Mikkel L. Ellertson wrote:
Les Mikesell wrote:
Mikkel L. Ellertson wrote:
Les Mikesell wrote:
Assume that your own application will take around a year to write/test
before deployment. Do you wait to start that work until after the
release of the OS and libraries you will be using, develop on something
you know is wildly different, or do you try to use something as close as
possible, knowing that backwards compatibility isn't taken very
seriously in the Linux kernel and distribution world?
At last, we get to the real problem - You want to develop for the
next release of RHEL, but Fedora is changing to fast for you to be
able use Fedora for that.
No, I have no problem with Fedora being used for testing and development
in any situation where being down for inderminate amounts of of time or
doing without security updates is acceptable. You are taking what I
said way out of context.
Strange - I could have sworn it was you who was complaining about
kernel updates causing a machine to not boot.
Yes, I did say that makes the system not usable for anything but testing
purposes.
I have yet to hear why
rebooting and picking the old kernel at the grub menu does not work.
I'm sure I mentioned the fact that you are doing without security
updates when you do this. Again, this may not be a problem on a test
system.
That should give you access to your data.
I can't remember now if the firewire drive issues were all fixed by
reverting the kernel or not. I think there were other problems but I
gave up on fedora on that box.
It is hard to follow your
arguments when your position keeps changing depending on what you
are trying to argue ageist. As for taking what you said out of
context, I included the entire paragraph, and replied to it. I
didn't take one sentence out of a paragraph and pretend it
represented the idea being expressed. Maybe you should be clearer in
what you write.
So you want Fedora development to change
to allow you to do that. Otherwise it is too much work to develop
your project. It must not be a user-space application, because there
you find things like compat libraries.
Sometimes, sometimes not. Suppose you had developed something as an
add-on to CIPE tunnels on the RHEL3/kernel 2.4. There's no concession
to backwards compatibility on RHEL4 and later.
Yes, some things do get thrown out, because they turn out to be a
dead end and/or a security risk. Redesigning thing when you find the
original design was not the best way to do it is part of what makes
Linux what it is.
No, copying the unix design and running all the applications that were
written for unix is what made Linux what it is. When it quits doing
that I'll quit using it.
--
Les Mikesell
lesmikesell@xxxxxxxxx
