I switched to Codeberg actually, the class would be here:
https://codeberg.org/goldmensch/jack/src/branch/dependencies-and-lockfile/src/main/java/io/github/goldmensch/cache/CacheLocalRepositoryManager.java
So, I have to create a root session, without any LocalRepositoryManager
set and then create a sub session from that root, which has the
LocalRepositoryManager set. Did I get this right?
Additionally, both sessions should only exist for the time of this
dependency fetch task or should the root session exists for the whole
lifespan of the application?
On 12.07.24 12:25, Tamás Cservenák wrote:
Yes, in that case, if you want to go "full resolver 2" way in Jack,
I'd do following:
- create a "root" session
- derive your own session from root (nested/derived session)
- used derived session
- once done, close derived and close root session
I don't see what CacheLocalRepositoryManager does, I guess it is new?
Is not in https://github.com/Goldmensch/jack
T
On Fri, Jul 12, 2024 at 12:17 PM Nick Hensel <nickhense...@icloud.com>
wrote:
Thanks for the fast reply :)
Actually the examples set the paths of the local repository
instead of
the LocalRepositoryManager directly. My problem here is, that I'm
actually
using an own implementation of LocalRepositoryManager that wraps 2
instances of EnhancedLocalRepositoryManager. I forgot to include
that in
my example code, my bad.
Here's an update version of the code I want to run:
```java
RepositorySystemSession session =
repositorySystem.createSessionBuilder()
.setLocalRepositoryManager(new
CacheLocalRepositoryManager(userConf.caching().global(), SESSION,
repositorySystem)).build();
```
The contructor of CacheLocalRepositoryManager looks like that:
```java
CacheLocalRepositoryManager(Global conf, RepositorySystemSession
session, RepositorySystem system) {
this.conf = conf;
this.local = system.newLocalRepositoryManager(session, new
LocalRepository(SpecialPaths.JACK_DIR.resolve(MAVEN_DIR)));
this.global = system.newLocalRepositoryManager(session, new
LocalRepository(SpecialPaths.GLOBAL_JACK_DIR.resolve(MAVEN_DIR))); }
```
I don't know where to get the SESSION from at this early point.
On 12.07.24 11:50, Tamás Cservenák wrote:
> Also, check out demo snippets:
>
https://github.com/apache/maven-resolver/blob/d13019ce5fa38c0951ca0fb0c62a3f5f1254cffc/maven-resolver-demos/maven-resolver-demo-snippets/src/main/java/org/apache/maven/resolver/examples/util/Booter.java#L68
>
> T
>
> On Fri, Jul 12, 2024 at 11:45 AM Tamás Cservenák
<ta...@cservenak.net>
> wrote:
>
>> Howdy,
>>
>> see here and below:
>>
>>
https://github.com/apache/maven-resolver/blob/d13019ce5fa38c0951ca0fb0c62a3f5f1254cffc/maven-resolver-api/src/main/java/org/eclipse/aether/RepositorySystemSession.java#L439
>>
>>
>> T
>>
>> On Fri, Jul 12, 2024 at 11:29 AM Nick Hensel
>> <nickhense...@icloud.com.invalid> wrote:
>>
>>> Hi,
>>> I'm currently migrating my Maven Artifact Resolver 1.x code to
2.0.
>>> As of version 2.0 the recommend way to create a
RepositorySystemSession
>>> is through a SessionBuilder.
>>> In my case, I need to set an own LocalRepositoryManager, which
wraps two
>>> EnhancedLocalRepositoryManager instanced. The problem here is,
that in
>>> order to create an instance of it I have to call
>>> RepositorySystem#newLocalRepositoryManager which needs a
>>> RepositorySystemSession, which I obviously cannot pass before
creating it.
>>>
>>> To demonstrate this problem, here's a piece of code.
>>> ```java
>>> RepositorySystemSession session =
>>> repositorySystem.createSessionBuilder()
>>>
.setLocalRepositoryManager(repositorySystem.newLocalRepositoryManager(SESSION,
>>> new LocalRepository(...)) // where should I get SESSION from?
>>> ```
>>>
>>> Best regards,
>>>
