[oe] [meta-networking][PATCH] mdns: improve numerous aspects of Posix backend
Apple's default implementation of the Posix backend for mDNSResponder has a number of weaknesses. Address several of them, most notably: * Improve interface tracking, preventing confusion to mdns's state machine. Prevents spurious removal/republication cycles whenever network interfaces are added or removed. * Support network interfaces whose indeces are great than 31. Indices grow past that range surprisingly quickly, especially with multi- homed, mobile, wifi, Bluetooth, VPN, VLANs, or other interfaces present. * Correctly handle edge cases during removal of a network interface. The fixes are kept as a patch series for clarity. Signed-off-by: Matt Hoosier --- ...utine-for-cleaning-recent-interfaces.patch | 60 + ...outine-for-tearing-down-an-interface.patch | 58 + .../0003-Track-interface-socket-family.patch | 50 + ...0004-Use-list-for-changed-interfaces.patch | 177 +++ .../0005-Handle-noisy-netlink-sockets.patch | 212 ++ .../files/0006-Remove-unneeded-function.patch | 51 + ...cate-loopback-interface-to-mDNS-core.patch | 129 +++ ...-deleted-interfaces-as-being-changed.patch | 39 .../0009-Fix-possible-NULL-dereference.patch | 45 ...0010-Handle-errors-from-socket-calls.patch | 62 + ...ic-allocation-to-file-scope-variable.patch | 51 + .../recipes-protocols/mdns/mdns_878.260.1.bb | 11 + 12 files changed, 945 insertions(+) create mode 100644 meta-networking/recipes-protocols/mdns/files/0001-Create-subroutine-for-cleaning-recent-interfaces.patch create mode 100644 meta-networking/recipes-protocols/mdns/files/0002-Create-subroutine-for-tearing-down-an-interface.patch create mode 100644 meta-networking/recipes-protocols/mdns/files/0003-Track-interface-socket-family.patch create mode 100644 meta-networking/recipes-protocols/mdns/files/0004-Use-list-for-changed-interfaces.patch create mode 100644 meta-networking/recipes-protocols/mdns/files/0005-Handle-noisy-netlink-sockets.patch create mode 100644 meta-networking/recipes-protocols/mdns/files/0006-Remove-unneeded-function.patch create mode 100644 meta-networking/recipes-protocols/mdns/files/0007-Indicate-loopback-interface-to-mDNS-core.patch create mode 100644 meta-networking/recipes-protocols/mdns/files/0008-Mark-deleted-interfaces-as-being-changed.patch create mode 100644 meta-networking/recipes-protocols/mdns/files/0009-Fix-possible-NULL-dereference.patch create mode 100644 meta-networking/recipes-protocols/mdns/files/0010-Handle-errors-from-socket-calls.patch create mode 100644 meta-networking/recipes-protocols/mdns/files/0011-Change-a-dynamic-allocation-to-file-scope-variable.patch diff --git a/meta-networking/recipes-protocols/mdns/files/0001-Create-subroutine-for-cleaning-recent-interfaces.patch b/meta-networking/recipes-protocols/mdns/files/0001-Create-subroutine-for-cleaning-recent-interfaces.patch new file mode 100644 index 0..692c344db --- /dev/null +++ b/meta-networking/recipes-protocols/mdns/files/0001-Create-subroutine-for-cleaning-recent-interfaces.patch @@ -0,0 +1,60 @@ +From 89ea6ac4a8840e8c2be0140a9805c6522c6c5280 Mon Sep 17 00:00:00 2001 +From: Nate Karstens +Date: Wed, 28 Jun 2017 17:30:00 -0500 +Subject: [PATCH 01/11] Create subroutine for cleaning recent interfaces + +Moves functionality for cleaning the list of recent +interfaces into its own subroutine. + +Upstream-Status: Submitted [d...@apple.com] + +Signed-off-by: Nate Karstens +--- + mDNSPosix/mDNSPosix.c | 24 ++-- + 1 file changed, 14 insertions(+), 10 deletions(-) + +diff --git a/mDNSPosix/mDNSPosix.c b/mDNSPosix/mDNSPosix.c +index 0e10bd5..ffc9696 100644 +--- a/mDNSPosix/mDNSPosix.c b/mDNSPosix/mDNSPosix.c +@@ -856,6 +856,19 @@ mDNSlocal int SetupSocket(struct sockaddr *intfAddr, mDNSIPPort port, int interf + return err; + } + ++// Clean up any interfaces that have been hanging around on the RecentInterfaces list for more than a minute ++mDNSlocal void CleanRecentInterfaces(void) ++{ ++PosixNetworkInterface **ri = ++const mDNSs32 utc = mDNSPlatformUTC(); ++while (*ri) ++{ ++PosixNetworkInterface *pi = *ri; ++if (utc - pi->LastSeen < 60) ri = (PosixNetworkInterface **)>coreIntf.next; ++else { *ri = (PosixNetworkInterface *)pi->coreIntf.next; free(pi); } ++} ++} ++ + // Creates a PosixNetworkInterface for the interface whose IP address is + // intfAddr and whose name is intfName and registers it with mDNS core. + mDNSlocal int SetupOneInterface(mDNS *const m, struct sockaddr *intfAddr, struct sockaddr *intfMask, const char *intfName, int intfIndex) +@@ -1010,16 +1023,7 @@ mDNSlocal int SetupInterfaceList(mDNS *const m) + + // Clean up. + if (intfList != NULL) free_ifi_info(intfList); +- +-// Clean up any interfaces that have been hanging around on the RecentInterfaces list for more than a minute +-PosixNetworkInterface **ri =
[oe] Method for specifying which among a recipe's multiple available licenses is desired?
Hi, I'm searching for an annotation that I can insert (perhaps with a .bbappend) into a recipe that indicates which of the multiple licensing terms I elect to use for a given recipe. The regular LICENSE_FLAGS_WHITELIST variable is somewhat related, but it works in the opposite way: make an exception to global preference against a license. What I had in mind would work as follows. Suppose foo.bb has: LICENSE="MIT BSD LGPLv2 LGPLv3" If I choose to take only the MIT or LGPLv2 terms of use for this one package (not as an overall global prohibition against the remaining licenses), the idea is that I would write something like: LICENSE_CHOSEN="MIT LGPLv2" The main desired outcome is that the build/tmp/deploy/license/ directory will only contain the texts of the license terms that I actually accepted, rather than all conceivable ones under which Foo is offered. It would be a bonus if some Bitbake magic behind the scenes would validate that there is a nonempty intersection between ${LICENSE} and ${LICENSE_CHOSEN}. If there's no overlap, an explicit fatal error would be nice. But this isn't a deal-breaker; I can infer this from an empty build/tmp/deploy/license/ directory. Are there any existing add-on bbclass'es that might do this? I had a brief try at rolling my own, but parsing the nested LICENSE text involving && and || operators looks non-trivial. -- ___ Openembedded-devel mailing list Openembedded-devel@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-devel
Re: [oe] [meta-qt5][PATCH] qtbase: fix Krogoth build regression from efa8aaf
On Mon, Aug 21, 2017 at 1:31 PM, Khem Rajwrote: > On Mon, Aug 21, 2017 at 2:51 AM, Martin Jansa wrote: >> I know it's not caused by this change, but it still creates "foodummy" file >> in ${TOPDIR} can we please get rid of it or at least move it somewhere else >> like ${WORKDIR}? > > ah thats where it comes from, I was wondering if its some of my workspace > tooling doing this. But thanks now I will less annoyed to see it. What branch of meta-qt5 is it that's getting built when you still see these build/foodummy files? There are no remaining occurences of the string 'foodummy' anywhere that I can see on jansa/master where my patch got applied: $ git remote -v github git://github.com/meta-qt5/meta-qt5 (fetch github git://github.com/meta-qt5/meta-qt5 (push) $ git checkout github/jansa/master $ git grep foodummy -Matt -- ___ Openembedded-devel mailing list Openembedded-devel@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-devel
Re: [oe] [meta-qt5][PATCH] qtbase: fix Krogoth build regression from efa8aaf
On Thu, Aug 17, 2017 at 7:27 AM, Matt Hoosier <matt.hoos...@gmail.com> wrote: > > From: Matt Hoosier <matt.hoos...@garmin.com> > > With > > commit efa8aaf82e580a7d32eaaab48eb92d436f2e222a > Author: Andreas Müller <schnitzelt...@googlemail.com> > Date: Thu Feb 9 00:26:09 2017 > > qmake5_base.bbclass: set qt.conf by environment variable again > > we stopped pointing ${OE_QMAKE_QTCONF_PATH} at a valid file and > instead directed it to a path which was intended not to exist. > > The motivation was to permit qtbase/qtbase-native/nativesdk-qtbase > to build again after Qt 5.8 started paying attention to the > contents of this file. > > The change as done in efa8aaf works well enough for Morty and > subsequent releases' copies of Bitbake, but fails on earlier > releases because they lack the following change: > > commit 2afcbfef2cd1ca568e5225884a8021df38ee3db0 > Author: Ross Burton <ross.bur...