[Insight-developers] Data submodule was reverted again ( staging check needed? )

[Bill Lorensen]

Brad,

I agree that history is a big problem introduced by our switch to git.

But the current setup with data as a submodule adds complexity to
checkins and is subject to unexpected abuse as recently reported by
Brad L. The midas absolution adds even more complexity, especially for
baselines.

I'm looking forward to a solution that keeps the footprint low but
keeps the workload on a developer at or near what it was in ITK 3.

Bill

On Tue, Jan 25, 2011 at 6:09 PM, Brad King <brad.king at kitware.com> wrote:
> On 01/25/2011 02:40 PM, Gaėtan Lehmann wrote:
>> I still can't see, nor did I get a good justification when I asked for
>> it on the mailing list, to have moved the data to a separate git
>> submodule.
>
> It's in a submodule because the size of the ITKData repository is a
> significant fraction of the source code size of ITK.  Assume for a moment
> that all data are in the source tree.  If a big file is added at any point
> in history and then removed, suddenly everyone's Github forks of ITK.git
> need space for that object in history forever and ever.  Without the data
> in the tree everyone can push around a much smaller repository.
>
>> I think we should discuss that at the meeting in Boston.
>
> I've been making progress on this front.  Someone from Kitware will talk
> about it in Boston.
>
> -Brad
>