This test hangs in a QEMU VM running LFS r12.0-102 (built with jhalfs
w/o any modification) even if -fsanitize=thread is not used... While
it's difficult to root cause the issue, let's not hang forever.
Link:https://github.com/llvm/llvm-project/issues/39012
To use clang with BPF for v4l-utils, libbpf must be installed but it's
not a BLFS package. And I doubt if any BLFS user really uses an IR
remote control on the BLFS system. So let's not make all BLFS users
build something useless.
Note that if both clang and libbpf both exist, v4l-utils will attempt to
use "clang -target bpf" anyway even if -Dbpf=disabled is explicitly
specified (we've got enough surprise from v4l-utils building system
these days!!), so add a note to v4l-utils for people installing non-BLFS
libbpf.
1. Allow building without gtk-doc.
2. Run "configure" manually instead of let autogen.sh run it. It's
because autogen.sh sets CFLAGS, producing binaries with optimization
off and debug on. More seriously, the debug feature it uses
(sanitizer) is not suitable for a release build per GCC security
policy.
3. Add a link to the external dependency libdbusmenu.
I've marked most of these as 'role=nodep' so that jhalfs doesn't pick up
multiples in cases where it's not needed.
Transmission, Libreoffice, CMake, Poppler, and Appstream now have Qt6
ports working well.
I have left Qca, gpgme, polkit-qt, and other KDE-related packages alone.
They do have support for Qt6, but we should not list them as they are
incompatible with the Qt5 versions and we could end up causing
unintentional breakage by doing this.
- Make some grammar improvements to the description
- Use &qt5-deps instead of just Qt5, since this package uses qt5-base.
This package will build with the alternate version as well
