On Dec/13, Bob Proulx wrote:
> Therefore I don't have a good idea of what to do here. I only know
> that it is an impossible system. I feel certain this can't be
> necessary.
While I appreciate your concern, and am also pained by seeing so many
versioned conflicts, what you *feel* is unfortunately not relevant. Let
me try to explain again why.
Upstream declares conflicts on many "recent" versions of the libraries
it depends upon. For each one of those, there are 2 cases:
1. the versioned conflict is legitimate, and mitmproxy would indeed
break if used with a library not satisfying the conflict. In this
case, there is nothing we in Debian can do, except write a patch to
support the new API exposed by the more recent version of the
library.
2. the versioned conflict is not legitimate, and mitmproxy works fine
with a newer library version. In this case, the versioned conflict
in the Debian package can be relaxed.
Whether each one of the mitmproxy dependencies falls in the first or the
second case is most definitely not a matter of feelings, but instead one
of manually checking. For each dependency. I certainly welcome help in
that department.
> Just today there are two more python packages that have been uploaded
> breaking the << dependencies of mitmproxy again. This is inevitable.
> Do you wish us to simply keep filing individual bug reports each time
> this happens?
Yes, please: until I have more time, I intend to look at breakages *as
they happen*, and see if it can be fixed accordingly for mitmproxy to
remain in stretch.
Cheers,
--Seb
