Re: [OE-core] [meta-java] icedtea7-native build fails on kirkstone and in master

[Bartosz Golaszewski <brgl@bgdev.pl>]

On Thu, Jun 9, 2022 at 12:14 PM Matt Madison <matt@madison.systems> wrote:
>
> (Dropping oe-core and leaving openembedded-devel, as that's the right
> list for this)
>
> On Thu, Jun 9, 2022 at 2:28 AM Bartosz Golaszewski <brgl@bgdev.pl> wrote:
> >
> > Hi!
> >
> > I'm trying to build java support with the following configuration:
> >
> > Build Configuration:
> > BB_VERSION           = "2.0.0"
> > BUILD_SYS            = "x86_64-linux"
> > NATIVELSBSTRING      = "universal"
> > TARGET_SYS           = "x86_64-poky-linux"
> > MACHINE              = "qemux86-64"
> > DISTRO               = "poky"
> > DISTRO_VERSION       = "4.0.1"
> > TUNE_FEATURES        = "m64 core2"
> > TARGET_FPU           = ""
> > meta
> > meta-poky            = "kirkstone:45d7615dfef8093a2467d52ca1130e62db6a1187"
> > meta-oe              = "kirkstone:fcc7d7eae82be4c180f2e8fa3db90a8ab3be07b7"
> > meta-java            = "kirkstone:1a8059f6b257ebe6fcae6416e499784d976afd24"
> >
> > ---
> >
> > PREFERRED_PROVIDER_virtual/java-initial-native = "cacao-initial-native"
> > PREFERRED_PROVIDER_virtual/java-native = "jamvm-native"
> > PREFERRED_PROVIDER_virtual/javac-native = "ecj-bootstrap-native"
> >
> > ---
> >
> > The build fails for icedtea7-native but it's hard to tell why exactly
> > from the logs. Here is the full output: https://pastebin.com/3dRt6J0Y
>
> I ran into this recently, too, after upgrading Ubuntu to 22.04.  Near
> as I can tell, there's a bug in the dynamic linker in glibc 2.35 that
> causes a segfault when it's invoked through ldd in some cases. One of
> my colleagues that uses Manjaro for his development host also saw
> this, and it, too, had glibc 2.35, so I think that's the common
> factor.
>
> I worked around the problem by disabling the ldd checks with
>
> EXTRA_OEMAKE:append = " LDD=:"
>
> in a bbappend.  This isn't specific to kirkstone, btw; my builds have
> been dunfell-based.
>
> The ldd checks in the makefiles don't seem particularly useful except
> to the JDK developers, so turning them off didn't seem too terrible.
>
> I see at least one bug open in the glibc bugzilla for something that
> looks similar to this problem, but apparently not easily reproducible
> with a small test case. I haven't had time to look into it further.
>

Thanks! It does indeed help with this specific issue.

Unfortunately, it still fails in a different way for me:

| make[5]: Leaving directory
'/home/brgl/workspace/target-build-yocto/build/tmp/work/x86_64-linux/icedtea7-native/2.1.3-r1.0/icedtea-2.1.3/build/openjdk-boot/jdk/make/sun/awt'
| make[5]: Entering directory
'/home/brgl/workspace/target-build-yocto/build/tmp/work/x86_64-linux/icedtea7-native/2.1.3-r1.0/icedtea-2.1.3/build/openjdk-boot/jdk/make/sun/xawt'
| /bin/sh: 1: free: not found
| /bin/mkdir -p
/home/brgl/workspace/target-build-yocto/build/tmp/work/x86_64-linux/icedtea7-native/2.1.3-r1.0/icedtea-2.1.3/build/openjdk.build-boot/tmp/sun/sun.awt.X11/xawt/sun/awt/X11/generator
| rm -f /home/brgl/workspace/target-build-yocto/build/tmp/work/x86_64-linux/icedtea7-native/2.1.3-r1.0/icedtea-2.1.3/build/openjdk.build-boot/tmp/sun/sun.awt.X11/xawt/sun/awt/X11/generator/WrapperGenerator.class
| /home/brgl/workspace/target-build-yocto/build/tmp/work/x86_64-linux/icedtea7-native/2.1.3-r1.0/icedtea-2.1.3/build/bootstrap/jdk1.6.0/bin/javac
-J-XX:ThreadStackSize=1536 -J-XX:-PrintVMOptions
-J-XX:+UnlockDiagnosticVMOptions -J-XX:-LogVMOutput -J-Xmx384m
-Xlint:-path -encoding ascii -source 6 -target 6
-XDignore.symbol.file=true -bootclasspath
/home/brgl/workspace/target-build-yocto/build/tmp/work/x86_64-linux/icedtea7-native/2.1.3-r1.0/icedtea-2.1.3/build/bootstrap/jdk1.6.0/jre/lib/rt.jar
-d /home/brgl/workspace/target-build-yocto/build/tmp/work/x86_64-linux/icedtea7-native/2.1.3-r1.0/icedtea-2.1.3/build/openjdk.build-boot/tmp/sun/sun.awt.X11/xawt/sun/awt/X11/generator
../../../src/solaris/classes/sun/awt/X11/generator/WrapperGenerator.java
| Annotation processing got disabled, since it requires a 1.6 compliant JVM
| ----------
| 1. ERROR in ../../../src/solaris/classes/sun/awt/X11/generator/WrapperGenerator.java
(at line 0)
| /*
| ^
| Inconsistent classfile encountered: The undefined type parameter E
is referenced from within Collection
| ----------
| 1 problem (1 error)make[5]: Leaving directory
'/home/brgl/workspace/target-build-yocto/build/tmp/work/x86_64-linux/icedtea7-native/2.1.3-r1.0/icedtea-2.1.3/build/openjdk-boot/jdk/make/sun/xawt'
| make[5]: *** [Makefile:280:
/home/brgl/workspace/target-build-yocto/build/tmp/work/x86_64-linux/icedtea7-native/2.1.3-r1.0/icedtea-2.1.3/build/openjdk.build-boot/tmp/sun/sun.awt.X11/xawt/sun/awt/X11/generator/WrapperGenerator.class]
Error 255
| make[4]: Leaving directory
'/home/brgl/workspace/target-build-yocto/build/tmp/work/x86_64-linux/icedtea7-native/2.1.3-r1.0/icedtea-2.1.3/build/openjdk-boot/jdk/make/sun'
| make[4]: *** [Makefile:94: all] Error 1
| make[3]: Leaving directory
'/home/brgl/workspace/target-build-yocto/build/tmp/work/x86_64-linux/icedtea7-native/2.1.3-r1.0/icedtea-2.1.3/build/openjdk-boot/jdk/make'
| make[3]: *** [Makefile:247: all] Error 1
| make[2]: Leaving directory
'/home/brgl/workspace/target-build-yocto/build/tmp/work/x86_64-linux/icedtea7-native/2.1.3-r1.0/icedtea-2.1.3/build/openjdk-boot'
| make[2]: *** [make/jdk-rules.gmk:79: jdk-build] Error 2
| make[1]: Leaving directory
'/home/brgl/workspace/target-build-yocto/build/tmp/work/x86_64-linux/icedtea7-native/2.1.3-r1.0/icedtea-2.1.3/build/openjdk-boot'
| make[1]: *** [Makefile:244: build_product_image] Error 2
| make: *** [Makefile:2252: stamps/icedtea-boot.stamp] Error 2
| ERROR: oe_runmake failed
| WARNING: exit code 1 from a shell command.

Bart