Re: Bug#789133: transition: ocaml 4.02.2
On 20/06/15 18:02, Stéphane Glondu wrote:
> Le 19/06/2015 12:56, Emilio Pozuelo Monfort a écrit :
>> > I see some of the failing packages have in the log:
>> >
>> > -> Finished parsing the build-deps
>> > Wrong version of OCaml!
>> >
>> > That does that mean the package couldn't be built because of the dependency
>> > problems you mention?
> Indeed.
>
>> > My only concern here is that with 41 failing packages, the transition may take
>> > quite a while to finish, blocking other stuff. That'd be different if most of
>> > those packages will just build fine after the binNMUs, but I have no idea if
>> > that's the case...
> No, it's not the case. However, having an old version of OCaml in
> unstable also blocks other stuff: new versions of OCaml-related stuff
> start picking up new features of OCaml so we cannot update them before
> OCaml. Moreover, sometimes, fixes for failing packages need the new
> version of OCaml. That's why I am in favour of removing packages from
> testing in order to update OCaml. IMHO, failing packages can be fixed later.
Sure, I'm fine with removing a few packages if necessary if those don't have
rdeps, and are not very important (e.g. they have low popcon). The usual stuff.
I'm just asking because I'd like to make sure the transition doesn't block for
too long because there are a bunch of FTBFS that we knew about before the
transition started. So I want to make sure the impact that those will have.
So, I'd like to know what the plan is for those packages that are "missing".
E.g. if those maintained by the ocaml team will be fixed promptly, and what will
happen to the others.
I'd like to see them analyzed and bugs filed (ideally with patches) before we
start this.
Or if you can give a more detailed explanation of what will happen after ocaml
is uploaded, binNMUs are scheduled, and we have ~30 packages that are holding
the transition.
Thanks for bearing with me with my first ocaml transition.
Cheers,
Emilio
>> > I do wonder how many of those are actual failures, of those, how many are
>> > maintained by the ocaml team and how many are not...
> I've recompiled everything with the final ocaml 4.02.2, fixing a few
> things on the way. The build logs are available at:
>
> http://ocaml.debian.net/debian/ocaml-4.02.2/
>
> There are 34 MISSING packages. I have attached a summary.
>
>> > BTW if you have filed bugs for the failing packages, please make them block this
>> > tracking bug.
> I will.
>
>
> Cheers,
>
> -- Stéphane
>
>
> missing.txt
>
>
> Not in testing:
> llvm-toolchain-3.6
> llvm-toolchain-snapshot
> ocamlduce
> janest-core
>
> Use compiler internals, should be removed from testing if needed:
> jocaml
> mingw-ocaml
> cmigrep
> otags
> cduce
> js-of-ocaml
> eliom (needs js-of-ocaml)
> nurpawiki (needs eliom)
>
> Maintained by the Debian OCaml Team:
> coq-doc (fix in coq)
> ocaml-fdkaac (dep issue, libfdk-aac-dev)
> coccinelle (dep issue, camlp4)
> lablgtk-extras (Some fatal warnings were triggered)
> ocaml-reins (Some fatal warnings were triggered)
> approx (Some fatal warnings were triggered)
> dose3 (issue in RPM bindings, #789354)
> ocaml-gettext (segfault, suspicious double linking of Unix)
> ocamldap
> matita (a class type should be virtual)
> ocsigenserver (dep issue)
> opam
> why
> liquidsoap
> coinst
> nss-passwords (int types, I am upstream)
>
> Maintained by others:
> monotone-viz
> plplot (configure error)
> libguestfs (needs ocaml-gettext)
> virt-top (needs ocaml-gettext)
> zeroinstall-injector (string/bytes discrepancy)
> botch (needs dose3)
>
Reply to: