[RFC]: preparing for Blob storage

[Martin Pool]

On 26 May 2006, John Arbash Meinel <john at arbash-meinel.com> wrote:

> Well, since we just had a stable release, I think we can do the 'get it
> into mainline so it doesn't bit rot'. Unless you object, that is how
> Aaron and I are planning on handling Changesets.

Yes, I think we should do that for both changesets and the refactoring
Robert discussed. 

> But it depends how invasive it is going to be. If it was only minimally
> invasive, a feature branch can be maintained without too much
> difficulty. I think we would need the Revision api changes to be in
> core, though. So that plugins, etc can realize that they shouldn't be
> using those functions.

It's not too hard to maintain them, but bitrot does occur and there is
some risk that they'll get lost or the work will need to be duplicated.
So, more than in the past, I would lean towards landing preparatory work
if it goes in some useful direction and passes the normal review
criteria (doesn't degrade code quality or design clarity, and increases
test coverage).

-- 
Martin