Re: Best practices for development workstations
John Goerzen <firstname.lastname@example.org> writes:
> 1. workstation running sid
> I've followed this model for over a decade. It works well, in general,
> and I keep up with development well enough that I can fix problems when
> they arise. However, it tends to lead to a certain amount of cruft over
> the years. Moreover, it's not really appropriate for a laptop or a
> situation in which Internet access isn't readily available to fix
> problems. I'm hoping to move away from that model.
> 2. workstation running squeeze or lenny
> I'm enjoying squeeze on my laptop and am considering switching my
> workstations to it. That, of course, means I will still need a sid
> environment somewhere for building. It also means that there are times
> when the packages I build for sid will not be installable on my
> workstation, which could be annoying. On the other hand, breakage
> should be more rare.
I use a combination of the two of these except with testing on my primary
workstation (the one that I can't afford to have go down), but list and
deprioritize sid in sources.list on the workstation running testing. Then
when testing builds of my own applications, I let apt resolve dependencies
from unstable where needed.
I've never had any trouble with running sid on my laptop other than
occasionally finding bugs (which is one big advantage of doing this; I
find and report bugs). I just don't do upgrades while I'm travelling. I
wait until I'm home again and have access to other systems and ways of
repairing the box.
> 2a. pbuilder
> pbuilder, or some other chroot such as schroot, can help. In theory, it
> is a good plan. I don't have to dedicate a lot of RAM to it. The
> problem is that a chroot doesn't establish terribly strict separation
> from the main environment. Despite promises to the contrary, I've had
> weird things happen; for instance, an MTA, database server, or some
> other daemon process might try to fire up from within the chroot, which
> can result in highly confusing situations. I am therefore somewhat
> uncomfortable with this prospect.
I do all my builds with pbuilder and cowdancer, but don't do testing in
that environment. I've never had problems with weird things happening
during the build itself.
Russ Allbery (email@example.com) <http://www.eyrie.org/~eagle/>