On May 19, 2011, at 1:25 PM, Torsten Bergmann wrote: > Stef wrote >> Now my main questions are > > Hey ... dont take me wrong.
I did not :) I just took your mail as a pretext to make us all think :) > I want the build server to > build the VMs too, I want a repeatable process and > a reliable VM. I appreciate that we change this!!! > > We all know that if VM development is dependent on a single > maintainer person then nobody would invest into Pharo, Squeak > or Cuis since the truck factor* is one. > > I support all this. Remember: I was the one who wrote the first > ConfigurationOfVMMaker so that it is easy to get all the bits > and pieces. I want to make it easier for people to dive into > VM building. I know :) > The windows build slave is a huge step forward ... we > can not only build the VM and installer ... we can also run > the tests on Windows too to see if there are platform specific > issues. Cool and hardly needed! > > But someone (person or team) has to wear the Win32 VM "hat" > to check contributions or help fixing a bug. So may question is what > is the direction we will take here? I have the impression that igor will have to have a look at it :) We did not tell him yet :) but there is a high chance that igor should do something :) > It looks like the standard SqueakWin32 VM is unmaintained since the > last release is from July 2010. We still have no Alien, we still > have no callbacks. Yes, there is Cog whose code base is newer > but I think Eliots focus is on speed rather than functionality. > > Will we fork from standard VM, will we participate in Eliots > work or will we fork an own using git and maintain it by picking pieces > from the other VMs? this is not the answer that we can give alone :) > Also: what can we do to reach the goal of feature complete VM's > with aligned version numbers. I think this means to participate on > lists like "vm-dev"... stef
