On Tue, Oct 7, 2008 at 9:59 AM, Greg Stein <gstein_at_gmail.com> wrote:
> On Tue, Oct 7, 2008 at 5:58 AM, Mark Phippard <markphip_at_gmail.com> wrote:
>>...
>> I am not sure the results would justify the effort. The benefit of a
>> framework would only be realized if several OSX Subversion clients
>> were going to use and share it. I am skeptical that would happen
>
> Um... that seems like false logic. If we package it as a Framework,
> *then* applications can start to use it.
>
> If we never package it as a Framework, then applications will *never*
> be able to use it (easily).
>
> Some time sunk into creating a Framework, in return for saving a lot
> of downstream users the time to use stuff right from our build? That
> seems like a win to me.
Perhaps using the Framework is easier, not sure, but since an
application cannot rely on it existing I'd expect each app to build
and ship their own framework. So I just do not see the benefit.
>> unless they were all also going to agree to work against the binaries
>> provided by CollabNet. The Apple docs you linked to seemed to suggest
>
> "provided by CollabNet" ... um... so? What about binaries provided by others?
I am not trying to be controversial here, I really do not care where
they are from. My point is that if there is not a canonical
distribution, and currently there isn't, I do not see the benefit.
Neither commercial or open source OSX apps are likely going to want to
direct you to download a package and install it. I suppose Apple
would eventually pick up these changes in the OS, but we know they are
not going to stay current with SVN releases so I do not see people
using their version either.
>> that using a Framework does not make sense unless there are going to
>> be applications build that use it. If it is just the command line
>> then it is unneeded overhead.
>>
>> Those same docs also discourage the use of Umbrella Frameworks and
>> seem to claim that is not the way to go.
>
> Hmm? Was Jeremy suggesting an Umbrella? It sounded like just one
> Framework Kinda unclear here, but I don't see how providing more
> options to downstream users can be a Bad Thing.
Jeremy said he envisioned using an Umbrella Framework.
I am not saying that a framework would be bad, although the Apple docs
imply there is some marginal cost to using them. I am saying the
improvements by doing this are likely to be small and it probably
involves a lot of work to get our build system modified to do this. I
do not see the value. Jeremy works on SVN on his own time and has two
young kids. I'd rather see him use his talents on areas that bring
more value. If this is something he is motivated to do, as he is with
the new Python bindings, then great he should go for it. If he is
asking for opinions, I am saying I do not see the value of this
justifying the effort. If he wants to make SVN a great OSX app, there
are certainly much bigger problems than this, such as the Unicode
encoding problems and the WC redesign you are working on.
--
Thanks
Mark Phippard
http://markphip.blogspot.com/
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe_at_subversion.tigris.org
For additional commands, e-mail: dev-help_at_subversion.tigris.org
Received on 2008-10-07 16:11:36 CEST