On Thursday 11 June 2020 23:28:26 Reinhard wrote: > Hi Gene > > On Donnerstag, 11. Juni 2020, 21:53:55 CEST Gene Heskett wrote: > > I find it totally unreal that most shop owners won't spend the > > sheckles to buy a router and maybe a switch to put the network into > > every machine in the building. Security is in how you do it, and > > I've not been touched in 20 some years. > > Yes, I know it the same way. Many shop owners have physically > separated networks. That's not the question and no problem. > > > Yet I can update any machine on my locale network everytime the > > buildbot makes a new version available to apt or synaptic. > > I know you run a build farm. And I know the cost of a release from > developer point of view. > > 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. > So if the buildbot generates that packages, you could simply declare a > snapshot to be a release. That would be a manageable job. > > But as I understand Andy, he does a lot of extra testing to deliver a > pretty bugfree release. That's what I'm talking about and I think, > that time could be better invested. > > > And I wouldn't have a milliseconds problem with an annual small fee > > to pay for buildbot electricity and server bandwidth. > > I know - but users that think the same way as you don't cry for new > versions or for many releases in short cycles. > You help a lot to push the project ahead - so it wasn't you, that I > was talking about. > I think of myself more as the canary in the coal mine, lightly exercising the buildbots output, occasionally even building my own armhf version via a git pull to check on things in the pi world. And if my canary dies, reporting the failure ASAP.
Sometimes, I won't catalog, my reports have been fixed in a few hours. So thats my contribution to the stability of the code seen by the installers using a "release". That is their choice. But its gaining new or better features they are missing out on too by sticking with a release. But real brokeness, even in a production shop, will be found to not have, IMNSHO, a very serious effect on the bottom line by running master, it quite simply is not that broken for that long. If I was better at scripts, a daily git pull and build of the install debs on that armhf would be running. With any failures generating an automatic email to raise and wave that famous hand. What I need, since I've broken the git pull and build to the make tests stage, and a separate deb build is a way to condition the deb build on the results of the make tests. Many thanks to Andy and probably others, LinuxCNC is likely the most tested software on the planet every time the buildbot makes a new master, with going on 250 "does it actually run right" tests of every function it has as part of the make tests at the end of my first script on the pi4b. I don't think new functions get added without a suitable test being written. On my pi, make tests takes more time to run than the actual build. I don't think the huge majority of "release" users know or appreciate that. The wiki maintainers are failing at their advertising efforts for not makeing that little detail a hell of a lot more promenant. There is not an RTAI for armhf, so all my builds are for a preempt-rt kernel. I get latency overruns even with RTAI, on my wheezy builds, short enough the machine doesn't seem to notice, but I don't get any more running uspace on the pi4b either. I've no clue about arm64 builds. But I am running, other than kernel related stuffs, 6 packages its not updating of a 100% uptodate buster build from raspbian on that pi4b. > cheers Reinhard > > > > > _______________________________________________ > Emc-developers mailing list > Emc-developers@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/emc-developers Cheers, Gene Heskett -- "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author) If we desire respect for the law, we must first make the law respectable. - Louis D. Brandeis Genes Web page <http://geneslinuxbox.net:6309/gene> _______________________________________________ Emc-developers mailing list Emc-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-developers