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

Re: [Backport - arch-dependent package] ruby-nokogiri



On Mon, Oct 28, 2019 at 08:04:51PM +0530, Pirate Praveen wrote:
> On 2019, ഒക്‌ടോബർ 28 4:12:44 PM IST, Utkarsh Gupta <guptautkarsh2102@gmail.com> wrote:
> >Hey,
> >
> >On 28/10/19 1:20 pm, Nilesh wrote:
> >> Hi,
> >> I had been trying to backport ruby-nokogiri, as it is one of the many
> >> dependencies of gitlab-12.2.8, and hence needs backporting in order
> >to
> >> backport gitlab.
> >> However, it fails with the error:
> >>
> >> <snip>
> >> dwz:
> >>
> >debian/ruby-nokogiri/usr/lib/x86_64-linux-gnu/ruby/vendor_ruby/2.5.0/nokogiri/nokogiri.so:
> >> Found compressed .debug_aranges section, not attempting dwz
> >compression
> >> dh_dwz: dwz -q --
> >>
> >debian/ruby-nokogiri/usr/lib/x86_64-linux-gnu/ruby/vendor_ruby/2.5.0/nokogiri/nokogiri.so
> >> returned exit code 1
> >> make: *** [debian/rules:9: binary] Error 1
> >> dpkg-buildpackage: error: fakeroot debian/rules binary subprocess
> >> returned exit status 2
> >>
> >> (pushed my changes to:
> >https://salsa.debian.org/gi-boi-guest/ruby-nokogiri/)
> >>
> >> It is due to debhelper version 12, and changing it to 11 its solves
> >the
> >> issue. So, should the compat level be changed to 11, and be
> >uploaded(but
> >> that will be too much work for backporting all arch dependent 
> >packages)
> >> or should ruby be backported in order to resolve the issue? It would
> >be
> >> great if it could be clarified.
> >
> >From what I know and what we discussed on IRC, I think there are two
> >ways to go about this.
> >First, either downgrade the version of dh-compat to 11 while
> >backporting
> >each package, or
> >Second, backport ruby2.5 2.5.5-4 (because it contains the fix).||||||
> >
> >However, given that there are many packages that'd be having the same
> >problems while backporting,
> >I'd be a +1 for backporting 2.5.5-4 (or 2.5.7-1) to buster-backports.
> >
> >What would other's opinion on this be?
> 
> I think its fine to backport as there is no soname change and we don't need to rebuild any native extensions (so no impact on existing packages). I'm wondering if 2.5.5-4 is suitable for a stable update as it'd mean we don't need to bump minimum ruby version for packages we need back porting.

2.5.5-4 *could* be made into a stable updade (and not a backport).
However, I'm not sure what the consequences of that would be. I would
like someone to run such interpreter on a stable system for a while and
report the results.

Meanwhile, just downgrading debhelper compat to 11 seems easy enough
when backporting. And note that's only necessary at all for C
extensions, so th eimpact on actual backporting efforts should not be
that large.

Attachment: signature.asc
Description: PGP signature


Reply to: