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

Re: Bug#776346: libpcap: build udeb for use in debian-installer



Hi Romain,

On Mon, Sep 7, 2015 at 6:52 AM, Romain Francoise wrote:
> libpcap is special because it's very close to the kernel; once it gets
> used in the d-i environment the burden of keeping it working there will
> fall on me.

It's fortunate that the udeb design is so simple then.  Each
effectively mirrors its equivalent deb package, so in almost all
imaginable cases a bug in one will be present in the other, so d-i
specific bugs have been surprisingly uncommon, although yes not
totally unheard of.

> In the past I've had very little help from the porters with
> kfreebsd issues (like #750836, #626232)

It is unfortunate that #750836 took 3 months, and didn't seem to get
the respect that you thought it deserved.  If I had to interpret that
situation, it looks like there was an unintentional communication
breakdown.  Since the workaround in libpcap was implemented so quickly
it appeared on the kfreebsd side that it was no longer a significant a
problem, so it wasn't given the severity you thought it deserved.

It looks like #626232 was fixed so rapidly (less than half a day) that
the porters didn't even have enough time to be needed.  It also looks
like Jakub Wilk solved that one, so it looks like you aren't totally
without help and interest from other developers.

So in conclusion, at least for these specific bugs, help did come
along, but not urgently enough to satisfy your expectations.

> So yes, in itself adding the udeb is trivial. But I'm not sure I want a
> udeb in the first place.
>
> Let me also state that NMUing a package without warning for a wishlist
> bug, even to DELAYED, is considered rude in my book.

Fortunately devref quite clearly disagrees with that sentiment [0],[1].

The NMU is now cancelled; a social stalemate.

I would like to try to find an amicable way to get the change
included.  Do you have any suggestions that could make it more
palatable from your perspective?  Specifically can anything be
improved based on technical merit, and what social changes do you
need?

Best wishes,
Mike

[0]https://www.debian.org/doc/manuals/developers-reference/ch05.en.html#kind-to-porters
[1]https://www.debian.org/doc/manuals/developers-reference/ch05.en.html#nmu-maintainer


Reply to: