I checked with Cristina A and the 60% (last week) passing was the number of components whose srpms had successfully been used to build rpms that passed the tests. These are the rpms that will be released. The directly built rpms are being used in tests for components prior to this.
John
-----Original Message-----
From: Testbed Support for GridPP member institutes [mailto:[log in to unmask]] On Behalf Of Ewan MacMahon
Sent: 13 April 2011 10:55
To: [log in to unmask]
Subject: Re: Middleware - release and updates summary
> -----Original Message-----
> From: Testbed Support for GridPP member institutes [mailto:TB-
> [log in to unmask]] On Behalf Of John Gordon
>
> From time immemorial ETICS has been building binary rpms from the source.
> I never really understood the technical reasons why they did not produce
> srpms but they didn't and the UK were foremost among those asking for
> them. Now that ETICS has started to produce them it seems a bit mean to
> say that this is worrying. Perhaps we should wait and see what is
> delivered.
>
It's not supposed to be mean, it's supposed to be constructive; if
etics is now generating srpms then it should be possible to take them
and run them through a conventional build process to generate the
binaries (after, all, if it's not possible to generate binaries form
them, then there isn't much point having them).
Based on past experience I worry slightly that something will be
delivered that ticks a box (srpms generated) while actually doing
so in a manner that makes it not terribly useful in practice.
I also don't agree that we should want and see; if we ask the question
"Are the released binaries generated from the released srpms?" and the
answer is 'yes' then everyone can be happy and we're all done with this.
If the answer is 'no' then it's easier to fix that at this stage of
proceedings than to wait until it's become established.
Ewan
|