Source code control

Eric W. Biederman ebiederman at lnxi.com
Mon Jan 5 17:10:01 CET 2004


"Bryan O'Sullivan" <bos at serpentine.com> writes:

> Not to bang the BK drum, but I'd hate to see the arguments for or
> against it misrepresented.

I agree that the way I stated the differences BK didn't sound too good.

> On Mon, 2004-01-05 at 06:15, Eric W. Biederman wrote:
> > - Distribution.  All you have to do to export an arch repository for
> >   all the world to see is to point an ftp or http server at it's
> >   directory.
> 
> BK has a built-in server that provides the same facilities.

Nice but not quite as convenient, as needing no special server.
Sort of the plain text versus the binary file argument, until it is a
performance bottleneck the extra convenience has a lot of nice
secondary side effects. 

> BK is pretty slow on a large tree.  However, it's several thousand (!)
> times faster than arch in most cases, which is by design just abominably
> slow at even the most trivial of common operations.

That arch is slow is something I will look into.  Although it should
be able to trivially be beat sourceforges anonymous CVS servers...

> On the other hand, there's a lot to be said for open source software in
> this realm.  
Right.

> The flip side is that arch is about a decade away from being as
> performant and mature as BK is today.

If the performance is not good enough to be usable we won't go there.
And if it is usable but annoying it can be fixed.

Eric



More information about the coreboot mailing list