@intel.com> > Date: 2016-07-14 13:56:22 > > bitbake: build: don't use $B as the default cwd for functions > > The result is that when we build with Krogoth or prior, the body > of do_generate_qt_config_file() runs with a cwd of ${B}, which was > _not_ the intent of efa8aaf. Because the working directory is ${B}, > ${OE_QMAKE_QTCONF_PATH} is written in there too. do_configure() -- > whose cwd is by design also ${B} -- then finds the file 'foodummy', > and the build breaks for the reasons outlined in efa8aaf. > > This change simply shifts the implementation tactics to suppress > the creation of ${OE_QMAKE_QTCONF_PATH} during qtbase rather than rely > on unspecified behavior about the cwd of do_generate_qt_config_file(). > > (cherry picked from commit a17ff281aa8d99f770b0a049cb798235005fb93e > at https://codereview.qt-project.org/yocto/meta-qt5.) > --- > recipes-qt/qt5/nativesdk-qtbase_git.bb | 2 +- > recipes-qt/qt5/qtbase-native_git.bb| 2 +- > recipes-qt/qt5/qtbase_git.bb | 2 +- > 3 files changed, 3 insertions(+), 3 deletions(-) > > diff --git a/recipes-qt/qt5/nativesdk-qtbase_git.bb > b/recipes-qt/qt5/nativesdk-qtbase_git.bb > index f2ff5f7..ffbb917 100644 > --- a/recipes-qt/qt5/nativesdk-qtbase_git.bb > +++ b/recipes-qt/qt5/nativesdk-qtbase_git.bb > @@ -94,7 +94,7 @@ OE_QMAKE_PATH_HOST_LIBS = "${libdir}" > > # for qtbase configuration we need default settings > # since we cannot set empty set filename to a not existent file > -export OE_QMAKE_QTCONF_PATH = "foodummy" > +deltask generate_qt_config_file > > do_configure() { > ${S}/configure -v \ > diff --git a/recipes-qt/qt5/qtbase-native_git.bb > b/recipes-qt/qt5/qtbase-native_git.bb > index 98cbecc..0884a19 100644 > --- a/recipes-qt/qt5/qtbase-native_git.bb > +++ b/recipes-qt/qt5/qtbase-native_git.bb > @@ -79,7 +79,7 @@ PACKAGECONFIG_CONFARGS = " \ > > # for qtbase configuration we need default settings > # since we cannot set empty set filename to a not existent file > -export OE_QMAKE_QTCONF_PATH = "foodummy" > +deltask generate_qt_config_file > > do_configure_prepend() { > # Avoid qmake error "Cannot read > [...]/usr/lib/qt5/mkspecs/oe-device-extra.pri: No such file or directory" > diff --git a/recipes-qt/qt5/qtbase_git.bb b/recipes-qt/qt5/qtbase_git.bb > index 5cfbcfa..a2e8917 100644 > --- a/recipes-qt/qt5/qtbase_git.bb > +++ b/recipes-qt/qt5/qtbase_git.bb > @@ -137,7 +137,7 @@ QT_CONFIG_FLAGS += " \ > > # for qtbase configuration we need default settings > # since we cannot set empty set filename to a not existent file > -export OE_QMAKE_QTCONF_PATH = "foodummy" > +deltask generate_qt_config_file > > do_configure() { > # Avoid qmake error "Cannot read > [...]/usr/lib/qt5/mkspecs/oe-device-extra.pri: No such file or directory" > during configuration > -- > 2.12.0 > Any comments? -- ___ Openembedded-devel mailing list Openembedded-devel@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-devel
[oe] [meta-qt5][PATCH] qtbase: fix Krogoth build regression from efa8aaf
