On Mon, Apr 05, 2004 at 05:36:22AM -0400, VaX#n8 wrote:
> Summary:
> What's the point of sup+import?  The "Tracking current" pages say I can always
> get an unmodified source tree with that method, but I could always do a full
> anoncvs checkout.
	sup has been around a lot longer than anoncvs has been.

> Scenario:
> I used to have a CVS project called "etc" with all my config files.
> I don't think I have a 1.1.1 branch on it, I think I just created it
> out of the blue one day.  This makes it a hassle to merge in changes
> to NetBSD, which I do manually.  I want to start basing my stuff off
> NetBSD so that I can automatically merge changes.
> 
> Can I automatically merge changes with the anoncvs method?
	You could do a similar thing with anoncvs, but to do anything reasonable
you'd still need the import step.  You would just replace the sup step
with a cvs checkout from anoncvs.  (or cvs update)
	If you don't want something quite as structured, you could get away
with having a checked out source tree that has your changes sitting as
local modifications.  Then you'd only need to do a "cvs update" and you could
skip the import step.  Of course, you wouldn't be able to keep track of 
any history of your own changes.

> I read the document on tracking -current and I decided to do sup+import.
> I'm having major problems with sup timing out 90% of the time.
> This has made me wonder, is there any real advantage to sup+import?
> In the "tracking -current" page they mention I can always build a unmodified
> source tree.  But I could always do a clean anoncvs checkout, right?
	Using sup _should_ be faster, since it doesn't have as much overhead
as cvs.  Either way will get you the same source tree.


eric