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

Re: Pandas new version



Yaroslav Halchenko <debian@onerussian.com> writes:
> On Tue, 05 Feb 2019, Ole Streicher wrote:
>> "Rebecca N. Palmer" <rebecca_palmer@zoho.com> writes:
>> > Has anyone checked whether this would break pandas' reverse dependencies?
>
>> I didn't yet. I just tried to update the packaging to 0.24, which
>> however has a number of test failures, which would need to be discussed
>> with upstream first.
>
> if not sever - I guess could be patched to be skipped for now and then
> patches picked up to address them?  or it was too many/too deep?

If you want: please do it as you like. My own workflow is to first
discuss all failures with upstream, especially when I can't estimate
the impact - sometimes it may be even fault of the package. There were
about ten failures, which makes this already quite an effort. Especially
since I did not work closely with upstream so far (so I don't know them
well, and they don't know me) -- this communication should IMO be done
by the regular maintainer. And, as I wrote, the package rules are a bit
complicated, so I don't want to touch them before they are simplified
(which would add more efforts on top of that). Bringing the other stuff
to gbp standards (pristine-tar etc.) even more.

And it also does not solve the problem that updating can introduce
regressions in reverse dependencies, which is not the best thing we can
do just before freeze. But finally you are the maintainer, if you think
that updating is the way to go, just do it ;-)

Best regards

Ole


Reply to: