71–80 of 91 posts
Earlier quoted context omitted.
Technically? Yes, Correto is it's own separate distribution that Amazon manages porting upstream changes into. Same as how Azul's Zulu is a "competitor" as well. Practically speaking, it shouldn't matter as they all pass the Java TCK.
1. AdoptOpenJDK does not run the TCK. 2. Corretto and Zulu each have their own patch set. Corretto backports a couple of G1 fixes and Zulu backports Flight Recorder and TLS 1.3.
It should be noted that AdoptOpenJDK does not run the TCK against these binaries so they can not be called a compatible Java implementation.
The TCK is not open, it isn't governed by a standards body, the TCL is completely in Oracle's control. If Java is to evolve as an open platform, it needs to break free from the TCK that disallowed Apache Harmony to exist. We don't need it and an independent TCK could be developed.
Earlier quoted context omitted.
Granted, Oracle does contribute lots of code. As the result, lots of Oracle-created code gets backported too. That said, I see people backport a lot of code/tests that they wrote themselves, and/or with the help of non-Oracle contributors. For example, Red Hat, SAP, Huawei backport a lot of code to maintain the target platforms they own (e.g. AArch64 and PPC), or their components (e.g. diagnostics support, Shenandoah…
I am not undervaluing anyone's contribution. Many companies contribute a lot of valuable work, and we at Oracle are thrilled to cooperate with them and actively encourage them to contribute more, but Oracle does currently contribute ~90% of the work on OpenJDK. So many companies and independent developers deserve full credit, but Oracle is in a separate category, contributing as much as all others combined, several t…
Implying applications always receive continued development...
Earlier quoted context omitted.
I am not undervaluing anyone's contribution. Many companies contribute a lot of valuable work, and we at Oracle are thrilled to cooperate with them and actively encourage them to contribute more, but Oracle does currently contribute ~90% of the work on OpenJDK. So many companies and independent developers deserve full credit, but Oracle is in a separate category, contributing as much as all others combined, several t…
> the recommended path is free perpetual support by using the current JDK version that, under the new gradual release model, means you'll never have to do another major upgrade again. Implying applications always receive continued development...
Earlier quoted context omitted.
The JDK requires no installation, and the concept of a "desktop system JRE" is no longer recommended; in fact, a JRE is no longer shipped. The recommended practice is for developers to bundle a custom, minimized version of the runtime with their application (using jlink). End users shouldn't install any Java runtime anymore. > After the bait and switch with the licensing for the original Java distribution The product…
Without installation on Windows `Preferences.systemRoot();` writes bunch of log errors in stderr. It tries to create HLKM\Software\JavaSoft\Prefs. Installer creates that registry node, so it works just fine after that. I don't know if it's a bug or intended behaviour.
Earlier quoted context omitted.
The TCK is not open, it isn't governed by a standards body, the TCL is completely in Oracle's control. If Java is to evolve as an open platform, it needs to break free from the TCK that disallowed Apache Harmony to exist. We don't need it and an independent TCK could be developed.
Yet any other OpenJDK distribution from Alibaba, Amazon, Azul, BellSoft, RedHat, SAP (I probably forgot some) all run the TCK. All except AdoptOpenJDK. And whenever you bring that up the only response is ranting about the TCK.
The existence and reliance on that TCK however means that Java is NOT an open platform, not as long as you insist on having implementations pass the TCK, compliance that's defined by whatever Oracle wants. And lets be clear about this, we are talking about what Oracle wants, not about JCP, which btw is only a farce of a standards body, not a real one.
This problem is old and I invite people to not be ignorant about the issue at play. Please read Apache's open letter to the former Sun Microsystems, as it does justice to this issue better than I could:
https://www.apache.org/jcp/sunopenletter.html
---
The irony of the situation is that the direct competition, .NET and C#, are managed by a real standards body and alternative open source implementations (e.g. Mono) have been allowed to exist, without backroom deals with Microsoft ;-)
Earlier quoted context omitted.
Without installation on Windows `Preferences.systemRoot();` writes bunch of log errors in stderr. It tries to create HLKM\Software\JavaSoft\Prefs. Installer creates that registry node, so it works just fine after that. I don't know if it's a bug or intended behaviour.
Which version are you using? This could be this issue[1], fixed in 9. [1]: https://bugs.openjdk.java.net/browse/JDK-8139507
Earlier quoted context omitted.
Without installation on Windows `Preferences.systemRoot();` writes bunch of log errors in stderr. It tries to create HLKM\Software\JavaSoft\Prefs. Installer creates that registry node, so it works just fine after that. I don't know if it's a bug or intended behaviour.
Which version are you using? This could be this issue[1], fixed in 9. [1]: https://bugs.openjdk.java.net/browse/JDK-8139507
Earlier quoted context omitted.
Which version are you using? This could be this issue[1], fixed in 9. [1]: https://bugs.openjdk.java.net/browse/JDK-8139507
I checked on latest 12.0.2+10 it's still reproduced.