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

Re: m68k port clears its Needs-Build queue

John Paul Adrian Glaubitz dixit:

>> Maybe one of the porters could explain the actual significance of this

Hm, it admittedly is still pretty impressive ;)

I’m using this as another chance to dist-upgrade my buildd,
both the chroot and outside, as it’s currently idle otherwise.

>> (I think it's not exactly the first time?);  and the overall status of
>> the port, to debian-publicity@ and/or a blog post via Planet Debian?  Or

Well, the port works, though you have to also use the unreleased
repository, and there’s no installer, though there’s currently
few enough core stuff in unreleased, so you can use debootstrap.
We also have base images, see wiki.d.o/M68k/Installing.

>> even just to me since I was curious enough to ask.
>Good idea, but I think we should rather wait until we have gotten above
>9000 or even 10.000 built packages.

We’re not getting there. The unfreeze is probably close. Why not now?

>There are still many packages that need to be fixed.

Right, but we can start from that… maybe the message gets fresh flesh.

>A good starter is too look at the packages in Failed state as these are
>packages that fail to build from source reproducibly.

Actually, those for which the buildd admin cared to reply with “fail”;
for some, the package maintainer already knows, plus most are small
fish. I posted a list of what I could say offhand in the other eMail.
I can dig down more later, but these are the best “enablers”:

boost, so we have the same boost-defaults version as other arches,
which gets rid of many these fails automatically.

OpenJDK, so we can finally be no GCJ arch any more… it’s got to be
still slow, but more compatible. Doko said most stuff should work,
but it fails somewhere late in the bootstrap-rebuild.

Erlang and GHC, as language enablers, too. No idea about Erlang.
We have GHC6 and GHC 7.0, but 7.2 FTBFS. The time in which GHC
7.2 was developed was the time the m68k port was dormant. Upstream
would actually like to help with that. Probably a codegen bug.

<igli> exceptions: a truly awful implementation of quite a nice idea.
<igli> just about the worst way you could do something like that, afaic.
<igli> it's like anti-design.  <mirabilos> that too… may I quote you on that?
<igli> sure, tho i doubt anyone will listen ;)

Reply to: