[HEADS-UP] OpenJDK 11 is now the default Java in rawhide

[Fabio Valentini]

(Sending this heads-up in a new thread for better visibility.)

Yesterday the f33-java11 side tag was merged into rawhide, which
brings the necessary changes to make java-11-openjdk the default Java
in fedora. All packages depending on Java should also have been
rebuilt in this side tag.

The "new" Java SIG (@java-maint-sig) is still dealing with a pretty
small amount of fallout from this change. For example, some packages
that are stuck on Java 8 for various reasons are failing to compile or
run because their dependencies have been built with bytecode that is
too new for OpenJDK 8.

We're tracking the packages that are stuck with OpenJDK 8 here:
https://pagure.io/java-maint-sig/issue/10

There's also a complete bytecode version analysis of all Java packages
in fedora in progress to identify more possible issues:
https://pagure.io/java-maint-sig/issue/7

I have tried to deal with the pretty long list of "easy fixes" for
Java 11 compatibility, having fixed over 50 FTBFS issues over the past
few days alone. The remaining "EasyFix" candidates are tracked here,
though we'll probably close this ticket soon.
https://pagure.io/java-maint-sig/issue/1

If you see any issues with Java 11 as default in rawhide (the change
is live in rawhide and will be part of the next successful compose,
possibly today), feel free to comment on the linked tickets if you
have similar issues, or open a new ticket with @java-maint-sig on
pagure if you need help.

Fabio

[Adam Williamson]

On Tue, 2020-07-21 at 12:02 +0200, Fabio Valentini wrote:
> (Sending this heads-up in a new thread for better visibility.)
>
> Yesterday the f33-java11 side tag was merged into rawhide, which
> brings the necessary changes to make java-11-openjdk the default Java
> in fedora. All packages depending on Java should also have been
> rebuilt in this side tag.
>
> The "new" Java SIG (@java-maint-sig) is still dealing with a pretty
> small amount of fallout from this change. For example, some packages
> that are stuck on Java 8 for various reasons are failing to compile or
> run because their dependencies have been built with bytecode that is
> too new for OpenJDK 8.

Well. I don't think "pretty small" is really accurate. libreoffice
failed build in the side tag:

https://koji.fedoraproject.org/koji/buildinfo?buildID=1540582

and, unsurprisingly, has again failed to rebuild since the side tag was
merged:

https://koji.fedoraproject.org/koji/buildinfo?buildID=1541384

libreoffice is part of our release-blocking package set. So now Rawhide
composes are failing because a release blocking image cannot be built
because libreoffice cannot be installed.

Merging the side tag with FreeIPA known broken was arguable enough, but
merging it with a release-blocking package broken seems worse. Please
try not to do this in future.

[Deepak Bhole]

* Adam Williamson  [2020-07-23 20:57]:
> On Tue, 2020-07-21 at 12:02 +0200, Fabio Valentini wrote:
> > (Sending this heads-up in a new thread for better visibility.)
> >
> > Yesterday the f33-java11 side tag was merged into rawhide, which
> > brings the necessary changes to make java-11-openjdk the default Java
> > in fedora. All packages depending on Java should also have been
> > rebuilt in this side tag.
> >
> > The "new" Java SIG (@java-maint-sig) is still dealing with a pretty
> > small amount of fallout from this change. For example, some packages
> > that are stuck on Java 8 for various reasons are failing to compile or
> > run because their dependencies have been built with bytecode that is
> > too new for OpenJDK 8.
>
> Well. I don't think "pretty small" is really accurate. libreoffice
> failed build in the side tag:
>
> https://koji.fedoraproject.org/koji/buildinfo?buildID=1540582
>
> and, unsurprisingly, has again failed to rebuild since the side tag was
> merged:
>
> https://koji.fedoraproject.org/koji/buildinfo?buildID=1541384
>
> libreoffice is part of our release-blocking package set. So now Rawhide
> composes are failing because a release blocking image cannot be built
> because libreoffice cannot be installed.

Not trying to downplay your concern about FreeIPA or other components but
neither of the errors above appear related to the JDK change. Both are errors
when building vcldemo.o which is being built with g++ and segfaulting during
build:

---

[ ... ]

[build CXX] vcl/workben/vcldemo.cxx
S=/builddir/build/BUILD/libreoffice-6.4.5.2 && I=$S/instdir && W=$S/workdir &&  mkdir -p $W/CxxObject/vcl/workben/ $W/Dep/CxxObject/vcl/workben/ && cd /builddir/build/BUILD/libreoffice-6.4.5.2 &&    g++ -DBOOST_ERROR_CODE_HEADER_ONLY -DBOOST_SYSTEM_NO_DEPRECATED -DCPPU_ENV=gcc3 -DLINUX -DNDEBUG -DOSL_DEBUG_LEVEL=0 -DUNIX -DUNX -DX86_64 -D_PTHREADS -D_REENTRANT   -fvisibility=hidden    -Wall -Wno-missing-braces -Wnon-virtual-dtor -Wendif-labels -Wextra -Wundef -Wunreachable-code -Wunused-macros -finput-charset=UTF-8 -fmessage-length=0 -fno-common -pipe -fstack-protector-strong  -Wduplicated-cond -Wlogical-op -Wshift-overflow=2 -Wunused-const-variable=1 -Wno-cast-function-type -fvisibility-inlines-hidden -fPIC -Wshadow -Woverloaded-virtual -std=c++17 -pthread  -O2 -flto -ffat-lto-objects -g -grecord-gcc-switches -Wp,-D_FORTIFY_SOURCE=2 -Wp,-D_GLIBCXX_ASSERTIONS -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -fstack-protector-strong -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -m64 -mtune=generic -fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection  -DEXCEPTIONS_ON -fexceptions -fno-enforce-eh-specs    -DLIBO_INTERNAL_ONLY  -c $S/vcl/workben/vcldemo.cxx -o $W/CxxObject/vcl/workben/vcldemo.o  -I$S/include  -I/usr/lib/jvm/java-11-openjdk-11.0.8.10-0.fc33.x86_64/include -I/usr/lib/jvm/java-11-openjdk-11.0.8.10-0.fc33.x86_64/include/linux -I$S/config_host  -I$W/UnoApiHeadersTarget/offapi/normal -I$W/UnoApiHeadersTarget/udkapi/normal -I/usr/include   -isystem /usr/include/harfbuzz -isystem /usr/include/glib-2.0 -isystem /usr/lib64/glib-2.0/include -isystem /usr/include/freetype2 -isystem /usr/include/libpng16      -I$S/vcl/inc

[ ... ]

/builddir/build/BUILD/libreoffice-6.4.5.2/vcl/workben/vcldemo.cxx: In destructor 'virtual DemoWin::RenderThread::~RenderThread()':
/builddir/build/BUILD/libreoffice-6.4.5.2/vcl/workben/vcldemo.cxx:1733:18: internal compiler error: Segmentation fault
 1733 |             join();

[ ... ]

make[1]: *** [/builddir/build/BUILD/libreoffice-6.4.5.2/solenv/gbuild/LinkTarget.mk:296: /builddir/build/BUILD/libreoffice-6.4.5.2/workdir/CxxObject/vcl/workben/vcldemo.o] Error 1
make[1]: *** Waiting for unfinished jobs....

[ ... ]

---

I could of-course misunderstanding the issue but is there a reason that you
believe this is JDK related?

Deepak
> Merging the side tag with FreeIPA known broken was arguable enough, but
> merging it with a release-blocking package broken seems worse. Please
> try not to do this in future.
> -- 
> Adam Williamson
> Fedora QA Community Monkey
> IRC: adamw | Twitter: AdamW_Fedora | XMPP: adamw AT happyassassin . net
> http://www.happyassassin.net
> _______________________________________________
> devel mailing list -- [e-mail]
> To unsubscribe send an email to [e-mail]
> Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/
> List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
> List Archives: https://lists.fedoraproject.org/archives/list/[e-mail]

[Adam Williamson]

