Re: maven.repo.local cf. localRepository

2024-07-14 Thread Timothy Stone
On 7/14/24 10:04 AM, Romain Manni-Bucau wrote: Hi Agree it should be absolute but there is no technical constraint - ultimately it must be known how it resolves. I know on ci it is common to use a relative path to avoid ci specific internal path issues accross os/workers. I concur. But as I n

Re: maven.repo.local cf. localRepository

2024-07-14 Thread Christoph Läubrich
Using relative path can have some surprises, so I would avoid it whenever possible. Especially on CI there is usually an (absolute) workspace path available. Technically the JVM will transform a relative path into an absolute one by using the currents process working directory but "relative" m

Re: maven.repo.local cf. localRepository

2024-07-14 Thread Romain Manni-Bucau
Hi Agree it should be absolute but there is no technical constraint - ultimately it must be known how it resolves. I know on ci it is common to use a relative path to avoid ci specific internal path issues accross os/workers. Le dim. 14 juil. 2024 à 09:52, Tamás Cservenák a écrit : > Howdy > >

Re: maven.repo.local cf. localRepository

2024-07-14 Thread Tamás Cservenák
Howdy Yes, usually you do want it absolute. The relative examples like "-Dmaven repo.local=../repo" you usually can find in some examples or reproducers. If local repo path relative, these will end up using different local repositories (imagine some multi module build): $ mvn install $ mvn -f s

maven.repo.local cf. localRepository

2024-07-13 Thread Timothy Stone
I help maintain Maven tooling for To Be Continuous, a GitLab CI/CD catalog of pluggable pipeline configurations. There is some discussion on the use of the JVM property, `-Dmaven.repo.local=path' and the note on "Configuring your Local Repository" (see [1]). Initially I was going to post to the u