On 2/3/19 11:55 AM, Fabian Groffen wrote:
> On 27-11-2018 10:20:52 +0100, Michael Haubenwallner wrote:
>> On 11/27/2018 09:37 AM, Sam Pfeiffer wrote:
>>> On Tue, Nov 27, 2018 at 7:20 PM Fabian Groffen <[email protected] 
>>> <mailto:[email protected]>> wrote:
>>>
>>>> I don't want to depress this entire discussion, but it would be really
>>>> nice if we could somehow interact with special machines people have at
>>>> their company or at home.  Prefix needs testing on many different
>>>> machines (non-Linux) which usually don't exist in docker images.
>>
>> I second this - and let me add a further aspect here:
>> What I know from buildbot setup is that the master does provide (mostly 
>> shell)
>> commands to be executed on the slave. This is fine as long as there is 
>> limited
>> visibility for the master. But when a public buildbot master is being 
>> hijacked,
>> it feels too easy to execute malicious commands even on the slave machines.
>>
>> So over a buildbot like setup, I would prefer a Jenkins like setup, where the
>> master does provide only trigger information to slaves. And even more 
>> appealing
>> would be a standalone slave setup, where the master does just receive the 
>> build
>> logs for the public, without access to slave machines at all.
> 
> So, with this in mind, I've started experimenting, here's my "progress":
> 
> http://bootstrap.prefix.bitzolder.nl/results/

Nice!

> The idea is to rsync the result after the bootstrap-prefix.sh call to the
> server.  I can have setup to be in an "upload" sense.  The current call
> (which assumes direct access) can be found in the dobootstrap script I
> currently use to fire off a bootstrap on a platform:
> 
> http://bootstrap.prefix.bitzolder.nl/dobootstrap  (see DOPUBLISH)

So I'm wondering how to enable myself to provide logs for some more CHOSTs.
What about rsync + ssh via pecker?

> None of these targets are RAP by the way.  I think the current CI is
> very good at that.

Absolutely. However, it would be nice if we could integrate the Linux/RAP
results into this overwiew as well - besides the Linux/Guest ones, even
if they share the same CHOST...

Also, just've found https://wiki.gentoo.org/wiki/Prefix/tested where the
'Last tried' column values seem outdated - maybe CI builds can provide
more recent dates there as well.

> By the way, no bootstraps succeeded recently, so that's the goal to get
> that triggered so we can focus on fixing it.  Just being able to pull in
> the CI success/fail for that would already be a start.

FWIW, I've created a gentoo-prefix project with Azure pipelines, but their
6 hours limit is too small for Prefix on Cygwin. So I've added my own
Windows VM there: https://dev.azure.com/gentoo-prefix/ci-builds/_build
However, I'm not sure if I should keep that for security concerns...

BTW, Cygwin 3.0.0-0.8 does have the fork() that works for Gentoo Prefix!

Thanks!
/haubi/

Reply via email to