chore(deps): updated go and github actions dependencies by in #511.fix: saving bookmarks inconsistencies by in #500.fix: sharing context between api and goroutine caused goroutine to fail by in #499.fix: remove CreateNewID usage from api handlers by in #486.fix: data dir permissions having execute for all by in #493.fix: import should respect add_date or last_modified fields by in #490.fix(psql): save bookmarks not using passed bookmark id for the insert by in #484.fix err in keyword searching by in #494.fix(import): check for no rows error by in #487.refactor:pass ctx to db layer, return err instead of recover panic by in #450.docs: fix remember represented as a number instead of a bool by in #472.fix: remove manual transaction on sqlite migration 2 by in #468.Improve performance for sqlite queries by in #425.Fix docker migration script example by in #451.feat: session expire time from backend by in #437.fix invalid assets path when use webroot by in #448.
If anything goes wrong, you will get an error on the output which you can use to open an issue and your database should be left as it is. In the end, and to avoid adding more complexity to the tool (which is one of the core features) I decided to rollback that change, meaning that everything should be done in the background while you use or serve Shiori.
What does this mean? Basically we added proper migrations into Shiori and added a custom command to perform them, but since our documentation is lacking in some ways most users where a bit confused about what a migration meant, specifically users of the CLI tool which do not rely on the server. After some discussion over #473, we've decided to roll back the change introduced in v1.5.3 that made the user take care of migrations manually.
0 kommentar(er)
