[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Development environments
I'm curious how often customers (that is, all you guys)
end up booting their (your) machines, and what you boot them for.
I end up booting my machine (or finding it rebooted) about three
times a week. Since other people may use the machine when I am not
around, it gets rebooted for insurance against unknown things that
other people may done to the environment. During the week I leave
it logged in (somewhat anti-social behavior) and make anyone who wants
to use it reboot, because (I agree) it is nice to everything loaded
when I arrive in the morning.
With the tendency to look at a continuing environment, there are
problems that loom larger and larger. Over the past couple of months
we have had power glitches that required all of the Lisp machines to
be rebooted. Fortunately, I didn't lose anything but I'm getting
worried that one of these days I'll lose a lot. When I'm making changes
to a system it can take several days to get the changes completed,
but there doesn't seem to be a good way to save intermediate changes
and still interact with the patch system in a reasonable way. My
compromise has been to save everything at the end of the day and
generate a patch file but not finish it until I'm really finished.
The result is that I get lots of error messages from multiple
definitions when I load the patches. Is there any way to generate
a patch file based on the differences since a specific time (default,
time of last patch), rather than just the differences in unsaved files?
Are there other ways to handle evolving systems that eliminate such