-
Notifications
You must be signed in to change notification settings - Fork 169
MSL v4.1.0 release
Here's the schedule for releasing MSL v4.1.0
- 2024-01-12: Feature freeze, no new PRs can be opened
- 2024-01-12 to 2024-02-02: close PRs that only need a review, all other PRs are moved to 4.2.0; regression testing with Dymola against results obtained with 4.0.0 is performed
- 2024-01-23: Branch off development branch as maint/4.1.x and tagged as v4.1.0-alpha.1
- 2024-02-09: Tag v4.1.0-beta.1
- 2024-02-02 - 2024-12-12 resolve issues with 4.0.0 - 4.1.0 regression testing and finalize 4.1.0 reference files
- 2024-12-12: Tag v4.1.0-beta.2
- 2024-12-12 - 2025-01-12 regression testing by tool vendors
- 2024-01-12?: Maintenance tagged as v4.1.0-rc.1, announcement for voting, rerun regression testing
- 2024-01-20?: start of voting (among MAP-Lib members) to accept the MSL v4.1.0-rc.1 as new MSL release
- 2024-02-03?: MA approval and MSL 4.1.0 release
- Reference results for 4.0.0 are stored in https://github.com/modelica/MAP-LIB_ReferenceResults/. For each model, the creation.txt file documents which simulation settings were used to generate the results
- Reference results for 4.1.0 are stored in https://github.com/modelica/MAP-LIB_ReferenceResults/tree/v4.1.0
- Regression testing is run on LTX servers, results are reported here, first the results obtained with the simulation settings contained in creation.txt, then the results obtained with the simulation settings of the respective version experiment annotation
- Regressions are reported by sub-library in these issues.
- Library officers should go through each model and:
-
identify the cause of regression and document it.
Possible reasons for regressions have been created by git bisect: Modelica, ModelicaTest, search the tag of the first bad commit on github, which typically points to a pull request with extensive explanation of the reasons for the difference. -
if there is a good reason for the regression (e.g. a bug was fixed), write "@GallLeo Please update reference". LTX will update the reference files in https://github.com/modelica/MAP-LIB_ReferenceResults/tree/v4.1.0 and check the checkbox. This will be later documented in the release notes
-
if there is a need to fix the MSL code, open an issue and resolve it. It will be resolved when the fix is pushed both to master and maint/v.4.1.0, then the library officer can check the checkbox.
-
- When all references are resolved, tool vendors can start with their regression testing. They can already do it now using results from https://github.com/modelica/MAP-LIB_ReferenceResults/tree/v4.1.0, disregarding regressions that show up in the regression reports
- Open pull requests: https://tinyurl.com/openpulls410
- Open issues: https://tinyurl.com/openissue410
- Especially, open features: https://tinyurl.com/openfeatures410
- Especially, regressions: https://tinyurl.com/regressions410
- Overview of regression tests against 4.0.0 run by LTX with Dymola
- Detailed report of regression tests against 4.0.0 run by LTX
- Regressions between 4.1.0 and 4.0.0, classified by sub-library