[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Replacing systemd in Jessie

Brian wrote:
On Tue 02 Dec 2014 at 16:52:46 -0500, Miles Fidelman wrote:

Brian wrote:
On Wed 03 Dec 2014 at 02:27:26 +1100, Scott Ferguson wrote:

Do you have a citation for this?
I'm glad you asked.
No - I "presumed" that amongst the "lots" of experts so opposed to the
late-command option, at least one of them would apply Kenshi's patch
(which apparently works) to d-i. Was my mistake an assumption that any
of them would do more than demand? (have any of them even, including
one of the noisier posters on this list who commented in that thread,
done any of the bug-tested he widely requested?)
You do us a service by raising this.

We are lead to believe there is a huge number of people who want to
preseed d-i to have a "clean install". Not one person on -user or -devel
has indicated any success with using the patch or given any detail which
would allow anyone to follow in their footsteps and test it.

Why not? Is it so difficult? Is it beyond the capabilities of a user
with technical skills? Looks like half an hour's work to me. Those who
have a vested interest in the issue seem reluctant to turn "apparently
works" into "does work" or "does not work".

Until we get some testing and substantial feedback, using this patch to
beat the anti-systemd drum should be seen as noise.

Well, actually, it does involve a little more than downloading
debootstrap, applying the patch, and compiling.  One has to build a
custom copy of d-i to actually make use of it.  That's  bit of a
complicated procedure.
Why does debootstrap have to be downloaded and a custom copy of d-i
built? Suppose the patch were applied to debootstrap in a running d-i.
Why wouldn't that be sufficient for testing?

That's a straightforward technical question, incidentally.

Also a straightforward technical question:  How would one actually do that?

It looks to me like d-i starts up, then pulls in debootstrap and starts it running. I've spent a little bit of time looking at how d-i figures out where to look for debootstrap, and got to the point of concluding that I either had to put the patched debootstrap in the repo (not going to happen) or build a customized d-i that looks for the patched debootstrap somewhere else. That's the point at which I decided that, not being much of a c coder, I really didn't want to mess with things.

If you have a straightforward suggestion, please....

                         Personally, I'd rather wait for the
installer team to fix a bug that has rather broad implications.
In other words, you'd rather not know whether the patch works (whether
the procedure is complicated or not). It cannot be that important to you

No... I kind of figure that deboostrap should actually work properly, and that as a core piece of system code, the maintenance team should care enough to actually fix this kind of bug. It was sitting there long enough. That, in itself, is a datapoint for me in making decisions about future reliance on Debian.

I also seem to recall seeing at least one report of someone who'd
done the test.
Not what I would call substantial feedback.

True enough.

Miles Fidelman

In theory, there is no difference between theory and practice.
In practice, there is.   .... Yogi Berra

Reply to: