check that get_data_stream distinguishes annotated and unannotated knits
Bug #165106 reported by
Martin Pool
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Bazaar |
Invalid
|
High
|
Martin Pool |
Bug Description
(mentioned by Robert)
Some knit files contain line annotation data, and some do not. This is not stored in the knit data but rather implied by the repository format and/or the identity of the knit.
We should make sure this is always taken into account when fetching a data stream from the smart server, to be sure it doesn't cause problems when pulling from knits to packs or vice versa.
(Follow-on from bug 164637)
Changed in bzr: | |
importance: | Undecided → High |
milestone: | none → 1.0rc1 |
status: | New → Incomplete |
Changed in bzr: | |
assignee: | nobody → mbp |
status: | Incomplete → Confirmed |
To post a comment you must log in.
Robert says he checked the code, and you do in fact get an error pulling from knits to packs or vice versa. get_data_stream generate a format that says whether it's annotated or not.
It looks like the client needs to detect there's no suitable data stream, and either transcode into an appropriate stream, or use vfs operations.