Re: The essential problems of moving from CVS

["D. Richard Hipp" <drh%hwaci.com@localhost>]

On Jan 18, 2010, at 2:22 PM, Joerg Sonnenberger wrote:

> On Mon, Jan 18, 2010 at 01:59:35PM -0500, D. Richard Hipp wrote:
> > I suppose my biggest concern with Fossil and NetBSD is how Fossil
> > might scale.  We use Fossil on projects with a 1GB or larger
> > check-out and with thousands and thousands of files.  But I didn't
> > really design fossil for truly huge projects.  How big is NetBSD?
>
> Working copy of src is around 1.7GB with 267k files and 40k  
> directories.
> Working copy of pkgsrc is around 334MB with 99k files and 27k
> directories.

The SqlLogicTest project (http://www.sqlite.org/sqllogictest) is about  
a 1.3GB checkout, but it only contains 633 files (the are mostly all  
huge 10MB test scripts).  So I know we can handle the 1.7GB project  
size.  What is untested is the 265K files and 40k directories.  I  
expect that it will work.  But if there are any inefficiencies, O(N^2)  
behaviors, or other factors about Fossil that do not scale, I suspect  
you will find them.  :-)  If you do run into sluggishness are don't  
want to give up on fossil, let me know and we can probably fix the  
problems, whatever they happen to be.

> > How large is your CVS repository?
>
> A cvs2svn trace of src from a while ago ended up with round 240k  
> revisions.
> The src tree itself is 3.9GB. The pkgsrc tree on the other hand is
> 913MB.
>
> > How many check-ins do you do per day?
>
> Typically a few dozen for each.
>
> > How many developers do you have?
>
> Maybe 20-30 developers with regular commits (e.g. typically at least  
> one
> commit per week).

These other factors are well within common practice and should not  
present any problems.

D. Richard Hipp
drh%hwaci.com@localhost