Hi Jiangli,
thanks for looking at the change.
'CDS_only' is only required for static fields because the
VMStructEntry for them contains a reference to the actual static field
which isn't present if we disable CDS, because the corresponding
compilations units (i.e. filemap.cpp) won't be part of li
Please review this small change that fixes the AIX build after "8203641:
Refactor String Deduplication into shared".
We are getting this compilation error :
/build_ci_jdk_jdk_rs6000_64/src/hotspot/share/gc/shared/stringdedup/stringDedup.hpp",
line 107.38: 1540-0063 (S) The text "1" is
On Thu, Jun 14, 2018 at 9:04 PM, Thomas Stüfe wrote:
> Hi Volker,
>
> http://cr.openjdk.java.net/~simonis/webrevs/2018/8204965/make/autoconf/hotspot.m4.udiff.html
>
> Seems like a roundabout way to have a platform specific default value.
>
> Why not determine a default value beforehand:
>
> if tes
Forwarding to build-dev.--- Begin Message ---
Just for the completeness, this is a normal openSUSE package build with
--with-system-jpeg and --with-system-lcms. Maybe the
LIBJPEG_HEADERS_FROM_SRC := false is culprit. Some other people were
able to build jdk-11+18 with bundled jpeg and lcms without
Hi Fridrich!
I can help you getting this issue fixed. If I remember correctly, you said
that SUSE has signed Oracle's OCA, so I it should be fine if I submit the
patch with you as the author.
Let me have a look at the problem first though.
Adrian
On 06/15/2018 10:27 AM, Severin Gehwolf wrote:
>
Hi Mattias,
the change looks good. Could you please also update the comment in the
line above which still reads "STAT".
Also, maybe "STATI" would be a better name choice (longer is better :)
because the probability of a clash is lower (and it would nicely align
with QUEUE in the comments :) But I
Hi , thanks for looking into it.
I think I will use STAT_IMPL .
Best regards, Matthias
> -Original Message-
> From: René Schünemann [mailto:[email protected]]
> Sent: Freitag, 15. Juni 2018 11:08
> To: Baesken, Matthias
> Cc: [email protected]; hotspot-...@openjdk.
Hi Matthias,
Good catch. Patch for me is good if you guys agree on a good uncommon name.
Gruß Thomas
On Fri, Jun 15, 2018, 09:48 Baesken, Matthias
wrote:
> Please review this small change that fixes the AIX build after
> "8203641: Refactor String Deduplication into shared".
>
> We are g
On 2018-06-15 10:57, John Paul Adrian Glaubitz wrote:
Hi Fridrich!
I can help you getting this issue fixed. If I remember correctly, you said
that SUSE has signed Oracle's OCA, so I it should be fine if I submit the
patch with you as the author.
Let me have a look at the problem first though.
On 06/15/2018 01:10 PM, Magnus Ihse Bursie wrote:
> I think I understand the underlying issue but I'm not sure how to try
> reproduce a build problem from it.
I have an openSUSE box here, so I should be able to reproduce it :).
> Here's a better patch. Please try and see if it solves your proble
Since b14 or b15, the client VM on x86 doesn't build anymore when building with
--with-jvm-variants=client,server --with-num-cores=4
building with one to three cores seems to work however. The server and zero VMs
build without issues and parallel builds. Any idea which dependencies got dropp
There was a race with the JFR build tools when building multiple JVMs.
:-( Erik produced a fix for this as part of JDK-8202384), unfortunately
this has not yet been pushed.
You can find the JFR fix part here:
http://cr.openjdk.java.net/~erikj/8202384/webrev.05/make/hotspot/gensrc/GensrcJfr.gm
On 2018-06-15 01:26, [email protected] wrote:
Locally tested the fix below and via mach5 jobs.
The ~50 failed tests all completed successfully.
It'd be good for Erik to repeat his full testing with
this patch included in case it blocked other issues
from being observed.
This fix can be integ
Hi,
This is a JDK 8u specific problem. It's not applicable to 10/11 since
the build system has changed. Make files in JDK 8 live in the hotspot
tree, hence, I'm also including hotspot-dev. The issue at hand is that
linker flags are not consistently passed down to individual library
builds. Specifi
(reviving an old review thread)
On 2018-05-25 18:17, Erik Joelsson wrote:
On 2018-05-25 06:52, Tim Bell wrote:
Magnus-
On 05/25/18 04:23, Claes Redestad wrote:
On 2018-05-25 12:06, Magnus Ihse Bursie wrote:
WebRev:
http://cr.openjdk.java.net/~ihse/JDK-8200867-remove-jdk9-references/webrev.0
On 2018-06-15 14:59, Severin Gehwolf wrote:
Hi,
This is a JDK 8u specific problem. It's not applicable to 10/11 since
the build system has changed. Make files in JDK 8 live in the hotspot
tree, hence, I'm also including hotspot-dev. The issue at hand is that
linker flags are not consistently pas
On 2018-06-12 17:26, Peter Budai wrote:
Hi,
You can find the patches here:
https://github.com/peterbud/MINGW-packages/tree/openjdk/mingw-w64-openjdk
I have managed to build OpenJDK on MSYS2 with gcc, both 32 and 64 bit,
but that was obviously the beginning. Around 10% of the test cases
were
Hi Magnus,
Yes, I have signed OCA last October, and Dalibor Topic has processed it.
I’m happy to work on the msys2 build part with you and Erik.
