sftp locks can get stuck

[Robey Pointer]

On 23 Dec 2005, at 5:59, John A Meinel wrote:

> Robey Pointer wrote:
>>
>> I think simpler is better here:  We record tracking info in the write
>> lock (email, IP, timestamp) and provide a way to explicitly break the
>> lock.  Maybe if an operation fails because of a write lock, and that
>> lock is pretty old (by some arbitrary standard), bzr could hint to  
>> the
>> user:
>>
>>     The branch was locked by <robey at lag.net> 5 hours ago.
>>     If the lock is stale, you can break it with: bzr break-lock
>>
>> But no magic of trying to guess if the lock should be automatically
>> broken, etc.  Stale locks should be relatively rare.
>
> I'm +1 for adding the username, and not automatically breaking the  
> lock.
>  I don't think the ip address helps much, but it is easy to record. I
> don't think we need the timestamp, because we can just stat() the lock
> file to figure out what time it was taken out. I suppose we could add
> the UCT timestamp to help with handling time-zone shift.

Exactly, yeah.

So probably what we should do is move the locking outward (to  
Branch?) and have Transport just provide something like

     create_lock_file(path, text_content)

and raise an exception if the lock file couldn't be created (via  
O_EXCL or whatever the transport's equivalent is).

If that sounds okay I can probably cook up a patch.

robey