Story #1150
Updated by bmbouter over 9 years ago
As This feature should be available on both a user, I can choose server or a node. It would allow the server or node to sync act as a repository lazily. When caching proxy. For now, lets call this lazy syncing is enabled for a specific repo, any associated metadata will be synced and published before any units are downloaded. If there is no metadata (ie: Python) then loading. In this model, the sync+publish will be server (or a no-op. Lazy syncing will be an optional feature and will not modify the normal sync workflow at all. Lazy sync will be configurable as a per-repo setting. Lazy sync will node), would have two modes: * lazy sync passive - This mode creates recent copies of the lazy repo and will metadata. However, it would not fetch content units until an actual client requests them. * lazy actually sync active - This mode creates the lazy repo and fetches content units using an asynchronous task. Missing content units package from the feed location until they are fetched lazily as in passive mode during requested by a client. This would allow for only the asynchronous sync so the repo can be used right away. This is being introduced as a platform feature so all content types will have support. This is not being integrated with nodes. This is a tracker story used to be pulled down to the file system and tracks for repositories to be created by only downloading the required work as child tasks/stories/refactors. metadata.
.