Merge lp:~mbp/bzr/2.0-stat-symlink into lp:bzr/2.0
Proposed by
Martin Pool
Status: | Rejected |
---|---|
Rejected by: | Martin Pool |
Proposed branch: | lp:~mbp/bzr/2.0-stat-symlink |
Merge into: | lp:bzr/2.0 |
Diff against target: |
209 lines (+113/-5) 5 files modified
NEWS (+5/-0) bzrlib/tests/per_transport.py (+39/-1) bzrlib/transport/__init__.py (+13/-1) bzrlib/transport/local.py (+31/-2) bzrlib/transport/sftp.py (+25/-1) |
To merge this branch: | bzr merge lp:~mbp/bzr/2.0-stat-symlink |
Related bugs: |
Reviewer | Review Type | Date Requested | Status |
---|---|---|---|
Martin Pool | Disapprove | ||
John A Meinel | Needs Fixing | ||
Review via email: mp+30210@code.launchpad.net |
Description of the change
Backport transport symlink support so that we can fix bug 32669 in 2.0.
To post a comment you must log in.
Unmerged revisions
- 4757. By Martin Pool
-
stat on a transport pointing directly to a symlink returns information about the symlink
- 4756. By Martin Pool
-
Raise TransportNotPos
sible when sftp server refuses to make symlinks - 4755. By Martin Pool
-
Backport transport support for symlinks
More justification: to fix bug 32669, we need to be able to tell whether A is a symlink, before we try to open it as a bzrdir. In 2.0 without this fix, you can see symlinks in the workingtree and locally, but there is no concept of creating or reading symlinks over a transport, which is the way we open bzrdirs/branches.
So there are two changes here:
* give transport readlink, hardlink, symlink
* Transport.stat should tell you about the link, not the target
This is getting a bit large for 2.0, but I think still reasonable and not too risky, because it's mostly adding new interfaces.