[09:19] <stefanct> apt-get source gitg ; sudo apt build-dep gitg ; cd gitg-3.17.1 ; debuild -b -uc -us -tc
[09:19] <stefanct> let's suppose gitg-3.17.1 is the directory that apt-get source gitg creates...
[09:19] <stefanct> how could that fail with wrong versions of a library?
[09:19] <stefanct> the concrete example is failing for me in xenial
[09:20] <stefanct> Requested 'libgit2-glib-1.0 < 0.24.0' but version of libgit2-glib is 0.24.0
[09:21] <stefanct> (commands paraphrased from memory... but that's essentially what i tried to do: build gitg from xenial from source)
[12:12] <elbrus> rbasak: Not sure if upstream can help here, as I don't want to ask them to check the cacti code.
[12:13] <elbrus> rbasak: One remark I'd like to make: I can see why changing the sql_mode is something good from MySQL perspective, but I would have expected the package/team in Ubuntu to handle it better, i.e. either a heads-up to the maintainers, or a grace period. Especiall the short period in Xenial was extremely bad for such potentially big impact change.
[12:14] <elbrus> rbasak: no offence, but both the php 7.0 change and the mysql change cought me by surprize and were both late
[12:14] <elbrus> and disruptive
[20:42] <rbasak> elbrus: I agree. Sorry. We should have communicated it better.
[20:42] <rbasak> elbrus: I'm also not sure we should have done the MySQL switch in hindsight.
[20:43] <rbasak> elbrus: I think I trusted tests too much. We don't have enough coverage to do that.