There are a very few places, where SAPMachine specific changes in shared code were made but those were not tagged or the diff is unnecessary.

Update the "Have an issue section?" in Readme.md

The change for JDK-8217773 needs to be cherry-picked because it conflicts with the temporary solution to fix test jdk/javadoc/doclet/testLinkOption/TestRedirectLinks.java of the langtools suite.

After the cherry-pick, source is identical with OpenJDK again.

The test jdk/java/net/Socket/ExceptionText.java currently fails in sapmachine11. Fix this.

There are a few typos in README.md which should be fixed.
One was reported Lucas Gabriel Schneider in PR #177.

Add a debug transport called dt_filesocket which uses safe ways to connect to a client.

It should use named pipes on Windows and Unix domain sockets on Unix. For the first version, no JDI connector which could handle this transport is needed, just the raw transport.

It's time to update the langtools problem list a little bit. One failing test has been fixed in the meanwhile.

Create a template for SapMachine pull requests.

The dt_filesocket debug transport in the sapmachine11 repository needs a few improvemts:

• A jtreg test
• A mechanism to avoid leaving file socket files behind
• Checking the connected user and not just relying on the file permissions. This is needed since on some platforms the connection permission is taken form the permission of the directory in which the domain sockets is created. Additionally it will be easier later to support abstract domain socket on unix later, since they cannot rely on file permissions at all.
  @

There are several places in OpenJDK where vendors can add their branding information. This gets effective in e.g. System properties, executable meta information, icons etc. SapMachine should make use of this.

There are a few places where SapMachine differs from OpenJDK. That might be different default values for VM options or configuration parameters or even added functionality.

For these differences, existing regression test cases had to be adapted or even additional test cases have been added.

Since these paths are not covered by OpenJDK community testing we shall add these tests to the tier1 layer in order to have them regularly executed already in pull request testing.

We want to deliver "8214892: Delayed starting of debugging via jcmd" in update 11.0.2.

We requested downport of this feature in OpenJDK, but the change will not make it into 11.0.2. in any case.

https://bugs.openjdk.java.net/browse/JDK-8214892

Hi,

I was at the FOSDEM talk from @simonis. This is how I become aware of the SapMachine project

Now to my request: A Windows Installer for SapMachine would be great. That's why I built one. Please have a look: https://github.com/Cyclenerd/SapMachineInstaller

With an installer a roll-out in companies is much easier and in general everything would be much easier.

Do you see a chance to implement the installer in your build process so that there is always an installer version for Windows besides the ZIP?

Best regards
Nils

The OpenJDK change JDK-8209942: [epsilon] range function for EpsilonTLABElasticity causes compiler warning contains a build fix that is required on Windows to be able to use WarningsAsErrors in the build.

This issue tracks development of JDK-8212618 "A low-cost, always on statistical value history".

This proposed feature has been rejected upstream since it is perceived to collide with JMC/JFR.
We consider it useful enough and the merge cost to be small enough to integrate it into the SapMachine.

Details:

SAP developed a supportability feature called "Statistics History" which is a low cost history of statistical values-of-interest.

These values contain parameters of the jvm (e.g. heap size, metaspace size, number of loaded classes) and the underlying platform (e.g. rss, swapping state, run queue length etc). At intervals of (by default) 60 seconds these values are measured and stored in a fifo buffer.

The fifo buffer has three parts, a short-, medium-, and long term fifo buffer. A fraction of the samples falling out of the short term fifo is transferred to the mid term fifo; again, a fraction of the samples falling out of the mid term fifo is transferred to the long term fifo. So, the short term fifo covers a short recent timespan (usually an hour) in comparativly short sample intervals (usually 60 seconds), whereas the long term fifo covers a very long time span (~10 days) with interval times of hours.

This feature has been very popular with our support folks and so we would like to contribute that. It enables us to easily analyze slowly developing situations like memory leaks, memory or cpu spikes, resource starvation etc.

[1] JDK-8212618 - Original JBS issue, closed as Wont Fix
[2] Patch on Github
[3] Example output

Missing quote in default definition of VENDOR_URL_BUG.

Hi. More a question than a defect....:
Is it allowed for firms (f.e. SAP partners) to deliver the SapMachine jre (bundelt) with their product (as private jre) without loosing their commerical rights? Or must they deliver on GPL (with copyleft)?
Thanks for the Information
Marco

After the changes for the following changes were pushed to openjdk main line, we can re-enable some tests.

https://bugs.openjdk.java.net/browse/JDK-8215568
https://bugs.openjdk.java.net/browse/JDK-8217473

Backports for sapmachine11.2:
8213086: Compiler thread creation should be bounded by available spac…
8214352: C1: Unnecessary "compilation bailout: block join failed" wit…

I'd like to backport 8213086 from 11u3 to have a better limit for compiler threads.
The backport of 8214352 from 13 is for better behavior with debugging on demand (JVMTI "can_access_local_variables" enabled). Both changes apply cleanly.

Support for Posix permissions with jdk.zipfs (as proposed in https://bugs.openjdk.java.net/browse/JDK-8213031) shall be added to SapMachine 11.

There are a few jtreg tests that currently fail and which have to be resolved in SapMachine (e.g. different default values in SapMachine vs. OpenJDK).