From: Matt Hoosier <matt.hoos...@garmin.com> With commit efa8aaf82e580a7d32eaaab48eb92d436f2e222a Author: Andreas Müller <schnitzelt...@googlemail.com> Date: Thu Feb 9 00:26:09 2017 qmake5_base.bbclass: set qt.conf by environment variable again we stopped pointing ${OE_QMAKE_QTCONF_PATH} at a valid file and instead directed it to a path which was intended not to exist. The motivation was to permit qtbase/qtbase-native/nativesdk-qtbase to build again after Qt 5.8 started paying attention to the contents of this file. The change as done in efa8aaf works well enough for Morty and subsequent releases' copies of Bitbake, but fails on earlier releases because they lack the following change: commit 2afcbfef2cd1ca568e5225884a8021df38ee3db0 Author: Ross Burton <ross.bur...@intel.com> Date: 2016-07-14 13:56:22 bitbake: build: don't use $B as the default cwd for functions The result is that when we build with Krogoth or prior, the body of do_generate_qt_config_file() runs with a cwd of ${B}, which was _not_ the intent of efa8aaf. Because the working directory is ${B}, ${OE_QMAKE_QTCONF_PATH} is written in there too. do_configure() -- whose cwd is by design also ${B} -- then finds the file 'foodummy', and the build breaks for the reasons outlined in efa8aaf. This change simply shifts the implementation tactics to suppress the creation of ${OE_QMAKE_QTCONF_PATH} during qtbase rather than rely on unspecified behavior about the cwd of do_generate_qt_config_file(). (cherry picked from commit a17ff281aa8d99f770b0a049cb798235005fb93e at https://codereview.qt-project.org/yocto/meta-qt5.) --- recipes-qt/qt5/nativesdk-qtbase_git.bb | 2 +- recipes-qt/qt5/qtbase-native_git.bb| 2 +- recipes-qt/qt5/qtbase_git.bb | 2 +- 3 files changed, 3 insertions(+), 3 deletions(-) diff --git a/recipes-qt/qt5/nativesdk-qtbase_git.bb b/recipes-qt/qt5/nativesdk-qtbase_git.bb index f2ff5f7..ffbb917 100644 --- a/recipes-qt/qt5/nativesdk-qtbase_git.bb +++ b/recipes-qt/qt5/nativesdk-qtbase_git.bb @@ -94,7 +94,7 @@ OE_QMAKE_PATH_HOST_LIBS = "${libdir}" # for qtbase configuration we need default settings # since we cannot set empty set filename to a not existent file -export OE_QMAKE_QTCONF_PATH = "foodummy" +deltask generate_qt_config_file do_configure() { ${S}/configure -v \ diff --git a/recipes-qt/qt5/qtbase-native_git.bb b/recipes-qt/qt5/qtbase-native_git.bb index 98cbecc..0884a19 100644 --- a/recipes-qt/qt5/qtbase-native_git.bb +++ b/recipes-qt/qt5/qtbase-native_git.bb @@ -79,7 +79,7 @@ PACKAGECONFIG_CONFARGS = " \ # for qtbase configuration we need default settings # since we cannot set empty set filename to a not existent file -export OE_QMAKE_QTCONF_PATH = "foodummy" +deltask generate_qt_config_file do_configure_prepend() { # Avoid qmake error "Cannot read [...]/usr/lib/qt5/mkspecs/oe-device-extra.pri: No such file or directory" diff --git a/recipes-qt/qt5/qtbase_git.bb b/recipes-qt/qt5/qtbase_git.bb index 5cfbcfa..a2e8917 100644 --- a/recipes-qt/qt5/qtbase_git.bb +++ b/recipes-qt/qt5/qtbase_git.bb @@ -137,7 +137,7 @@ QT_CONFIG_FLAGS += " \ # for qtbase configuration we need default settings # since we cannot set empty set filename to a not existent file -export OE_QMAKE_QTCONF_PATH = "foodummy" +deltask generate_qt_config_file do_configure() { # Avoid qmake error "Cannot read [...]/usr/lib/qt5/mkspecs/oe-device-extra.pri: No such file or directory" during configuration -- 2.12.0 -- ___ Openembedded-devel mailing list Openembedded-devel@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-devel
[oe] [meta-qt5] Pull request for 'qtbase: fix Krogoth build regression'
https://github.com/meta-qt5/meta-qt5/pull/62 -- ___ Openembedded-devel mailing list Openembedded-devel@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-devel
Re: [oe] [meta-qt5] Minimum required Yocto version for 5.9 branch?
