Skip to content
This repository has been archived by the owner on Oct 11, 2023. It is now read-only.

Did the BundleID change? #56

Closed
waynebeaton opened this issue Sep 29, 2023 · 1 comment
Closed

Did the BundleID change? #56

waynebeaton opened this issue Sep 29, 2023 · 1 comment

Comments

@waynebeaton
Copy link

The Eclipse Dash License Tool uses this repository to figure out how to map a bundle Id (e.g., com.atlassian.commonmark-gfm-tables) to the corresponding GAV (e.g., com.atlassian.commonmark:commonmark-ext-gfm-tables). The scanner walks this repository and pulls this information out of the pom.xml files. It generally works (with some small amount of hinting).

The expectation is that as we encounter new versions of things, they are likely to have similar IDs to the older versions. The exceptions are handled with the aforementioned small amount of hinting.

The Orbit build (the only information that I have is that the issue "was created 2 days ago by orbit bot user) recently created a handful of reviews for which the scanner cannot find source code.

For example: https://gitlab.eclipse.org/eclipsefdn/emo-team/iplab/-/issues/10695 (there's actually a bunch of similar "commonmark" entries).

Here... we need to map the bundle ID org.commonmark-gfm-tables. The closest that I can find in this repository is a bundle named com.atlassian.commonmark-gfm-tables.

Are these the same thing? Did the name change intentionally? Is this a mistake?

@merks
Copy link
Contributor

merks commented Sep 30, 2023

I think this ip lab issue was asking an analogous question so the same information applies:

https://gitlab.eclipse.org/eclipsefdn/emo-team/iplab/-/issues/10729

The dash scanner does significantly more than look at pom.xml files. It also appears to monitor Tycho's processing of *.target files including Tycho's processing of the BND instructions used to wrap a Maven non-OSGi artifact as an OSGi bundle.

I see this in the log for the processing of the Maven GAV org.commonmark:commonmark-ext-gfm-tables:0.21.0:

[INFO] org.commonmark:commonmark-ext-gfm-tables:0.21.0 is wrapped as a bundle with bundle symbolic name org.commonmark-gfm-tables
[INFO] The artifact can be referenced in feature files with the following data: <plugin id="org.commonmark-gfm-tables" version="0.21.0.v20230927-1332" download-size="0" install-size="0" unpack="false"/>
...
[INFO] A review request already exists https://gitlab.eclipse.org/eclipsefdn/emo-team/iplab/-/issues/10732 .
[INFO] A review is required for p2/orbit/p2.eclipse.plugin/org.commonmark-gfm-tables/0.21.0.v20230927-1332.

This naming/version pattern is used because that was what was used in the past by Orbit EBR-baed builds which presumably haven been IP reviewed in the past:

So yes, the latter is derived from the former, and the naming is often driven by what was done previous by the Orbit EBR builds so that the resulting bundles retain the same symbolic name as was historically the case.

@merks merks closed this as completed Sep 30, 2023
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants