Sgallagh suggested opening up the current functionality of coprs to proven packagers - to discover more bugs and to help us figure out what else is needed.
I'd like to suggest two things:
1. open up what we have to provenpackagers as sgallagh suggested
2. setting up a second 'dev' instance so we can test there before rolling over into the more public instance. - Also it will help us verify the setup process.
I can easily setup a be and fe dev instances in the private openstack cloud we have setup. It would also let us refresh those instances and be sure we're still functional.
-sv
----- Original Message -----
Sgallagh suggested opening up the current functionality of coprs to proven packagers - to discover more bugs and to help us figure out what else is needed.
I'd like to suggest two things:
open up what we have to provenpackagers as sgallagh suggested
setting up a second 'dev' instance so we can test there before
rolling over into the more public instance. - Also it will help us verify the setup process.
I can easily setup a be and fe dev instances in the private openstack cloud we have setup. It would also let us refresh those instances and be sure we're still functional.
-sv
Hey Seth and all, although I have no problem with this, I'd like to suggest that we rethink the design a bit (at least of frontend) and I'd like to get rid of writing new DB migrations and just say that the current stuff will not be transfered to future testing instance. Would that be ok with everyone? I can keep creating migrations and testing that nothing breaks, but it takes a lot of time and I think this time could be spent better on creating new functionality. As for the redesign, I'll post my ideas in a couple of days for comments, hopefully.
Does anyone disagree? Thanks, Slavek.
On Tue, 18 Jun 2013 01:42:18 -0400 (EDT) Bohuslav Kabrda bkabrda@redhat.com wrote:
----- Original Message -----
Sgallagh suggested opening up the current functionality of coprs to proven packagers - to discover more bugs and to help us figure out what else is needed.
I'd like to suggest two things:
open up what we have to provenpackagers as sgallagh suggested
setting up a second 'dev' instance so we can test there before
rolling over into the more public instance. - Also it will help us verify the setup process.
I can easily setup a be and fe dev instances in the private openstack cloud we have setup. It would also let us refresh those instances and be sure we're still functional.
-sv
Hey Seth and all, although I have no problem with this, I'd like to suggest that we rethink the design a bit (at least of frontend) and I'd like to get rid of writing new DB migrations and just say that the current stuff will not be transfered to future testing instance. Would that be ok with everyone? I can keep creating migrations and testing that nothing breaks, but it takes a lot of time and I think this time could be spent better on creating new functionality. As for the redesign, I'll post my ideas in a couple of days for comments, hopefully.
Does anyone disagree? Thanks, Slavek.
I have setup two new dev instances that have all the prerequsites but are more or less blank.
copr-be-dev.cloud.fedoraproject.org copr-fe-dev.cloud.fedoraproject.org
these boxes are available, using your fedora account ssh keys for you to login to - but as root - not as your username.
These are throw-away cloud instances. If you decide it is messed up and you want me to reset the whole thing, just say the word - it takes very little time.
These instances are not backed up, of course, but they are dev instances.
they should have the same ports and access to things as anything else.
yell at me if you need something or what not. -sv
On 06/18/2013 06:48 AM, seth vidal wrote:
- open up what we have to provenpackagers as sgallagh suggested
a) I agree with what Bohuslav said. b) until we provide build logs then COPR is not useful if build fail.
- setting up a second 'dev' instance so we can test there before
rolling over into the more public instance. - Also it will help us verify the setup process.
+1 I would be more than willing to do that as part of boot camp. But I would need your guidance for sure. For starter where I can get machine from Fedora Cloud?
Mirek
On Tue, 18 Jun 2013 08:53:27 +0200 Miroslav Suchy msuchy@redhat.com wrote:
On 06/18/2013 06:48 AM, seth vidal wrote:
- open up what we have to provenpackagers as sgallagh suggested
a) I agree with what Bohuslav said. b) until we provide build logs then COPR is not useful if build fail.
We provided build logs now.
What are you talking about?
-sv
On Tue, 18 Jun 2013 00:48:51 -0400 seth vidal skvidal@fedoraproject.org wrote:
Sgallagh suggested opening up the current functionality of coprs to proven packagers - to discover more bugs and to help us figure out what else is needed.
I'd like to suggest two things:
open up what we have to provenpackagers as sgallagh suggested
setting up a second 'dev' instance so we can test there before
rolling over into the more public instance. - Also it will help us verify the setup process.
I can easily setup a be and fe dev instances in the private openstack cloud we have setup. It would also let us refresh those instances and be sure we're still functional.
I've been talking with patrick uiterwijk about the openid teams extension for python. Once that is in place we will be able to specify group memberships for users who can log in and use the copr instance(s) we have setup.
-sv
----- Original Message -----
On Tue, 18 Jun 2013 00:48:51 -0400 seth vidal skvidal@fedoraproject.org wrote:
Sgallagh suggested opening up the current functionality of coprs to proven packagers - to discover more bugs and to help us figure out what else is needed.
I'd like to suggest two things:
open up what we have to provenpackagers as sgallagh suggested
setting up a second 'dev' instance so we can test there before
rolling over into the more public instance. - Also it will help us verify the setup process.
I can easily setup a be and fe dev instances in the private openstack cloud we have setup. It would also let us refresh those instances and be sure we're still functional.
I've been talking with patrick uiterwijk about the openid teams extension for python. Once that is in place we will be able to specify group memberships for users who can log in and use the copr instance(s) we have setup.
-sv
How exactly will this work? The information will be available in the extra info that OpenID can ask for when authenticating?
On Tue, 2 Jul 2013 04:39:20 -0400 (EDT) Bohuslav Kabrda bkabrda@redhat.com wrote:
----- Original Message -----
On Tue, 18 Jun 2013 00:48:51 -0400 seth vidal skvidal@fedoraproject.org wrote:
Sgallagh suggested opening up the current functionality of coprs to proven packagers - to discover more bugs and to help us figure out what else is needed.
I'd like to suggest two things:
open up what we have to provenpackagers as sgallagh suggested
setting up a second 'dev' instance so we can test there before
rolling over into the more public instance. - Also it will help us verify the setup process.
I can easily setup a be and fe dev instances in the private openstack cloud we have setup. It would also let us refresh those instances and be sure we're still functional.
I've been talking with patrick uiterwijk about the openid teams extension for python. Once that is in place we will be able to specify group memberships for users who can log in and use the copr instance(s) we have setup.
-sv
How exactly will this work? The information will be available in the extra info that OpenID can ask for when authenticating?
That's pretty much it, yes
-sv
copr-devel@lists.fedorahosted.org
-
Bohuslav Kabrda -
Miroslav Suchý -
Seth Vidal