kicad-developers team mailing list archive
Mailing list archive
Re: ideas to improve patch and CI integration
On 10/22/2018 9:00 AM, Mark Roszko wrote:
>> I would bet money this will happen at some point in the future so it
>> could be an issue when it does unless we just limit the kicad project to
>> using the community edition.
> No need for anything other than community edition. I use it just fine at
> my company for 20 or so devs ;)
If the community edition meets the projects needs, all the better.
> Also in terms of hosting it, etc. We have access to something now from
> one of our great benefactors that'll let us run it there :#
If it's the same great benefactor that I'm thinking of, then they would
most likely be doing the gitlab code hosting as well.
> I've been busy doing some of the website stuff for that. The website is
> ready. I'm still working on a fancy new docs website to go into it to
> replace the current setup.
> The plan from Nick was to first migrate existing stuff over before we go
> entirely crazy with new stuff. He was going to move Jenkins to it too.
It makes sense to do this in manageable pieces so as not to overload our
> On Mon, Oct 22, 2018 at 8:56 AM Wayne Stambaugh <stambaughw@xxxxxxxxx
> <mailto:stambaughw@xxxxxxxxx>> wrote:
> On 10/22/2018 5:59 AM, Maciej Sumiński wrote:
> > Hi Thomas,
> > I like Gitlab, but it is a matter of taste and obviously the final
> > belongs to Wayne. I think the main problem would be to transfer
> the bug
> > tracker and decide what to do regarding the mailing list.
> We have actually discussed this possibility privately so it's on our
> radar. I personally would like to see the entire project on a single
> site instead of our current situation where we have repos on both
> launchpad and github. I haven't use gitlab so I cannot speak about it's
> pros and cons but I have looked over some the documentation and it seems
> comparable to github so I suspect it would be more than acceptable.
> > Another point is a plan for a case when Gitlab turns evil. Currently
> > they offer Gold version to open source projects for free , but you
> > never know what will happen in a few years. Either we pay or host a
> > limited version (Community Edition) privately. I realize the same
> > may happen with Launchpad, the point is to decide whether we would be
> > satisfied with Gitlab Community Edition features.
> I would bet money this will happen at some point in the future so it
> could be an issue when it does unless we just limit the kicad project to
> using the community edition.
> > You may also find insightful a discussion we head about moving KiCad
> > from Bazaar to Git .
> Fun times! As always, there are no silver bullets. If we choose to go
> down this path, there will be issues that will have to be resolved.
> > Cheers,
> > Orson
> > 1. https://about.gitlab.com/pricing/#gitlab-com
> > 2. https://lists.launchpad.net/kicad-developers/msg25668.html
> > On 10/21/18 7:11 PM, Thomas Pointhuber wrote:
> >> Hi,
> >> I think this topic will have many different point of views, but I
> >> it is important to discuss it anyway.
> >> I'm now about 5 years on the project, started as library
> maintainer and
> >> now sometimes with code contributions as well. I experienced at
> least a
> >> few different types of how software projects can be design of a
> >> management perspective to get work done.
> >> Currently, there is the time where many big organizations (GNOME and
> >> freedesktop.org <http://freedesktop.org> for example) moved their
> development to custom GitLab
> >> instances. And I would like to propose to think about that for
> KiCad as
> >> well. Not because everyone does it, but because it would benefit the
> >> project itself. There are various reasons I can think of right now:
> >> * patches are sometimes simply forgotten on the mailing list
> >> * it is pretty hard to find old patches on the mailing list
> unless you
> >> have the e-mails stored on your account. You cannot search on
> >> * CI integration of patches is missing. People need to run them
> >> before merging, or the automated build will fail at some point
> >> * getting CI runs for various systems requires explicit request
> due to
> >> missing automatization
> >> * no automated linter checks (would require us to reformat the
> >> * it's hard to track changes of patches, especially when modified,
> >> squashed,...
> >> * way better possibilities to view/edit/show history online
> >> * review tools on Launchpad are cumbersome and rarely used. This
> >> be the standard in point of fact.
> >> * easier to handle a proper review than with e-mail
> >> * easier to keep track how discussions evolved
> >> * the project is currently split between Launchpad and GitHub, this
> >> would allow us to combine all parts on one platform
> >> * GitLab would theoretically allow us add custom differs for
> library review
> >> * new contributors likely know GitHub/Gitlab workflow and can
> join more
> >> easily
> >> I know this is quite some heavy proposal due to change of
> workflow and
> >> bug-migration. Nevertheless, we should at least discuss it.
> >> PS: A college built a working GitLab CI script with export of build
> >> artifacts. GitLab integration of Jenkins would be possible as well.
> >> Regards,
> >> Thomas
> > _______________________________________________
> > Mailing list: https://launchpad.net/~kicad-developers
> > Post to : kicad-developers@xxxxxxxxxxxxxxxxxxx
> > Unsubscribe : https://launchpad.net/~kicad-developers
> > More help : https://help.launchpad.net/ListHelp
> Mailing list: https://launchpad.net/~kicad-developers
> Post to : kicad-developers@xxxxxxxxxxxxxxxxxxx
> Unsubscribe : https://launchpad.net/~kicad-developers
> More help : https://help.launchpad.net/ListHelp