As far as I recall, the hotspot tests were really close, the number of failed
test cases seemed to be manageable:
==
Test
Hello Matthias,
I believe I know the problem. I fixed it in my patch currently under
review here (see GensrcJfr.gmk):
http://cr.openjdk.java.net/~erikj/8202384/webrev.05/index.html
That patch will take a while before it gets in though because it needs
to go through the JEP process first. Ple
Looks good.
/Erik
On 2018-06-15 05:59, Severin Gehwolf wrote:
Hi,
This is a JDK 8u specific problem. It's not applicable to 10/11 since
the build system has changed. Make files in JDK 8 live in the hotspot
tree, hence, I'm also including hotspot-dev. The issue at hand is that
linker flags are
Hi Matthias,
the name SI seems also quite "common" and may result in other naming
clashes in the future.
Maybe something more readable like STAT_IMPL?
Please also change the name in the comment:
// STAT: String Dedup Stat implementation
Regards,
Rene
On Fri, Jun 15, 2018 at 9:47 AM, Baesken,
Hi,
Whatever the name you come up, could you please also update
stringDedup.inline.hpp to use the same names?
Thomas Schatzl pointed out the inconsistent, I made change to use S and
Q, but apparently, I messed up in final patch.
Otherwise, looks good to me too.
-Zhengyu
On 06/15/2018 06
Hello, Magnus,
Your original patch had a problem, since the second statement overwrote
the $1_SRC_HEADER_FLAGS instead of appending to it. With that, things
like osSupport.hpp, jawt_md.h and so on could not be found in build of
tests. So, I modified your patch to read as attached and the build
fin
Hello,
I am trying to setup project panama on my local mac, and I am not able to
build the project.
Configuration summary:
* Debug level:release
* HS debug level: product
* JDK variant:normal
* JVM variants: server
* OpenJDK target: OS: macosx, CPU architecture: x86, address length: 64
Hi Volker,
> On Jun 15, 2018, at 12:43 AM, Volker Simonis wrote:
>
> Hi Jiangli,
>
> thanks for looking at the change.
>
> 'CDS_only' is only required for static fields because the
> VMStructEntry for them contains a reference to the actual static field
> which isn't present if we disable CDS,
Hi Volker,
On Fri, Jun 15, 2018 at 10:05 AM, Volker Simonis
wrote:
> On Thu, Jun 14, 2018 at 9:04 PM, Thomas Stüfe wrote:
>> Hi Volker,
>>
>> http://cr.openjdk.java.net/~simonis/webrevs/2018/8204965/make/autoconf/hotspot.m4.udiff.html
>>
>> Seems like a roundabout way to have a platform specif
Hello Sebastian,
Please try running configure with --disable-warnings-as-errors.
When using compiler versions (or OS versions) other than the ones listed
in doc/building.md it's common to have warnings triggered and we run
with warnings treated as errors by default.
/Erik
On 2018-06-11 12:
Due to a typo in spec.gmk.in, the platform specific include files have
lost their platform sub directory in the jdk image include dir.
Bug: https://bugs.openjdk.java.net/browse/JDK-8203667
Webrev: http://cr.openjdk.java.net/~erikj/8203667/webrev.01/
/Erik
+1
Mandy
On 6/15/18 11:12 AM, Erik Joelsson wrote:
Due to a typo in spec.gmk.in, the platform specific include files have
lost their platform sub directory in the jdk image include dir.
Bug: https://bugs.openjdk.java.net/browse/JDK-8203667
Webrev: http://cr.openjdk.java.net/~erikj/8203667/we
Erik:
Looks good to me as well.
/Tim
On 06/15/18 11:20, mandy chung wrote:
+1
Mandy
On 6/15/18 11:12 AM, Erik Joelsson wrote:
Due to a typo in spec.gmk.in, the platform specific include files have
lost their platform sub directory in the jdk image include dir.
Bug: https://bugs.openjdk.jav
+ build-dev
On Fri, Jun 15, 2018 at 2:17 PM Sebastian Demian <[email protected]>
wrote:
> Thanks for the reply Erik.
>
> running with --disable-warnings-as-errors helped.
>
> I also ran make dist-clean before building again.
>
> The build failed at a later stage now.
>
>
>
> Here is the new
On 2018-06-15 18:02, Fridrich Strba wrote:
Hello, Magnus,
Your original patch had a problem, since the second statement overwrote
the $1_SRC_HEADER_FLAGS instead of appending to it. With that, things
like osSupport.hpp, jawt_md.h and so on could not be found in build of
tests. So, I modified
On 2018-06-15 21:18, Sebastian Demian wrote:
+ build-dev
On Fri, Jun 15, 2018 at 2:17 PM Sebastian Demian <[email protected]>
wrote:
Thanks for the reply Erik.
running with --disable-warnings-as-errors helped.
I also ran make dist-clean before building again.
The build failed at a
https://bugs.openjdk.java.net/browse/JDK-8174050
I switched from panama/panama to panama/dev instead and was able to build
it successfully.
Thank you for your help,
Sebastian.
On Fri, Jun 15, 2018 at 3:16 PM Magnus Ihse Bursie <
[email protected]> wrote:
>
>
> On 2018-06-15 21:18, Sebastian Demian wrote:
> > + build-dev
> >
> > On
I have just started getting a reliable build failure on macOS, even with a new
clone of the client repo.
I’m not sure what might have changed, other than using a new SDK (from macOS
10.14) and Xcode 10.
The key lines are:
Compiling 4 files for BUILD_JIGSAW_TOOLS
make[3]: ***
[/Volumes/A/JDK/c
36 matches
Mail list logo