This is addressed with pull request #62: From: Matt Hoosier <matt.hoos...@garmin.com> > > Date: Mon, 14 Aug 2017 08:04:03 -0500 > > Subject: [PATCH 1/2] qtbase: fix Krogoth build regression from efa8aaf > > >> With > > >> commit efa8aaf82e580a7d32eaaab48eb92d436f2e222a > > Author: Andreas Müller <schnitzelt...@googlemail.com> > > Date: Thu Feb 9 00:26:09 2017 > > >> qmake5_base.bbclass: set qt.conf by environment variable again > > >> we stopped pointing ${OE_QMAKE_QTCONF_PATH} at a valid file and > > instead directed it to a path which was intended not to exist. > > >> The motivation was to permit qtbase/qtbase-native/nativesdk-qtbase > > to build again after Qt 5.8 started paying attention to the > > contents of this file. > > >> The change as done in efa8aaf works well enough for Morty and > > subsequent releases' copies of Bitbake, but fails on earlier > > releases because they lack the following change: > > >> commit 2afcbfef2cd1ca568e5225884a8021df38ee3db0 > > Author: Ross Burton <ross.bur...@intel.com> > > Date: 2016-07-14 13:56:22 > > >> bitbake: build: don't use $B as the default cwd for functions > > >> The result is that when we build with Krogoth or prior, the body > > of do_generate_qt_config_file() runs with a cwd of ${B}, which was > > _not_ the intent of efa8aaf. Because the working directory is ${B}, > > ${OE_QMAKE_QTCONF_PATH} is written in there too. do_configure() -- > > whose cwd is by design also ${B} -- then finds the file 'foodummy', > > and the build breaks for the reasons outlined in efa8aaf. > > >> This change simply shifts the implementation tactics to suppress > > the creation of ${OE_QMAKE_QTCONF_PATH} during qtbase rather than rely > > on unspecified behavior about the cwd of do_generate_qt_config_file(). > > > This is a backport from the Qt fork of meta-qt5. On Fri, Aug 11, 2017 at 5:16 PM, Matt Hoosier <matt.hoos...@gmail.com> wrote: > On Fri, Aug 11, 2017 at 3:11 PM, Matt Hoosier <matt.hoos...@gmail.com> > wrote: > >> Hi, >> >> I observe that qtbase-native from jansa/master-5.9 fails on releases >> versions of Yocto prior to 2.2 with a message like: >> >> | + /home/hoosier/projects/consumer/tools/yocto/build/tmp/work/ >>> x86_64-linux/qtbase-native/5.9.1+gitAUTOINC+160533328c-r0/build/bin/qmake >>> /home/hoosier/projects/consumer/tools/yocto/build/tmp/work/ >>> x86_64-linux/qtbase-native/5.9.1+gitAUTOINC+160533328c-r0/git -- >>> -opensource -confirm-license -sysroot /home/hoosier/projects/consume >>> r/tools/yocto/build/tmp/sysroots/x86_64-linux -no-gcc-sysroot >>> -system-zlib -no-libjpeg -no-libpng -no-gif -no-accessibility -no-cups >>> -no-gui -no-qml-debug -no-sql-mysql -no-sql-sqlite -no-opengl -no-openssl >>> -no-xcb -no-icu -verbose -release -prefix /home/hoosier/projects/consume >>> r/tools/yocto/build/tmp/sysroots/x86_64-linux/usr -hostprefix >>> /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr >>> -bindir /home/hoosier/projects/consumer/tools/yocto/build/tmp/ >>> sysroots/x86_64-linux/usr/bin/qt5 -hostbindir >>> /home/hoosier/projects/consumer/tools/yocto/build/tmp/ >>> sysroots/x86_64-linux/usr/bin/qt5 -libdir /home/hoosier/projects/consume >>> r/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/lib -hostlibdir >>> /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/lib >>> -headerdir /home/hoosier/projects/consumer/tools/yocto/build/tmp/ >>> sysroots/x86_64-linux/usr/include/qt5 -archdatadir >>> /home/hoosier/projects/consumer/tools/yocto/build/tmp/ >>> sysroots/x86_64-linux/usr/lib/qt5 -datadir >>> /home/hoosier/projects/consumer/tools/yocto/build/tmp/ >>> sysroots/x86_64-linux/usr/share/qt5 -hostdatadir >>> /home/hoosier/projects/consumer/tools/yocto/build/tmp/ >>> sysroots/x86_64-linux/usr/lib/qt5 -docdir /home/hoosier/projects/consume >>> r/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/share/doc/qt5 >>> -sysconfdir >>> /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/etc/qt5 >>> -no-glib -no-iconv -silent -nomake examples -nomake tests -no-rpath >>> -platform linux-oe-g++ >>> | Command line: -opensource -confirm-license -sysroot >>> /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux >>> -no-gcc-sysroot -system-zlib -no-libjpeg -no-libpng -no-gif >>> -no-accessibility -no-cups -no-gui -no-qml-debug -no-sql-mys
Re: [oe] [meta-qt5] Minimum required Yocto version for 5.9 branch?
