere-opensource-src-4.8.7/mkspecs/linux-g++ -o Makefile.WebKit
Project ERROR: Package gstreamer-app-0.10 not found
make[1]: *** [Makefile.WebKit:72: WebCore/Makefile.WebKit] Error 2
@dviktor: Can you add that the right script for the devtools is extra-x86_64-build?
@Tharbad no, it's not proper way to perform clean chroot build. Please follow instructions in wiki
Used yay clean build option. Got:
In file included from ./wtf/FastAllocBase.h:93,
from ./wtf/RefCounted.h:25,
from ./wtf/CrossThreadRefCounted.h:36,
from ./wtf/text/StringImpl.h:28,
from ./runtime/UString.h:26,
from yarr/YarrPattern.h:30,
from yarr/YarrInterpreter.h:29,
from yarr/YarrInterpreter.cpp:28:
./wtf/TypeTraits.h:186:69: error: ‘std::tr1’ has not been declared
template<typename T> struct HasTrivialConstructor : public
std::tr1::has_trivial_constructor<T> { };
^~~
./wtf/TypeTraits.h:186:74: error: expected ‘{’ before ‘has_trivial_constructor’
template<typename T> struct HasTrivialConstructor : public
std::tr1::has_trivial_constructor<T> { };
^~~~~~~~~~~~~~~~~~~~~~~ ./wtf/TypeTraits.h:187:68: error: ‘std::tr1’ has not been declared template<typename T> struct HasTrivialDestructor : public std::tr1::has_trivial_destructor<T> { }; ^~~ ./wtf/TypeTraits.h:187:73: error: expected ‘{’ before ‘has_trivial_destructor’ template<typename T> struct HasTrivialDestructor : public std::tr1::has_trivial_destructor<T> { };
^~~~~~~~~~~~~~~~~~~~~~
yarr/YarrInterpreter.cpp: In member function ‘bool
JSC::Yarr::Interpreter::backtrackParenthesesOnceEnd(JSC::Yarr::ByteTerm&,
JSC::Yarr::Interpreter::DisjunctionContext*)’:
yarr/YarrInterpreter.cpp:707:13: warning: this statement may fall through [-
Wimplicit-fallthrough=]
if (backTrack->begin == notFound) {
^~
yarr/YarrInterpreter.cpp:711:9: note: here
case QuantifierNonGreedy:
^~~~
make[2]: *** [Makefile.WebKit:1062: .obj/release-static/YarrInterpreter.o] Error 1
make[2]: Leaving directory '/tmp/yay/qt4/src/qt-everywhere-opensource-src-4.8.7/src/3rdparty/webkit/Source/JavaScriptCore'
make[1]: *** [Makefile.WebKit:43: sub-JavaScriptCore-JavaScriptCore-pro-make_default-ordered] Error 2
make[1]: Leaving directory '/tmp/yay/qt4/src/qt-everywhere-opensource-src-4.8.7/src/3rdparty/webkit/Source'
make: *** [Makefile:747: sub-webkit-make_default-ordered] Error 2
==> ERROR: A failure occurred in build().
Aborting...
Also, when trying to remove qt4 I noticed it's dependency of kde.
Just for example there will be never alternative for KBGOffice Dictionary for it is abandoned years ago but is the only one Eng<->Bg dictionary for Linux. Some packages just doesn't have alternatives.
Except none of that makes sense. sip does not need qt4, and it doesn't need qt5 either -- it just generates code for them. And the real issue with sip is that in the latest version of sip, the upstream developer of sip, removed all qt4 support. https://www.riverbankcomputing.com/hg/sip/rev/328f2e872d64
All three of those packages are AUR packages... it's not unheard of for one AUR package to depend on another. And spflashtool-bin requires qtwebkit in addition to qt4... do you realize qtwebkit has been in the AUR for two years now?
This is not about "drying the well", and we're not trying to force anyone to port anything. We're simply saying that for years we've only begrudgingly supported qt4 at all, and we finally decided to stop. There's no use getting upset about the lack of support for really old things; we don't support KDE3, gnome2, linux 3.x either, now it is qt4's turn.
My advice is to port it or ask someone else to port it, because getting it ported will make it easier to run without requiring AUR packages, and will be beneficial in the long run either way.
Alternatively, look for modern alternatives to these packages.
@Tharbad recheck your /etc/makepkg.conf and also show us steps how do you create clean chroot. Please post full relevant error so we can inspect it.
@piedro I think it's clear now how to build this package. For those who can't build it under usual environment method with clean chroot exists and works just fine.
It's inevitable that some software becomes obsolete and out-of-date, new versions will replace the old one - this is normal process of software development. Great example are GCC compilers version 6 and 7 (and, in near future, 8) - they are also moved out to the AUR and also requires a lot of time to build but I don't see anyone who complains about it. If you're not satisfied with AUR solution you can always find binary repository for qt4 and install it directly.
@piedro +1
Pinned Comments
dviktor commented on 2022-09-28 18:33 (UTC)
PKGBUILDfor this package is being maintained there. You can install pre-built packages from my repodviktor commented on 2019-06-18 12:50 (UTC) (edited on 2022-01-27 12:28 (UTC) by dviktor)
Qt4 binary package has moved to my personal repository located here. You can add it to your
/etc/pacman.conffor convenience, if you want (don't forget to import my GPG key).eschwartz commented on 2019-05-09 13:24 (UTC)
@semeion,
qt4 was expelled from the official repositories due to https://lists.archlinux.org/pipermail/arch-dev-public/2019-April/029560.html
It will not be restored, because software needs to stop using qt4. It's fine for people to still use it via the AUR if they have old software that is not ported, but the proper solution is to get that software ported to qt5.
"It takes a long time to compile" is not a reason to move it to community.
@xuanruiqi,
Only Developers and Trusted Users have access to pkgbuild.com, and we will not be uploading qt4 there. If we wanted qt4, we would upload it to community, but we don't -- we have managed to move every package still being actively maintained in the official repos, over to qt5, and we want to stay that way.
...
Again, the proper long-term solution is to get software ported over to qt5.
dviktor commented on 2019-05-05 17:49 (UTC) (edited on 2019-05-15 19:02 (UTC) by dviktor)
For those who have problems with
‘std::tr1’ has not been declarederror: build in clean chroot withextra-x86_64-buildscript.