Rich Freeman <rich0 <at> gentoo.org> writes:
> > liveUSB, where folks can download "Gentoo Fever" onto a usb stick and stick > > into their current hardware and boot up a killer code development system. > > Building a liveUSB version of Gentoo is almost completely orthagonal > to building an automatic ebuild testing system. Agreed from where you sit. Where I sit, when I explain to folks about this project, a liveUSB stick running the latest in what gentoo-fever is, speaks volumes to encourage non-gentoo folks to take it for a test drive. > > A Gentoo CI system doesn't even have to be hosted on Gentoo, or on > Linux for that matter. Of course, if one were ever to become official > it most likely would be hosted on Gentoo, but most likely not on a box > booted from a USB. Um, I never saw "CI" defined, so please define specifically, then use the abbrev? > I'm not saying that a liveUSB version of Gentoo wouldn't be nice to > have. It just has nothing to do with solving this particular problem. OK, see above; you are right technically. Do you want a few dozen participates or a few thousand? > Sure, you need an OS to host a CI solution just like you need hardware > to host a CI solution on, and I wouldn't focus on the OS for the same > reason that I wouldn't focus on the hardware. LikeWahoa has solved this problem already. So all we need are a few additional packages to build 'gentoo-fever'. Those difficult codes you alluded to early need to get started. Sam has some ideas we can put on the first 'gentoo-fever' liveusb now. This is so, the messed up open source projects I participate in can quickly evaluate gentoo. Lots of folks know how to code. They mostly never have thick enough skin to stick with gentoo long enough to become productive coders in a gentoo-centric environment. Gentoo-fever can change that. Sure folks can work on other distros for this project. Those sorts of folks are rare, imho. Many folks need a clean, coherent environment to work on problems. MS, OSX and a myriad of other linux distros are clumsy and often opaque in the areas of development, in my decades of coding experience. Please don't underestimate that problem, since you sit in a very strong chair (basis) of ability. > You might very well > want to run the CI solution on a cluster, but inventing a new > clustering solution is also orthagonal to building a CI solution. Correct! But a very exciting twist for those who are interested in clustering, methinks. I for one, can hardly wait! > By all means work on those projects if you want to, but I wouldn't do > it in the context of building a CI solution. I can follow leadership on this issue. I cannot lead on this issue. Leadership is fleeting, as we discuss earlier. Don't stress over it. Somebody will show and 'Carpe Diem'. You should not have to do everything at Gentoo, so relax a bit....... As usual, your acumen is keenly correct. It's a party, so let folks 'party with ideas' a bit longer and let's see if somebody steps up with a liveUSB for (CI what ever that is) automated tools testing of ebuilds and then auto testing of the codes themselves. OH, Alex is well underway with etest..(grin). I'm learning how to package up a liveUSB for those Vikings buddies who are coding on subterranean physics..... but I am slow. If you think we have issues, dude, they looking to me for clarity on how to run a heterogenous coding project (ah ha ha ha ha ha). It would be funny, twer it not true. They put out a stable release 18 months ago and it's a fight because the module developers rarely test their modules against one another. None use gentoo, so build me up a gentoo-fever liveUSB, so I can put some ebuilds on it and share it with them? Add some packages that a typical gentoo dev would use? Share a few secrets? hth, James
