Improvements to make AOSD reports better comparable #9615
Conversation
sschuberth
changed the title
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #9615 +/- ##
=========================================
Coverage 68.02% 68.02%
Complexity 1285 1285
=========================================
Files 249 249
Lines 8828 8828
Branches 921 921
=========================================
Hits 6005 6005
Misses 2433 2433
Partials 390 390
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. |
…thing Omit the `selectedLicense` if it is the same as `spdxId` and might still offer a choice. Also, an absent `selectedLicense` should be encoded as the empty string, not as SPDX `NOASSERTION`. Signed-off-by: Sebastian Schuberth <[email protected]>
sschuberth
force-pushed
the
aosd-comparison
branch
from
7a5dcec to
03d7ec5
Compare
| @@ -62,6 +62,7 @@ class Aosd21Reporter(override val descriptor: PluginDescriptor = Aosd21ReporterF | |||
| indexedPackages[it]?.index?.toLong() | |||
| }, | |||
| components = indexedPackages.toComponents(input, project) | |||
| .toSortedSet(compareBy { it.componentName }) | |||
There was a problem hiding this comment.
There was this larger effort to eliminate the use of sorted collections in favor of using annotation to do the sorting for the serialization only. I believe we should annotate the components property instead.
There was a problem hiding this comment.
Please note that the model still only uses a Set, not a SortedSet. I.e. users of the data model are not forced to use a sorted set, esp. when only using the data model in memory. And with my change even users that want to serialize are not forced to use a sorted set. In that sense my change is even less strict than enforcing sorting on any serialization.
There was a problem hiding this comment.
Please note that the model still only uses a Set, not a SortedSet.
Was aware of that.
And with my change even users that want to serialize are not forced to use a sorted set.
I suppose with users your speaking of callers from within the same component. It's not obvious to me
whether there is need for using different orderings within the reporter.
...anyhow, I thought the annotation approach would be favorable by default, because it yields a more deterministic serialization.
There was a problem hiding this comment.
Anyway, I'm now sorting on serialization so that files can always be easily compared, even if the data model would be serialized by another caller.
| @@ -11,8 +11,8 @@ | |||
| ], | |||
| "components": [ | |||
| { | |||
| "id": 0, | |||
| "componentName": "no-license-file", | |||
| "id": 3, | |||
There was a problem hiding this comment.
commit message: Why is it favorable to align with [1] ?
There was a problem hiding this comment.
As said in the title: To more easily compare the JSON output files from both tools with a diff tool, while focusing on real differences.
There was a problem hiding this comment.
I had read that title of course. To me it wasn't clear the motivation behind being able to compare to that tool?
There was a problem hiding this comment.
I've also added to the commit message that spdx2aosd actually is a tool that's officially used by Audi.
sschuberth
force-pushed
the
aosd-comparison
branch
from
03d7ec5 to
057b658
Compare
sschuberth
force-pushed
the
aosd-comparison
branch
from
057b658 to
7e108f6
Compare
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
This eases comparison with the files generated by [1], which is a tool officially used by Audi. [1]: https://gitlab.com/digitalwerk/tools/spdx2aosd Signed-off-by: Sebastian Schuberth <[email protected]>
This eases file comparison. Do not trim leading whitespace as some license texts (like for `Apache-2.0`) actually start with indentation. Signed-off-by: Sebastian Schuberth <[email protected]>
sschuberth
force-pushed
the
aosd-comparison
branch
from
7e108f6 to
b4dd81d
Compare
Please have a look at the individual commit messages for the details.