The Jelmer "fork" of bzr.dev

[Jelmer Vernooij]

Robert Collins wrote:
> On Sun, 2009-04-12 at 13:50 +0200, Jelmer Vernooij wrote:
>   
>> Robert Collins wrote:
>>     
>>> On Sun, 2009-04-12 at 09:31 +0100, Russel Winder wrote:
>>>   
>>>       
>>>> The appears a strong prima facie case for reverting the decision to
>>>> remove svn+http.
>>>>     
>>>>         
>>> I think having svn+http present is useful for cases such as these;
>>> autodetection is great too, and they are not exclusive as far as I can
>>> tell.
>>>   
>>>       
>> svn+http:// doesn't bypass the other detection though, it just happens
>> to not implement all methods on Transport (at the moment) causing the
>> bzr format probes to fail in different way than 401s.
>>     
>
> It won't ever implement get_smart_* though, will it?
>   
It might. It's not unlikely that SvnHttpTransport will subclass
HttpTransport in the future, and add a couple of extra methods to
support DeltaV HTTP requests. Or alternatively, SvnHttpTransport might
go away altogether if webdav support is integrated in bzr core.

Also, if Google happened to start returning 401's to GET or HEAD
requests (neither of which are required for svn access) just like they
do for POST things will still break, as attempts to access .bzr/format
will fail.

Cheers,

Jelmer