Re: Proposed conversion strategy

[Joerg Sonnenberger <joerg%britannica.bec.de@localhost>]

On Mon, Oct 27, 2014 at 03:51:23PM -0400, Eric S. Raymond wrote:
> Joerg Sonnenberger <joerg%britannica.bec.de@localhost>:
> > On Mon, Oct 27, 2014 at 11:31:55AM -0700, Jeff Rizzo wrote:
> > > The conversion from fossil to git (which the most-used git mirror of
> > > NetBSD uses) takes nearly twice as long as from CVS->fossil.
> > 
> > The most annoying aspect for us is that I could never get incremental
> > fast-import to work reliable on the git side. Having to write all the
> > blobs on every round just for git to drop them as redundant sucks. The
> > commits are several order of magnitudes less problematic. Last time I
> > tried, git would not find objects it declared as present in its own mark
> > file. So if there is anything testing and fixing is mostly needed on the
> > git side, it is that.
> 
> I'll run some incremental-export tests here.  I've had success in the past
> not using mark files but with the new blobs in the stream and using
> the magic cookie indicating a branch join. cvs-fast-export does this
> and it seens to work.

Yeah, I wanted to avoid having to compute the git hashes for the objects
myself. Talk about smart protocol design...

Joerg