Design decision: use bzr commit --local ?
Bug #240670 reported by
Daniel Holbach
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
five-a-day |
Fix Released
|
Medium
|
Unassigned |
Bug Description
I'm pondering using
bzr commit --local
for the cases where we don't push to the branch (after one hour of not pushing).
I realise that the code path will be a bit longer for local commits but it has obvious advantages:
1) less conflicts because of using 5-a-day on various machines (imagine two .5-a-day-data directories with unpushed changes)
2) commit --local is guaranteed to work and if the push (or 'commit' in this case) fails, we have a local state that makes sense
Can you let me know what you think?
Changed in five-a-day: | |
importance: | Undecided → Medium |
status: | New → Confirmed |
Changed in five-a-day: | |
status: | Confirmed → Fix Released |
To post a comment you must log in.
Well, this bugreport is surprising me. I always thought that 5-a-day is doing a local commit if the last commit was done by the same user within one hour, but obviously I'm wrong.
I'm not using bound branches that often, so I also don't have much experiences or knowledge about local commits, but I can't see any cons against using 'commit --local' in this cases.
Markus