On Wed, Feb 01, 2017 at 06:40:45PM -0800, Julian Rendell wrote:
If the other interested parties are ok with using gitlab then let’s get
it
setup.
Let’s give them a couple of days to respond.
Ok, sounds good.
Does the gitlab outages affect this suggestion at all?
Our current gitlab endeavors are primarily aimed at information-gathering to support a decision in around a month about migrating the core Inkscape codebase there. So in that context the recent gitlab outage is certainly an important data point to take into consideration.
However, in the context of nearer term efforts such as inkscape_web and this mac os porting work, so long as the outage has been resolved (as it appears it's been), I wouldn't let that color the decision.
Is the project planning on using Gitlabs CI system? Skimming < https://about.gitlab.com/gitlab-ci/%3E it looks like you have to self-host the runners (build agents.) Ok, found this thread ( https://github.com/travis-ci/travis-ci/issues/5931) that gives some details. There are free, shared runners, but only for linux.
I like CI, it keeps me honest. But worst case, primary hosting on gitlabs, secondary on github, use Travis with the github mirror for Mac builds. Git makes multiple remotes pretty straight forward :-)
Yes, gaining good CI capabilities is definitely going to be part of our decision process. Experimentation towards achieving multi-platform CI in the gitlab ecosystem would thus help a lot towards making the right decision - even failed experiments here will provide good data.
The Travis OS X support page looks pretty good ( https://docs.travis-ci.com/user/osx-ci-environment/) - looks like they have the current OS release, and the previous two available, and brew available to install tools and dependencies.
Hosting on Gitlabs and using Github as a mirror/CI system could be a nice backup strategy.
Yep, I could easily see us ending up with a bit of a hybrid of the two.
It depends of course on how you want to invest your energies and how urgent your needs are. That said, from the projects' perspective having Gtk3 OS X packaging ready for 0.93 is the more strategically important since it will be the longer term need. And to my knowledge no one is working on it at present, so it's a clear area where help is needed.
I’d much rather work on something that helps this long term. Right now, the brew tap that was created solves my personal and immediate needs. I’ll look into means of supporting other local Mac users when that crops up. By then I might have enough knowledge to at least build everything into a single directory and provide a bash script + icon for them to use in the interim.
Cool, I'll try to keep tabs on how things go.
Bryce