On Fri, Aug 11, 2017 at 3:11 PM, Matt Hoosier <matt.hoos...@gmail.com> wrote: > Hi, > > I observe that qtbase-native from jansa/master-5.9 fails on releases > versions of Yocto prior to 2.2 with a message like: > > | + /home/hoosier/projects/consumer/tools/yocto/build/ >> tmp/work/x86_64-linux/qtbase-native/5.9.1+gitAUTOINC+160533328c-r0/build/bin/qmake >> /home/hoosier/projects/consumer/tools/yocto/build/ >> tmp/work/x86_64-linux/qtbase-native/5.9.1+gitAUTOINC+160533328c-r0/git >> -- -opensource -confirm-license -sysroot /home/hoosier/projects/ >> consumer/tools/yocto/build/tmp/sysroots/x86_64-linux -no-gcc-sysroot >> -system-zlib -no-libjpeg -no-libpng -no-gif -no-accessibility -no-cups >> -no-gui -no-qml-debug -no-sql-mysql -no-sql-sqlite -no-opengl -no-openssl >> -no-xcb -no-icu -verbose -release -prefix /home/hoosier/projects/ >> consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr -hostprefix >> /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr >> -bindir /home/hoosier/projects/consumer/tools/yocto/build/ >> tmp/sysroots/x86_64-linux/usr/bin/qt5 -hostbindir /home/hoosier/projects/ >> consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/bin/qt5 -libdir >> /home/hoosier/projects/consumer/tools/yocto/build/ >> tmp/sysroots/x86_64-linux/usr/lib -hostlibdir /home/hoosier/projects/ >> consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/lib -headerdir >> /home/hoosier/projects/consumer/tools/yocto/build/ >> tmp/sysroots/x86_64-linux/usr/include/qt5 -archdatadir >> /home/hoosier/projects/consumer/tools/yocto/build/ >> tmp/sysroots/x86_64-linux/usr/lib/qt5 -datadir /home/hoosier/projects/ >> consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/share/qt5 >> -hostdatadir /home/hoosier/projects/consumer/tools/yocto/build/ >> tmp/sysroots/x86_64-linux/usr/lib/qt5 -docdir /home/hoosier/projects/ >> consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/share/doc/qt5 >> -sysconfdir /home/hoosier/projects/consumer/tools/yocto/build/ >> tmp/sysroots/x86_64-linux/etc/qt5 -no-glib -no-iconv -silent -nomake >> examples -nomake tests -no-rpath -platform linux-oe-g++ >> | Command line: -opensource -confirm-license -sysroot >> /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux >> -no-gcc-sysroot -system-zlib -no-libjpeg -no-libpng -no-gif >> -no-accessibility -no-cups -no-gui -no-qml-debug -no-sql-mysql >> -no-sql-sqlite -no-opengl -no-openssl -no-xcb -no-icu -verbose -release >> -prefix >> /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr >> -hostprefix >> /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr >> -bindir /home/hoosier/projects/consumer/tools/yocto/build/ >> tmp/sysroots/x86_64-linux/usr/bin/qt5 -hostbindir /home/hoosier/projects/ >> consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/bin/qt5 -libdir >> /home/hoosier/projects/consumer/tools/yocto/build/ >> tmp/sysroots/x86_64-linux/usr/lib -hostlibdir /home/hoosier/projects/ >> consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/lib -headerdir >> /home/hoosier/projects/consumer/tools/yocto/build/ >> tmp/sysroots/x86_64-linux/usr/include/qt5 -archdatadir >> /home/hoosier/projects/consumer/tools/yocto/build/ >> tmp/sysroots/x86_64-linux/usr/lib/qt5 -datadir /home/hoosier/projects/ >> consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/share/qt5 >> -hostdatadir /home/hoosier/projects/consumer/tools/yocto/build/ >> tmp/sysroots/x86_64-linux/usr/lib/qt5 -docdir /home/hoosier/projects/ >> consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/share/doc/qt5 >> -sysconfdir /home/hoosier/projects/consumer/tools/yocto/build/ >> tmp/sysroots/x86_64-linux/etc/qt5 -no-glib -no-iconv -silent -nomake >> examples -nomake tests -no-rpath -platform linux-oe-g++ >> | Cannot read /home/hoosier/projects/consumer/tools/yocto/build/ >> tmp/sysroots/x86_64-linux/usr/lib/qt5/mkspecs/linux-oe-g++/qmake.conf: >> No such file or directory > > > Does anybody know what might be going on with this error? > > After a bit of digging, I find that the following commit in poky is the one that's required to let the 5.9 branch of meta-qt5 build correctly with Yocto <= 2.1: Author: Ross Burton <ross.bur...@intel.com> 2016-07-14 13:56:22 > Committer: Richard Purdie <richard.pur...@linuxfoundation.org> > 2016-07-19 02:56:50 > Parent: faa726824dfb5bc0141919bb16c0d0069a3da69e (bitbake: > lib/bb/build.py: decode the command as UTF-8) > Branches: remotes/origin/morty, remotes/origin/pyro > Follows: 2.2_M1 > Precedes: 2.2_M2 > bitba
[oe] [meta-qt5] Minimum required Yocto version for 5.9 branch?
