Bug#635504: ITP: flashcache -- write back block device cache for Linux
-----BEGIN PGP SIGNED MESSAGE-----
On 18.09.2011 16:25, onlyjob wrote:
> When you mention technical problems, I wish you could be more specific.
You made some semantic errors Lintian can't detect in its purpose as
static analysis tool. As I said, I didn't extensively check your
package. Among the most obvious things are:
* For example the VCS fields do not denote upstreams VCS, but the
packaging VCS in Debian.
* You provide no mechanism to setup a flashcache volume on boot.
* You are not supposed to copy the DKMS script into your package, but to
invoke them at runtime.
* You close an ITP bug you don't own ... mine.
* Code of Flashcache is GPL-2 _only_ but you release your Debian
packaging as GPL-3+. That will cause trouble as soon as you need to
merge patches (at least).
> I don't quite see the point in merging something working into
> something what doesn't work, and not the other way.
Fair enough. Please note this is not how we work in Debian, though. I am
the owner of the ITP bug and I made visible progress since I filed the
ITP. You will very likely have a hard time in finding a sponsor while
trying to hijack someone else's packages.
This does not give you very much alternatives to my offer, besides of
stepping down as maintainer (in Debian) completely. Your approach wasn't
exactly polite and yet I offered you to form a team together with me.
> (With all due respect there are more than some minor things you didn't
> address i.e. your package is not usable yet even though you did a lot
> of work.)
I didn't say my package would be ready to use. That's the reason why I
didn't try to introduce it to Debian yet. However I believe, the package
is (sort of) usable as is. If you find some problems, I am eager for
letting me know.
> For practical reasons giving up working package may be not the best
> option for people who might need it.
Merging is not exactly the same as giving up. Also, Debian is not only
about working packages, but also about quality packages, and that takes
time. Something which "just works" somehow is trivial to achieve.
Note, by that I didn't want to imply at all that your package would not
be a quality package. I am just not confident enough about mine yet, as
I know it lacks certain features including but not limited to tested
init.d support and manpages.
with kind regards,
IRC: daemonkeeper on Freenode/OFTC
GnuPG Key-ID: 0x9D80F36D
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
-----END PGP SIGNATURE-----