> Le 17/04/2024 à 09:54, Timo Röhling a écrit : > >> FastDDSGen currently depends on OpenJDK 11 (or more precisely, any >> OpenJDK before 15) because it needs the Nashorn JavaScript engine, so >> could OpenJDK 11 be kept in trixie for the time being? > > That's a dead end, if OpenJDK 11 is kept there will be no incentive to > update softwares to follow the evolution of Java, and the old JDK will > just hang there forever.
For many years Java's philosophy used to be maintaining backwards compatibility as long as humanly possible. This was sometimes frustrating and limiting, but OTOH this was also one of the most important reasons why Java used to be one of the most popular choices for enterprise and other build-to-last systems. Many of these systems still run on older versions of Java, sometimes maybe even on hardware that is no longer supported by newer versions: according to New-Relic stats in 2024 60% of deployments using their software were running on Java 11 or 8 (32.9% and 28.8% respectively, source: https://newrelic.com/resources/report/2024-state-of-the-java-ecosystem , the first graph). After Java-11, this approach was somewhat relaxed, causing some problems here and there from time to time (for example some 11-compliant code cannot be built using 17+ versions). This in turn _severely_ decreased migration rate of existing systems to newer versions: for sys-ops teams maintaining systems that have been running undisturbed for several years, there's a fundamental difference between "can be replaced without any additional changes" and even "requires also this 2-line sed-script/patch/etc". As a consequence of the 2 above facts, OpenJDK-11 is currently the most popular choice for building libraries and infrastructure software. Therefore as `openjdk-11` package seems to be still maintained anyway, it would be beneficial for Debian adoption to include it in trixie before it becomes stable. Regards PMK
