-
Notifications
You must be signed in to change notification settings - Fork 6
Build refactoring for generic connector #13
Comments
At some point we need to update the gradle build files/cicd to publish a maven artefact - update here if/when you need help with that/are ready to do it. The hardest part is probably deciding the name(s) of the artifacts(assuming we keep them all distinct) This will mean the connector gets published -- even as a snapshot -- so it will be downloadable. Later we can consider released versions. At that time we should probably add an 'enforcedPlatform' reference in the top level gradle file -- this will cause us to use the same versions as egeria (this is how we did it in the hive megastore connector) |
Agreed in team meeting 30 Aug that I will make the change to get the new connectors published as maven artifacts |
These are all being published - but with names based on the directory name. Either we need to move the directory to use as the default component name (our usual pattern in egeria) or explicitly specify the project name (probably the simplest for now) ^^ These can be accessed NOW (since I merged the PR that I wrote before my vacation, but was awaiting review so only merged this morning) I'll update the artifact names later in the week |
@planetf1 Thanks - I looked in both Maven Central and Sonatype for a jdbc.snapshot.jar but didn't see one? Am I looking for the wrong thing or looking in the wrong place? |
The artifacts are only available on the SNAPSHOT repository, since we have not yet done a release for the database connectors (at least not the new ones). I did ask at the f2f but the consensus was not yet. In the snapshot URL, as an example, the resource connector can currently be found under: However, as these are snapshots, all files include the timestamp ,making it awkward to fetch So you can issue a query like
What this shows, however, is that the actual connector is not published ie: We have javadoc, sources, but no actual code ;-) Thanks for spotting. |
Signed-off-by: Nigel Jones <[email protected]>
#180 Correct maven publish - artifact names
Connectors now being published with new names:
And...
As an example we can see this also now includes the jar correctly: At this point the http query isn't working - this may depend on an index update (or another bug?) |
The http query will work -- for the jar, do not specify a classifier ie:
Note that both jars are currently built without dependencies. The integration connector does require the resource connector. Both require a jdbc connector for the database being used. I've not verified the runtime behaviour to see if anything else extra is required, though there should be no change here from previously |
(extra points if anyone spots my error in the current code. cosmetic only!) |
Inspecting .module file -- this is a gradle addition that provides additional metadata for gradle consumers The top level metadata looks good:
The java jar is properly represented with:
Checking the pom
And the files
And the maven-metadata.xml
By manual inspection I do not see issues..... at least when publishing to local maven repo |
Checking retrieving the jar via curl:
This looks ok (repeat for any other connectors) Next we check with the proper maven tools - and can see that again, the maven artifact is correctly downloaded:
|
In summary - for this repo's publish flow:
More broadly we need to think about usability
I suggest these general issues are followed up in Does that help @dwolfson @mandy-chessell ? |
Tested with postgres connectors: Using a file ~/etc/jdbc.yaml:
We can install an instance of the helm chart ie:
This
After the container is started, we can see the connector is present:
|
Update on above - the connectors are the wrong content .The reason is that the specific URL specified utilizes a query which results in a http redirect as a response. In the charts we do not follow redirects, so the http redirect is downloaded, not the target of the redirect.. |
This error was a user problem - missing quotes in the URL which then failed (incorrect end result, not a failure....) on the curl request The correct jdbc.yaml is:
|
This issue has been automatically marked as stale because it has not had recent activity. It will be closed in 20 days if no further activity occurs. Thank you for your contributions. |
This issue has been automatically marked as stale because it has not had recent activity. It will be closed in 20 days if no further activity occurs. Thank you for your contributions. |
this was completed as part of the original work |
The text was updated successfully, but these errors were encountered: