Hi Tobias,
Tobias Ringström wrote:
> Erik Huelsmann wrote:
>
>> Hi Patrick,
>>
>>> r8009
>>> A recent change in svn_auth_ssl_server_trust_prompt_func_t made
>>> the javahl binding uncompilable. This changes my implmentation of
>>> that function type.
>>
>>
>> If you follow the changes in the Subversion API, could you mention the
>> revision number of the change which you are following? There has been
>> quite a
>> large checkin by Tobias. If you are following that checkin with this
>> change, then
>> it's not in 0.35.0 yet. As soon as he creates an issue for applying
>> it to
>> 1.0, you could mention your dependency in the issue tracker.
>
>
> Thanks Erik, but this particular change is an adaption for r7975 which
> is already in the 0.35 branch.
>
> Patrick, I committed another API change in r8006 that I hope to get
> into the 1.0 branch but not the 0.35 branch. I have one more API
> change lined up that I will commit shortly. After that the API should
> be stable for a long time.
I have commited 8012, which aplied that changes to javahl. I hope both
change can be put into the 1.0 branch together.
For the next change, could send me a reminder by personal mail, so that
I can follow up with my commit as fast as possible. Or could you send me
a diff before you commit? I would answer with a diff which would contain
the matching javahl changes-
>
>>> r8010
>>> I have found a crash in the javahl module, which happened because
>>> I was passing native path styles into the svn_client_* functions.
>>>
>>> r8011
>>> A small addition the README file, which is important for building
>>> javahl on MacOS X.
>>
>>
>> These two seem like the type of change we also applied to 0.34.0 after
>> branching. I'd say they should go into 0.35.0.
>
>
> I second that (i.e. +1), but there is a lot of whitespace confusion
> going on in those files...
I have to do a lot of cleanup to do. But that is not relevant for 1.0.
>
> As mentioned last time, fixing things like this is why whe have one
> week from branching until release. Compile and crash fixes should go
> in as long as the risk of side-effects is minimal, and that is
> definately the case here.
>
> Note that everything that goes into the 0.35 branch will automatically
> end up in the 1.0 branch since it will be created from the 0.35 tag
> (when it has been created). I think it's best to wait for one more +1
> before Jostein merges this into the 0.35 branch, for formalitys sake.
> A veto from Jostein is possible of course, but not likely in this case.
>
> /Tobias
>
Patrick
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Tue Dec 16 10:04:41 2003