Marcus Denker

This is now fixed

Yes, and it needs some more work in Pharo9 in general (and even more in Pharo10)

There are lots of failing tests: ``` [osx-64 / Tests-osx-64 / MacOSX64.PharoDocComment.Tests.CommentTestCase.nil](https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%20and%20branch%20Pipeline/job/PR-16637/6/testReport/junit/MacOSX64.PharoDocComment.Tests/CommentTestCase/osx_64___Tests_osx_64___nil/) [osx-64 / Tests-osx-64 / MacOSX64.PharoDocComment.Tests.CommentTestCase.nil](https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%20and%20branch%20Pipeline/job/PR-16637/6/testReport/junit/MacOSX64.PharoDocComment.Tests/CommentTestCase/osx_64___Tests_osx_64___nil_2/) [osx-64 / Tests-osx-64 / MacOSX64.PharoDocComment.Tests.CommentTestCase.nil](https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%20and%20branch%20Pipeline/job/PR-16637/6/testReport/junit/MacOSX64.PharoDocComment.Tests/CommentTestCase/osx_64___Tests_osx_64___nil_3/) [osx-64 / Tests-osx-64 / MacOSX64.UnifiedFFI.Tests.FFIAutoReleaseOptionCalloutTest.testExternalStructure](https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%20and%20branch%20Pipeline/job/PR-16637/6/testReport/junit/MacOSX64.UnifiedFFI.Tests/FFIAutoReleaseOptionCalloutTest/osx_64___Tests_osx_64___testExternalStructure/) [osx-64 / Tests-osx-64...

Are you sure the failing tests are not related? It is hard to me as we now have 76 (!) failing tests. Just checking that there is not one more...

Splitting the commits at a class boundary is much faster (25 seconds): ``` Smalltalk globals allClasses do: [ :cl | | tr | tr := soil newTransaction. cl localMethods do:...

This is the current list. - [ ] SoilPagedFileIndexStore>>#hasStream - [ ] SoilMetrics>>#objectRecordCacheMaximumSize - [ ] SoilCommitTransactionEntry>>#objectIds - [ ] SoilNewBehaviorEntry>>#objectIds - [ ] SoilPluggableIndexRewriter>>#itemBlock: - [ ] SoilPluggableIndexRewriter>>#indexBlock: -...

I updated the list and updated the code snippet to generate the list of currently unused methods. This is a good starting point to improve code coverage / find untested...

It would be nice to have a symmetric readHeaderFrom: for every #writeHeaderOn:, but of course this is not really that important (it would aide understanding an simplify changes in the...

One possibility would be to use a special Slot class that encodes the meta data. ("this slow is named "x" but is used to be "z"). Similar to #102 ,...

We can add a RenamedSlot -> it should be virtual (subclass of Slot, nit IndexedSlot, thus it takes no offset) -> when writing, it should reflectivey write to the new...