Re: [Monotone-devel] New cvssync.attrs branch

[William Uther]

Welcome back :)

On 20/03/2007, at 1:29 PM, Christof Petig wrote:

> Unfortunately my laptop broke down, Siemens/Nixdorf was unable to
> repair/replace it and I did not get around to buying a new desktop for
> the refund. My current machine takes half an hour to compile mtn  
> (orange
> iBook).
>
> Combine that with a new job, a stage play and some family parties and
> you see why cvssync development has slowed down a bit.

:)

You mean Monotone isn't your entire life *shocked face*.  :)

> I'm perfectly fine with William fixing/finding problems in a shared
> infrastructure (most of the contents of mtn_cvs are not cvssync3
> specific) and trying out Nathaniels proposal. Perhaps cvssync4 is
> superior to cvssync3 ;-)

Well, we'll see.  At the moment the .refactor branch has some major  
problems with the sync encoding.  It only works at all for the small  
cases through blind luck.  Once you get a few revs in a row that it  
has to pull across, it fails.  See the push_pull_with_changes test  
case I committed.

> There's also partial pull crying for some spare time ;-)

You may be better off spending your time there.  I'm getting to know  
the cvs-sync code better, and although I personally need it, I don't  
consider it a high priority for mtn in general - certainly not as  
high as partial pull.

It would be cool if you could have a brief look over my changes and  
see if I've done anything majorly stupid, but that's up to you.  I  
also don't understand how the takeover stuff works.  I was about to  
rip it out - it is failing tests and I couldn't be bothered fixing  
it.  If you want to keep that, then you could look at it and point me  
in the direction of what might be failing.

Be well,

Will          :-}