Hi, I observe that qtbase-native from jansa/master-5.9 fails on releases versions of Yocto prior to 2.2 with a message like: | + > /home/hoosier/projects/consumer/tools/yocto/build/tmp/work/x86_64-linux/qtbase-native/5.9.1+gitAUTOINC+160533328c-r0/build/bin/qmake > /home/hoosier/projects/consumer/tools/yocto/build/tmp/work/x86_64-linux/qtbase-native/5.9.1+gitAUTOINC+160533328c-r0/git > -- -opensource -confirm-license -sysroot > /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux > -no-gcc-sysroot -system-zlib -no-libjpeg -no-libpng -no-gif > -no-accessibility -no-cups -no-gui -no-qml-debug -no-sql-mysql > -no-sql-sqlite -no-opengl -no-openssl -no-xcb -no-icu -verbose -release > -prefix > /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr > -hostprefix > /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr > -bindir > /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/bin/qt5 > -hostbindir > /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/bin/qt5 > -libdir > /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/lib > -hostlibdir > /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/lib > -headerdir > /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/include/qt5 > -archdatadir > /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/lib/qt5 > -datadir > /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/share/qt5 > -hostdatadir > /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/lib/qt5 > -docdir > /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/share/doc/qt5 > -sysconfdir > /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/etc/qt5 > -no-glib -no-iconv -silent -nomake examples -nomake tests -no-rpath > -platform linux-oe-g++ > | Command line: -opensource -confirm-license -sysroot > /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux > -no-gcc-sysroot -system-zlib -no-libjpeg -no-libpng -no-gif > -no-accessibility -no-cups -no-gui -no-qml-debug -no-sql-mysql > -no-sql-sqlite -no-opengl -no-openssl -no-xcb -no-icu -verbose -release > -prefix > /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr > -hostprefix > /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr > -bindir > /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/bin/qt5 > -hostbindir > /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/bin/qt5 > -libdir > /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/lib > -hostlibdir > /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/lib > -headerdir > /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/include/qt5 > -archdatadir > /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/lib/qt5 > -datadir > /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/share/qt5 > -hostdatadir > /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/lib/qt5 > -docdir > /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/share/doc/qt5 > -sysconfdir > /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/etc/qt5 > -no-glib -no-iconv -silent -nomake examples -nomake tests -no-rpath > -platform linux-oe-g++ > | Cannot read > /home/hoosier/projects/consumer/tools/yocto/build/tmp/sysroots/x86_64-linux/usr/lib/qt5/mkspecs/linux-oe-g++/qmake.conf: > No such file or directory Does anybody know what might be going on with this error? -- ___ Openembedded-devel mailing list Openembedded-devel@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-devel
[oe] [meta-qt5] [meta-mingw] Any success using these two layers together?
Does anybody have experience using meta-qt5 and meta-ming together to produce a usable SDK including Qt development tools for use on a Windows host together? The stock recipes for producing the nativesdk flavors of certain Qt packages would seem to require some modification from vanilla meta-qt5 upstream to be able to target mingw. Cheers, Matt -- ___ Openembedded-devel mailing list Openembedded-devel@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-devel