How exactly does setting the sysprop for only 18 and 19 allow folks to test their stuff? If Ant currently depends on the security manager to operate, why not set the sysprop regardless, then remove in future when a replacement API exists?
Since I work on a project that embeds Ant and uses it API, I am trying assess what I need to do on my end to mitigate the problem. I do not use the launcher scripts, but invoke Ant directly as an embedded service (same JVM) or via an external JVM process (most common usage). Regards, --ewh On Mon, Feb 7, 2022, 4:18 AM Jaikiran Pai <jaiki...@apache.org> wrote: > Hello Stefan, > > I was planning to send out a mail about this change later tonight. But > good you brought this up. To give some background, the security manager > changes starting Java 18 make it such that setting of the security > manager at runtime now throws an exception, which effectively fails all > builds since Ant by default sets up a security manager. After various > experimentations and checking over with the JDK team, it appears that we > (Ant) can't get rid of setting the security manager till the JDK team > provides an API to prevent System.exit(...) calls. So in the meantime to > allow users of Ant to build their projects and test for any issues > against Java 18 (and now Java 19 EA), I decided to specifically set this > system property only for these specific versions. Initially I had only > done it for Java 18, hoping that a new API for System.exit(...) > prevention would be availbale in 19, but it isn't ready so far. So the > launch scripts (the Linux one and the .bat for Windows one) have both > been updated to set this system property only for Java 18 and 19. I > expect this to be a temporary thing till the new API is available. Once > the new API is available, I think we should just get rid of this setting > of system property even for Java 18 and 19 (since I don't expect many to > be using these releases once the newer versions are available). > > Now coming to the actual implementation of this, it took me multiple > weekends to get the .bat version to correctly work. Mainly due to lack > of easy access to a Windows setup plus my general lack of knowledge of > bat scripting and some gotchas when it comes to .bat parsing and the > "errorlevel" values. Ultimately the one I committed ant.bat now launches > the Java command twice and expects it to dump certain property values, > which are then used by "find" to see if the version is 18 or 19. So > essentially, launching Ant (on Windows) now additionally triggers > lauching of two Java process (they do exit) just to check the version. > It's not the best way, but I couldn't find any other way to do this. > > As for the Linux version of the ant launch script it does a similar > thing but in its case it just launches the Java program once and figures > out if the version is 18 or 19. > > With these changes the CI builds which run Ant tests against Java 18, 19 > and previous version like Java 17 now work fine. However, like I said my > scripting skills are minimal, so if any of these changes in these > scripts can be done in a better way, please feel free to do so. I would > do it myself, but it's going to take me trial and error methods to get > it right :) > > -Jaikiran > > On 07/02/22 12:26 pm, Stefan Bodewig wrote: > > On 2022-02-07, <jaiki...@apache.org> wrote: > > > >> - if [ "$JAVA_SPEC_VERSION" = "java.specification.version=18" ]; then > >> + if [ "$JAVA_SPEC_VERSION" = "java.specification.version=18" ] || [ > "$JAVA_SPEC_VERSION" = "java.specification.version=19" ]; then > > Bourne shell's case may make this more > > readable (not sure whether there is a Windows batch file equivalent) > > > > case $JAVA_SPEC_VERSION in > > java.specification.version=18 | \ > > java.specification.version=19 ) > > ... > > ;; > > esac > > > > But I'm afraid this is not going to scale :-) > > > > In the long run we probably are better off by enumerating the JDKs where > > we don't want to set the property (ten from 8 to 17, but this is a fixed > > list that doesn't need to change with every release). > > > > Stefan > > > > --------------------------------------------------------------------- > > To unsubscribe, e-mail: dev-unsubscr...@ant.apache.org > > For additional commands, e-mail: dev-h...@ant.apache.org > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@ant.apache.org > For additional commands, e-mail: dev-h...@ant.apache.org > >
