Hi Tibor,
My offer to help stands; I sincerely hope that the suggested command line is
not the way things go; in my regular projects, we have been working very hard
to avoid the use of various —add-* switches, as they are stopgaps. Of course,
one of my real hopes is to have all of this work
Hi Russ,
All maven-compiler-plugin 3.6.x are already based on this idea. So if you
use maven-compiler-plugin 3.6.2 (current) it'll work as explained.
And yes, surefire needs some help on this, but I like to leave Tibor in
the lead for this. JUnit5 and Java9 both have impact, but the focus
Hi Robert,
That sounds great. Any idea on timing of the release? What about the surefire
plugin? Is there something I can do to help? I am actively working on Java 9
issues in a few projects, including both the module system and multi-release
jars, so I am very motivated to do what I can to
Hi,
On 27/08/17 14:38, Robert Scholte wrote:
Hi Russ,
this is how it works:
in case there's a module descriptor, the module-path with be used. With
the help of plexus-java it is possible to divide all jars over the
module-path and classpath. This mechanism is implemented in the
Hi Russ,
this is how it works:
in case there's a module descriptor, the module-path with be used. With
the help of plexus-java it is possible to divide all jars over the
module-path and classpath. This mechanism is implemented in the
maven-compiler-plugin (will do an official release soon
A bit more research turned up this page
https://www.sitepoint.com/maven-cannot-generate-module-declaration/ in which
Robert explains why my “first thought” below is a non-starter, unless the pom
spec is upgraded. So it sounds now as though we are waiting on ASM to read
module-info files…? Is
I’m sure this must have been described somewhere; is there an overall direction
document for supporting the java platform module system?
case 1)
I want to compile against another project on the module path rather than the
class path. I do this in order to ensure compile-time checking for
