I know many user need rtai 32bit performances .... but more interesting machinery today was build on 64bit SO .... rt-preempt work ok with hight speed delta with servo motors .... servo motors actually is not so expensive than sme years ago ... so rt-preempt is not a bad choice for all user .. there is also to add that since it is no longer possible to use ubuntu 10.04 rtai Lcnc can no longer achieve the performance of the time .... so the problem does not arise for me. rt-preempt is enough .... if you have cheap and high performance stepper motors just add a mesa card that helps to solve the gap.
*** obviously I do not claim to be an exhaustive or complete argument .... I think it is interesting only as a statistical document **** Il giorno gio 18 giu 2020 alle ore 07:45 John Morris <j...@zultron.com> ha scritto: > > On 6/12/20 5:03 PM, andy pugh wrote: > > On Fri, 12 Jun 2020 at 04:28, Reinhard <reinha...@schwarzrot-design.de> > wrote: > > > >> So let's ask for the other side of view - what's the difference between > a > >> snapshot from master to a ordinally rolled out release - from the user > side of > >> view? > >> Afaik all they want, are packages to install from without compiling. > > > > A major release has a matching ISO file for a scratch install, is > > thought to be largely bug-free and has documentation that matches the > > actual behaviour. > > At least in principle. > > > > Machinekit abandoned the idea of releases. I have heard that that > > makes it hard to find a version that actually works. > > Correct, the Machinekit project doesn't make releases, per C4 [1]. The > idea is that anyone who wants something like a "stable release" is free > to fork the project and do so themselves. In practice, this has been > done to support commercial products, not directly by the community. The > only *public* stable releases I'm aware of are Robert C. Nelson's > Beaglebone images and perhaps one by The Cool Tool; otherwise, we > maintain private stable releases in projects I'm involved with, and I've > heard of others doing the same. > > However, in the past, I haven't heard of the master branch remaining > broken for very long at all, I imagine similar to the LCNC project's > master branch. That's not exactly true at the moment, however, when the > code is being restructured as part of a scramble to replace the CI and > package distribution infrastructure with systems that can be trivially > reproduced, after the old system running on private hardware suddenly > went down. Lesson learned. > > Andy's comment points out an important difference between the two > projects' approaches: Machinekit offloads the burden of making releases > (as well as simplifies code reviews, again per C4) in order to speed up > the pace of accepting contributions (both new features and bug fixes), > whereas LinuxCNC prioritizes providing a rock-solid, turn-key binary > distribution with long-term support for use by its large user community. > Delaying the RTAI ISO in order to get 2.8 out sooner hardly seems like a > significant step towards Machinekit's "no-releases" model, even while it > is a move to temporarily streamline one release requirement. > > [1]: https://rfc.zeromq.org/spec/42/ > > John > > > _______________________________________________ > Emc-developers mailing list > Emc-developers@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/emc-developers > _______________________________________________ Emc-developers mailing list Emc-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-developers
