Neohapsis is currently accepting applications for employment. For more information, please visit our website www.neohapsis.com or email email@example.com
Date: Thu Aug 19 2010 - 11:21:09 CDT
Quoting Luis Useche <usechegmail.com>:
> Hi Guys,
> I have been meaning to follow current for a couple of weeks now. I read the
> "Building Sources" page and it seems like I should follow the process of:
> cvs up src xenocara ports -> compile -> install, where install includes
> merging of configuration files. Moreover, I should also keep an eye on the
> "Following -current" webpage for any change I should make. This looks like
> lot of work every-time you run cvs up (mainly the compilation of ports and
> merging of conf files).
> I was wondering how do you usually work on current and if you all follow
> this process through-fully. If not, what kind of tricks do you use to make
> the process easier.
> For now, I am using snapshots with binary packages.
> Thanks in advance,
It really is't that hard to do. If you follow the examples for all this
in the release man page, it is straightforward, and when you
crash into something odd, it's at least a 90% chance that you
botched something, as opposed to the tree being borked.
Subscribing to the src changes list is really very very usefu.
This lets you see the changes to the system at the atomic level,
and in time you'll recognize the stuff that needs only a kernel
recompile, or even just a part of userland to be rebuilt. At first
though, rebuilding eveything is the right way to go.
It really isn't a lot of work. If you run script(1) when doing a cvsup
you can go back to that and see whats new, which might be a
big help in determining if you want to rebuild stuff or not. Depening
on whats being changed the most in a given week, you might
not bother rebuilding things for a while. Other times, a single
change to a library where the major number is bumped you'd
want to rebuild everything.
A certain amount of experimentation on a spare machine is a
great way to blow things up and then let you figure out how to
repair stuff in a non pressure-cooker environment. By taking
the time to learn and messing things up you'll be in much better
shape for real systems that you depend on.
A door-stop 400MHz Dell with 256M is a great system for doing
with with if you're on a budget. Building your own packages is
harder, in that you can screw up in lots of ways. You also want
as fast a machine as you can get for that. My 2.1GHz package
machine takes about 60 hours for a run. But before you do that
learn about OpenBSD proper first.