Re: [Savannah-hackers-public] bzr access problems

[Tim Cross]

On Mon, Oct 17, 2011 at 3:35 PM, Tim Cross <address@hidden> wrote:
> On Mon, Oct 17, 2011 at 3:02 PM, Martin Pool <address@hidden> wrote:
>> On 17 October 2011 14:56, Tim Cross <address@hidden> wrote:
>>> I will talk to our network guys and see what I can find out. However,
>>> it would seem weird if its my network as I don't have any problems
>>> accessing bzr repositories anywhere else - for example, I updated from
>>> a launchpad repository today without any problems. It seems to be
>>> restricted to just repositories hosted on savannah.
>>
>> Launchpad doesn't speak bzr-over-tcp ('bzr://'), only bzr-over-ssh and
>> http.  So if there is a network or firewall problem I wouldn't be
>> surprised if the behaviour was different.
>>
>> If you have an ssh account on Savannah can you try using that?
>>
>> m
>>
>
> True, but bzr pull http:// fails on savannah, but works fine on launchpad
>
> I don't have a savannah ssh access.
>

I need to correct my previous assertion regarding http.

I have two copies of emacs from bzr. One Iwhere I use to use
http://bzr.savannah.gnu.org/r/emacs/trunk and one where I use
bzr://bzr.savannah.gnu.org/emacs/trunk.  Currently, I cannot do a pull
in either of these copies. The http based one returns an error about
bad permissions and the bzr version just returns a general error about
permissions/connecitn/bugs.

I had tried a fresh bzr branch bzr://, but with no luck. However, I
had not tried a fresh bzr branch http://, which I just did and it
appears to be WORKING - at leas, it has run a lot longer than any
previous branch commands for the last week and looks good.

So, it would seem that my previous copy created with bzr branch
http:// is somehow screwed up - perhaps unrelated to other issues or
perhaps I got confused and did a pull on top of it over the bzr://
protocol and screwed it up. At any rate, branching over http is
looking like it will work (over half way so far). This would seem to
confirm that it is an issue with the bzr:// protocol that is the
problem. This may help with examination from our network perspective.

I also wanted to mention that I've upgraded from 2.3.4 to 2.4.1 and
see the same issue (just for the record).

Tim
-- 
Tim Cross
Phone: 0428 212 217