[Qgis-psc] Discussion our financial situation

[Andreas Neumann]

Hi Matthias, 

Outsourcing infrastructure is preferred - I think. But if the outsourced
services are so limited that they can't handle a large project like
QGIS, or take waaayyyyy ttoooooo loooooongggggg to build, then
self-hosted might be superior? A six or 8-core machine at Hetzners is
certainly affordable. 

Certainly something to investigate. 

Andreas 

On 2019-11-14 09:40, Matthias Kuhn wrote:

> Hi Andreas, 
> 
> It would certainly be an option. And there are others as well (like good old Jenkins for example). 
> 
> Are you asking for self-hosted because of performance or for other reasons? I think in the past there was a shift towards outsourcing such things to external services that do this as their core business instead of having to manage infrastructure on our own (e.g. github, travis). 
> 
> As mentioned before, I think someone will need to invest some real time to exploring possibilities and limitations to be able to come forward with a good plan of action. I think it would be good to have a week of time to test different options. 
> 
> Matthias
> 
> On 11/14/19 9:28 AM, Andreas Neumann wrote: 
> 
> Hi Matthias, 
> 
> Would it be an option to "self-host" appveyor at our own (to be ordered) Hetzner Windows machine? 
> 
> I see that Appveyor offers a self-hosted version, free for one team and with only community support: https://www.appveyor.com/self-hosted/ 
> 
> Andreas 
> 
> On 2019-11-14 09:22, Matthias Kuhn wrote: 
> 
> Hi Andreas, 
> 
> Yes, I remember this. I also invested some time into AppVeyor back then and the timeouts. 
> 
> Finding a reliable solution and exploring possibilities will be an investment in itself already. I think that with a good financial situation, it's also a good possibility for the project to invest into such an area with a potentially great benefit. But it's ultimately the PSC's decision if the project prefers to sponsor this work or hope for it to happen and outsource the work and risks. 
> 
> Matthias
> 
> On 11/14/19 8:59 AM, Andreas Neumann wrote: 
> 
> Hi, 
> 
> The Windows CI thing would be worth-wile to invest it. But I think we already had it in the past and because of a lot of problems (e.g. timeouts when building, etc.) we stopped using it. I know that we definitely paid for such a service in the past. It was called "Appveyor" back then. Nyall might know more about it. I think he paid for it upfront and then I reimbursed him. That was in 2016. 
> 
> I think we would first have to find a reliable solution for Windows CI before investing into it. 
> 
> Andreas 
> 
> On 2019-11-14 08:51, Anita Graser wrote: 
> 
> Hi Andreas, 
> 
> On Wed, Nov 13, 2019 at 9:57 PM Andreas Neumann <a.neumann at carto.net> wrote: Any ideas? Do we have useful and important work that could be finished 
> until January for around 15k? Should we invest more QGIS bug fixing? Documentation (if we find one)? 
> Our infrastructure? Or upstream qt5 improvements? 
> 
> On the infrastructure side, there seems to be interest in Windows based CI. See current thread on [Qgis-psc] Direct push forbidden to master. However, I'm not 100% sure how this relates to a previous discussion from 2018 in the thread [QGIS-Developer] Windows compilation infrastructure. 
> 
> Regards, 
> Anita

_______________________________________________
Qgis-psc mailing list
Qgis-psc at lists.osgeo.org
https://lists.osgeo.org/mailman/listinfo/qgis-psc

_______________________________________________
Qgis-psc mailing list
Qgis-psc at lists.osgeo.org
https://lists.osgeo.org/mailman/listinfo/qgis-psc
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/qgis-psc/attachments/20191114/43d62490/attachment.html>