Merge lp:~bzoltan/click/vivid-transition_mirrors into lp:ubuntu/vivid-proposed/click
Proposed by
Zoltan Balogh
Status: | Merged |
---|---|
Merge reported by: | Michael Vogt |
Merged at revision: | not available |
Proposed branch: | lp:~bzoltan/click/vivid-transition_mirrors |
Merge into: | lp:ubuntu/vivid-proposed/click |
Diff against target: |
100 lines (+28/-5) 2 files modified
click/chroot.py (+26/-3) debian/control (+2/-2) |
To merge this branch: | bzr merge lp:~bzoltan/click/vivid-transition_mirrors |
Related bugs: |
Reviewer | Review Type | Date Requested | Status |
---|---|---|---|
Michael Vogt | Approve | ||
Review via email: mp+246913@code.launchpad.net |
Commit message
Transition of locally used mirror servers to the chroots.
Description of the change
Transition of locally used mirror servers to the chroots.
To post a comment you must log in.
Some explanation on the MR:
I have chosen to use the local ports and archive mirror because other options seems to be less optimal:
a) There is a way to measure bandwidth to/from the apt's mirror support, but that give a geographically close mirror and not a mirror with a fast access
b) Measuring the bandwidth to mirrors and trying to find the most optimal would need way much more code and would be too complex for a fairly simple problem
I decided to use a strict and limited expression to find the locally used mirror. I accept only the /XX.archive. ubuntu. com $series main/ source line. In case I do not find the mirror the code falls back to the central archive what is a safe choice.
All in all, the job here is to make the Chinese app developer's life easy. I am sure that we can make a more generic and smarter solution to support Ubuntu mirrors, but this one I believe is safe and good enough for now. Also I think it is legit idea to use the same mirror in the chroots what is used on the SDK's host environment.