On July 23, 2020 6:17:04 p.m. PDT, Deepak Bhole  wrote:
> Adam Williamson  [2020-07-23 20:57]:
> > On Tue, 2020-07-21 at 12:02 +0200, Fabio Valentini wrote:
> > > (Sending this heads-up in a new thread for better visibility.)
> > >
> > > Yesterday the f33-java11 side tag was merged into rawhide, which
> > > brings the necessary changes to make java-11-openjdk the default
>
> Java
> > > in fedora. All packages depending on Java should also have been
> > > rebuilt in this side tag.
> > >
> > > The "new" Java SIG (@java-maint-sig) is still dealing with a pretty
> > > small amount of fallout from this change. For example, some
>
> packages
> > > that are stuck on Java 8 for various reasons are failing to compile
>
> or
> > > run because their dependencies have been built with bytecode that
>
> is
> > > too new for OpenJDK 8.
> >
> > Well. I don't think "pretty small" is really accurate. libreoffice
> > failed build in the side tag:
> >
> > https://koji.fedoraproject.org/koji/buildinfo?buildID=1540582
> >
> > and, unsurprisingly, has again failed to rebuild since the side tag
>
> was
> > merged:
> >
> > https://koji.fedoraproject.org/koji/buildinfo?buildID=1541384
> >
> > libreoffice is part of our release-blocking package set. So now
>
> Rawhide
> > composes are failing because a release blocking image cannot be built
> > because libreoffice cannot be installed.
>
> Not trying to downplay your concern about FreeIPA or other components
> but
> neither of the errors above appear related to the JDK change. Both are
> errors
> when building vcldemo.o which is being built with g++ and segfaulting
> during
> build:
>
> [ ... ]
>
> [build CXX] vcl/workben/vcldemo.cxx
> S=/builddir/build/BUILD/libreoffice-6.4.5.2 && I=$S/instdir &&
> W=$S/workdir &&  mkdir -p $W/CxxObject/vcl/workben/
> $W/Dep/CxxObject/vcl/workben/ && cd
> /builddir/build/BUILD/libreoffice-6.4.5.2 &&    g++
> -DBOOST_ERROR_CODE_HEADER_ONLY -DBOOST_SYSTEM_NO_DEPRECATED
> -DCPPU_ENV=gcc3 -DLINUX -DNDEBUG -DOSL_DEBUG_LEVEL=0 -DUNIX -DUNX
> -DX86_64 -D_PTHREADS -D_REENTRANT   -fvisibility=hidden    -Wall
> -Wno-missing-braces -Wnon-virtual-dtor -Wendif-labels -Wextra -Wundef
> -Wunreachable-code -Wunused-macros -finput-charset=UTF-8
> -fmessage-length=0 -fno-common -pipe -fstack-protector-strong 
> -Wduplicated-cond -Wlogical-op -Wshift-overflow=2
> -Wunused-const-variable=1 -Wno-cast-function-type
> -fvisibility-inlines-hidden -fPIC -Wshadow -Woverloaded-virtual
> -std=c++17 -pthread  -O2 -flto -ffat-lto-objects -g
> -grecord-gcc-switches -Wp,-D_FORTIFY_SOURCE=2 -Wp,-D_GLIBCXX_ASSERTIONS
> -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -fstack-protector-strong
> -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -m64 -mtune=generic
> -fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection 
> -DEXCEPTIONS_ON -fexceptions -fno-enforce-eh-specs   
> -DLIBO_INTERNAL_ONLY  -c $S/vcl/workben/vcldemo.cxx -o
> $W/CxxObject/vcl/workben/vcldemo.o  -I$S/include 
> -I/usr/lib/jvm/java-11-openjdk-11.0.8.10-0.fc33.x86_64/include
> -I/usr/lib/jvm/java-11-openjdk-11.0.8.10-0.fc33.x86_64/include/linux
> -I$S/config_host  -I$W/UnoApiHeadersTarget/offapi/normal
> -I$W/UnoApiHeadersTarget/udkapi/normal -I/usr/include   -isystem
> /usr/include/harfbuzz -isystem /usr/include/glib-2.0 -isystem
> /usr/lib64/glib-2.0/include -isystem /usr/include/freetype2 -isystem
> /usr/include/libpng16      -I$S/vcl/inc
>
> [ ... ]
>
> /builddir/build/BUILD/libreoffice-6.4.5.2/vcl/workben/vcldemo.cxx: In
> destructor 'virtual DemoWin::RenderThread::~RenderThread()':
> /builddir/build/BUILD/libreoffice-6.4.5.2/vcl/workben/vcldemo.cxx:1733:18:
> internal compiler error: Segmentation fault
> 1733 |             join();
>
> [ ... ]
>
> make[1]: ***
> [/builddir/build/BUILD/libreoffice-6.4.5.2/solenv/gbuild/LinkTarget.mk:296:
> /builddir/build/BUILD/libreoffice-6.4.5.2/workdir/CxxObject/vcl/workben/vcldemo.o]
> Error 1
> make[1]: *** Waiting for unfinished jobs....
>
> [ ... ]
>
> I could of-course misunderstanding the issue but is there a reason that
> you
> believe this is JDK related?
>
> Deepak
> > Merging the side tag with FreeIPA known broken was arguable enough,
>
> but
> > merging it with a release-blocking package broken seems worse. Please
> > try not to do this in future.
> > -- 
> > Adam Williamson
> > Fedora QA Community Monkey
> > IRC: adamw | Twitter: AdamW_Fedora | XMPP: adamw AT happyassassin .
>
> net
> > http://www.happyassassin.net
> > _______________________________________________
> > devel mailing list -- [e-mail]
> > To unsubscribe send an email to [e-mail]
> > Fedora Code of Conduct:
>
> https://docs.fedoraproject.org/en-US/project/code-of-conduct/
> > List Guidelines:
>
> https://fedoraproject.org/wiki/Mailing_list_guidelines
> > List Archives:
>
> https://lists.fedoraproject.org/archives/list/[e-mail]
> _______________________________________________
> devel mailing list -- [e-mail]
> To unsubscribe send an email to [e-mail]
> Fedora Code of Conduct:
> https://docs.fedoraproject.org/en-US/project/code-of-conduct/
> List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
> List Archives:
> https://lists.fedoraproject.org/archives/list/[e-mail]

