On 2/26/19 4:58 PM, Sérgio Basto wrote:
On Tue, 2019-02-26 at 16:09 -0800, Kevin Fenzi wrote:
> On 2/26/19 11:11 AM, Sérgio Basto wrote:
>> On Tue, 2019-02-26 at 10:38 -0500, Ben Cotton wrote:
>>> According to the Fedora 30 schedule[1], the 100% code complete
>>> deadline[2] for Changes is Tuesday, 5 March. The beta freeze[3]
>>> takes effect on this date as well. All Changes should be in
>>> "ON_QA"
>>> state by then.
>>
>> I still haven't any compose for F30 completed, we should have time
>> to
>
> We have. The compose on the 24th finished and synced out.
>
>> build some packages against f30 buildroots , this situation also
>> happened in F29 .
>
> The buildroot is in koji and is always updated, or do you mean
> something
> else?
yes, without composes, buildroots on copr, mock and RPMFusion aren't
updated .
Yeah, sadly true. You could use the fedora koji buildroot repo, but I
don't think we want to have too many people hit it.
Without buildroots updated is much more difficult test the builds.
Sorry for my bad mood , I'm tired, but since I'm on reflection , I
think after mass rebuild with new gcc we should wait a little more
until mass branching, new gcc always have a lot of FTBFS
Well, I don't think any of the compose failures were actually due to
gcc. It's mostly I think everyone dropping in there work so it lands in
f30, and it's up to us to integrate it and get composes working.
No branched composes also means no testing installations, dnf broken
deps and debug a lot of things .
I don't disagree, but we are trying as best we can to get composes
working. I don't think a delay would have helped in particular. People
finishing and landing things earlier so we had less of a chain of issues
would have helped, but I have no idea how to get that to happen. ;(
kevin