diff --git a/dd20d3ddc86ee802fe7b15e2c91dc160/comments/9c4dc8aa38ba0a9a6f3fbc014b05f7ba/description b/dd20d3ddc86ee802fe7b15e2c91dc160/comments/9c4dc8aa38ba0a9a6f3fbc014b05f7ba/description deleted file mode 100644 index 9918ec7..0000000 --- a/dd20d3ddc86ee802fe7b15e2c91dc160/comments/9c4dc8aa38ba0a9a6f3fbc014b05f7ba/description +++ /dev/null @@ -1,33 +0,0 @@ -There are `entomologist-data` branches "in the wild" that have issues -with tags files, so we need to migrate those somehow. - -We discussed a couple of ways to do this: - -1. Have the ent library do it automatically and transparently when it - comes across an old-style tags file. In `Issue::new_from_dir()` we'd - detect the old format, parse it, then immediately delete the file, - write the new format directory, and commit the result. - - This is good: - - It's easy for the human - - This is bad: - - entomologist will need to carry along this crufty code until we - think there are no more old-style repos in existence - - -2. Make a stand-alone tool (like the `set-done-time` tool in - . - The ent binary does *not* know how to read the old tags file, and will - error out if it comes across an issue with such a file. The human - would then manually run the tags-representation updater tool. - This tool would make the git worktree, walk the issues, find the - old-style tags, rewrite them in the new format, and commit the result. - - This is good: - - Keeps the core entomologist code clean and "modern". - - This is bad: - - This tool does manual/direct manipulation of the on-disk format, - which means it duplicates a lot of what the entomologist library - does.