The specific compile error isn't really relevant, the problem is that we need a new LO because the deps of the existing build cannot be satisfied. However, now I look closer, I think I was wrong to blame java for the deps changing, I apologize for that - the timing lined up, and there was an LO build failure in the java side tag like I mentioned, so I figured the new java was the issue based on that, but it actually seems to be coincidental. It looks like the dep that changed was actually poppler - a new build of poppler was done two weeks ago but I guess was only tagged to f33 recently.

[Deepak Bhole]

* Adam Williamson  [2020-07-24 00:52]:
> On July 23, 2020 6:17:04 p.m. PDT, Deepak Bhole  wrote:
> > Adam Williamson  [2020-07-23 20:57]:
> > > On Tue, 2020-07-21 at 12:02 +0200, Fabio Valentini wrote:

[ ... ]
> > I could of-course misunderstanding the issue but is there a reason that
> > you
> > believe this is JDK related?
> >
> > Deepak
> > > Merging the side tag with FreeIPA known broken was arguable enough,
> >
> > but
> > > merging it with a release-blocking package broken seems worse. Please
> > > try not to do this in future.

[ ... ]
> The specific compile error isn't really relevant, the problem is that we need a new LO because the deps of the existing build cannot be satisfied. However, now I look closer, I think I was wrong to blame java for the deps changing, I apologize for that - the timing lined up, and there was an LO build failure in the java side tag like I mentioned, so I figured the new java was the issue based on that, but it actually seems to be coincidental. It looks like the dep that changed was actually poppler - a new build of poppler was done two weeks ago but I guess was only tagged to f33 recently.

Ah okay, thanks for clarifying, I was unaware of the additional dep issue.

Deepak
> -- 
> Adam Williamson
> Fedora QA Community Monkey
> IRC: adamw | Twitter: AdamW_Fedora | XMPP: adamw AT happyassassin . net
> http://www.happyassassin.net