The zipfs enhancements that allow for storing POSIX permissions in zip archives via the zipfs file system provider of SapMachine11 have a compatibility issue with earlier SAP-closed versions of this.

The original closed version would restore an empty set of permissions as default return value of ZipFileAttributes::permissions while the initial version in SapMachine11 would throw an UnsupportedOperationException. The later needs to be changed to keep compatibility.

There's still an issue with jdk/nio/zipfs/TestPosixPerms.java, seen on Linux/Unix machines where the Default Filesystem is POSIX.

I've added SapMachine to my list of OpenJDK builds. Other builds have a page describing how long each build will have security updates. Is there such a page/commitment for SapMachine? Thanks

test

(Cherry-picks #220)

"8211845: A new switch to control verbosity of hs-err files" introduces the flag
ExtensiveErrorReports to OpenJDK. We contributed this change to add more
verbose output to hs_err files that is not desired in OpenJDK by default.
An example usage is "8211326: add OS user related information to hs_err file"

To improve the servicability, we enable ExtensiveErrorReports in SapMachine11 by default.

Downport 8207849: Allow the addition of more number to the Java version string

It would be nice if you could support macOS too, since quite a few devs use it and it's also a supported OS inside SAP.

runtime/SharedArchiveFile/serviceability/ReplaceCriticalClasses.java is failing in OpenJDK.
Exclude the test until fixed.

SapMachine11 includes a preliminary solution for accessing poxis permisisons in zip archives (https://bugs.openjdk.java.net/browse/JDK-8213031).
I'd like to push a testcase cleanup and a missing tag in ZipFileSystem.java.

In 11.0.2, a row of tests and bugs were fixed. Remove these from the ProblemList.
Add some cert tests which fail because the certificates expired.
Update our tags, remove initials.

The fix for VM Option -XX:ThreadPriorityPolicy shall go into SapMachine12 as well.

OpenJDK bug: https://bugs.openjdk.java.net/browse/JDK-8215962

I have also requested OpenJDK 12u downport .

SDKMan (https://sdkman.io/) allows you to install and manage different versions of JDKs (and other Java-related tools) via CLI, the same way RVM does for Ruby. They have a well-documented procedure describing how to add development kits and their versions to it at https://sdkman.io/vendors.

Supporting installation via SDKMan may increase SAPMachine popularity and adoption outside SAP customer base, which would help improve the product itself.

SapMachine modified the default setting of java.security property jdk.includeInExceptions, as documented in https://github.com/SAP/SapMachine/wiki/Differences-between-SapMachine-and-OpenJDK#other-minor-differences.

Due to an integration issue we are missing the value "jar" in the sapmachine branch (JDK12) and for branch sapmachine11 (JDK11) this was not integrated at all.

The build on Windows is broken because libCreateNewProcessGroupOnSpawnTest.c gets compiled on Windows when it shouldn't. The test is non Windows only.

Going through the diffs of SapMachine to OpenJDK, there are a few places, where
a) the license headers are not fully correct
b) SapMachine specific changes are not correctly tagged
c) trailing whitespace was introduced
d) Readme.md in sapmachine11 and sapmachine12 branches should be updated to the version from sapmachine

This shall be fixed/cleaned up

Downport 8215962: Support ThreadPriorityPolicy mode 1 for non-root users on linux/bsd

"8211845: A new switch to control verbosity of hs-err files" introduces the flag
ExtensiveErrorReports to OpenJDK. We contributed this change to add more
verbose output to hs_err files that is not desired in OpenJDK by default.
An example usage is "8211326: add OS user related information to hs_err file"

To improve the servicability, we enable ExtensiveErrorReports by default.
We will enable this in jdk 11, 12, too.

There are still some reference to Alpine Linux left in the README.md.
These should be removed.

The SapMachine specific problem lists need an update.

• Merge with the internal problem lists that SAP uses for OpenJDK regression testing
• Add SapMachine infrastructure specific exclusions to reduce the number of false positives when running platform builds

The test test/jdk/java/net/Socket/ExceptionText.java is failing because jdk.includeInExceptions is set to "hostInfo,jar" in SapMachine, while it's not set in OpenJDK.

A new switch is added, -Dsap.jdk.lang.Process.createNewProcessGroupsOnSpawn. If this switch is set to true, new processes started with ProcessBuilder.start or with Runtime.exec() will be leader of an own process group instead of belonging to the process group of the parent VM.

Code review did uncover some minor issues but #241 was already pushed.

https://bugs.openjdk.java.net/browse/JDK-82197212 was submitted to jdk12u but it is not part of the JDK12 release. SapMachine should include it already then.

JDK-8034802 needs to be cherry-picked to sapmachine11. It is kind of a prerequisite for the enhancements to zipfs regarding Posix permissions. It is needed to fix issues shown by its jtreg testcase jdk/nio/zipfs/TestPosixPerms.java. A downport to jdk11u will be requested in the OpenJDK as well.

Fix failing pr validation on linuxppc64le by excluding
gc/g1/humongousObjects/TestHumongousClassLoader.java
gc/g1/humongousObjects/TestHumongousNonArrayAllocation.java

Remove some tests that don't exist any more or are part of the common exclude list.

Do some cleanups.