NightOwl888 commented on issue #396: URL: https://github.com/apache/lucenenet/issues/396#issuecomment-734254722
> Yep all of what you said makes sense. The converter code could even be removed entirely from the master branch if we wanted and only live in the conversion branches. I have no experience with what `nbgv prepare-release` so I don't really have a comment about that one (but seems like an interesting tool!) I poked around under the hood of `nbgv` and then started experimenting to see what options Git may have to solve the issue with keeping `JavaDocToMarkdownConverter` updated on each of the `converted` branches. Then I realized something: this is not a VCS problem, it is a versioning problem. This issue can be addressed simply by: 1. Moving `JavaDocToMarkdownConverter` to a repo of its own so there is only 1 copy to deal with (https://selfserve.apache.org/) 2. Converting it into a [dotnet tool](https://docs.microsoft.com/en-us/dotnet/core/tools/global-tools-how-to-create) so it can be packaged and versioned with NuGet 3. Making all of the scripts that utilize it always uninstall and reinstall the latest version of the tool from the feed (see [dotnet tool update](https://docs.microsoft.com/en-us/dotnet/core/tools/dotnet-tool-update)) or use a specific version in cases where version drift is a problem We don't have to use the official NuGet.org feed for this, it can be MyGet, Azure Devops, GitHub, or whatever as long as it is Internet-accessible. > I'm happy to get started on this if we agree with the branch naming? Sure, that's fine. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected]
