Re: RFS: mdadm
On 09.09.2011 01:16, Tim Gokcen wrote:
> I have uploaded a new version of mdadm which, on top of the previous fixes, also fixes the FTBS bug 625392 and bug 63380. I am therefore still/again seeking a sponsor for my NMU of package "mdadm".
>
> This upload is a fix for two trivial bugs opened against the current mdadm package, as well as an FTBS in unstable and a small bug in the kernel version detection system that could prevent boot (on non-Debian systems).
What is your intended target for this fix?
I think that for a squeeze point release (if we target that), the package
should have minimal difference - like the two bugs mentioned. If you want
to target wheezy/sid, my guess is that it is way better to start with a
more recent upstream version instead of doing cosmetic NMUs.
You included some stuff in your package which is not mentioned in the
changelog - notable some very questionable mdadd.sh thing (which comes
without a manpage too, but with some inline docs), and alot of documentation
about various raid levels, superblock formats and whatnot. I haven't done
a review of this yet, but a good review of mdadm.sh thing is definitely
needed.
Also, you have some debian/.pc/ leftovers in your package, which should not
be there.
But that's all mostly cosmetics still. The main question is the first one,
which is your target. Because the currently offered package does not fit
stable and does not fit testing/unstable. Just IMHO anyway.
I hope I will collect a small NMU for squeeze later today, which includes
just the fixes for the two annoying bugs mentioned plus removal of
-Werror option from makefile - so it will be buildable on gcc-4.6 too,
without a more intrusive patch for the source.
Thanks,
/mjt
Reply to: