This seems like something that would be easier to support with a modular log4j-core, though that's a much larger effort than fixing missing class issues.
On 13 September 2017 at 10:37, Ralph Goers <[email protected]> wrote: > Great. But how would you do it? > > Ralph > > > On Sep 13, 2017, at 7:11 AM, Gary Gregory <[email protected]> > wrote: > > > > I would like to see support for Android, all in our one code base, for > now. > > > > Gary > > > > On Sep 13, 2017 08:02, "Ralph Goers" <[email protected]> wrote: > > > >> We are getting Jira issues about getting log4j to work in Android. At > >> first, all I thought was required was getting the API to function on > top of > >> Android’s logging system. However, it seems that there are some who > want to > >> use the RollingFileAppender or perhaps other appenders. The issue I > have is > >> that I am pretty sure there are other things inside core that won’t > work in > >> Android, which was why my first attempt was to just get the API working > >> there. > >> > >> Right now there are 8 open issues that mention Android. LOG4J2-1915, > >> LOG4J2-1920 LOG4J2-1921 all say that 2.8.2 doesn’t work, which was > before > >> the Java 9 support was added. LOG4J2-2042 says 2.9 doesn’t work because > of > >> a dependency on java.lang.management classes. LOG4J2-1920 says it has an > >> error because the ScriptEngineManager is not available. > >> > >> I am not sure how to solve this. We could just say Android is not > >> supported although that would likely cause problems for a lot of > people. We > >> could only support API integration, as I first tried to do, We could > put > >> in a bunch of code to try to detect android and disable stuff that > doesn’t > >> work there, but that is quite a bit of work and as far as I can tell > none > >> of the committers work with Android. Finally, we could create a > >> log4j-android that contains the API combined with a minimal version of > core. > >> > >> Any thoughts on this? > >> > >> Ralph > >> > > > -- Matt Sicker <[email protected]>
