river-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Greg Trasuk <tras...@stratuscom.com>
Subject Does it really make sense to have a "Convenience Binary" artifact?
Date Thu, 21 Jan 2016 08:51:00 GMT

In going through the exercise of cleaning up the release artifacts, I’ve started to wonder
if it actually makes sense to publish a “binary distribution” of the JTSK separately from
publishing the artifacts to Maven Central.

Basically, there is nothing in the JTSK that you can "just run”.  Contrast this with something
like Tomcat, where you might download the binary distribution and “just run” the web server.
 All you can do with the JTSK as it stands is run the QA suites.  To do that essentially requires
starting from the source distribution, since the Ant scripts do a build before running the
integration testing (and it really isn’t practical to run the tests without the Ant scripts).
 The browser jar is there, but frankly should probably be taken out (as we did in the 2.2
branch), because to actually use it you need binaries, policy files, etc, which haven’t
been maintained in the JTSK for years.  People should start from the river-examples project,
or Rio, or Harvester or StartNow if they want to setup a Jini infrastructure and play with
it.

Any useful examples or applications will be getting the compiled jars from Maven Central (via
Maven, Gradle, or Ivy).  I suppose one might argue that it’s useful to ship the collection
of compiled jars with their dependencies (Groovy and high-scale-lib), but I suspect that most
people are using dependency management theses days.  So I’m not sure if it’s worth the
effort to maintain the packaging scripts and the alternate license and notice files that we
would need for a binary release.

Opinions?

Cheers,

Greg Trasuk
Mime
View raw message