On Sunday 26 April 2020 05:18:53 andy pugh wrote: > On Sun, 26 Apr 2020 at 08:55, Gene Heskett <[email protected]> wrote: > > EXCEPT! Its not in either script so it has to be in > > dpkg-buildpackage. > > Buildpackage does not care about the tests, look at the contents of > http://buildbot.linuxcnc.org/dists/buster/2.8-rtpreempt/Contents-armhf >.gz > > It seems more likely that, for some reason, you have a commit in your > git tree with the file removed.
If I do, it came with the git clone that downloaded it. git checkout has been run 2 times, once to set it to 2.8, and once to set it back to master. A git pull has been done 10+ times. On each machine. > You said that "make clean" removes the file? It shouldn't, this is the > code: >From my backtrace history on the wheezy box. # To run the software type # # linuxcnc # # # ###################################################################### cd src && /usr/bin/make clean -s rm -f Makefile.inc rm -f src/config.log src/config.status dh_clean rm: cannot remove `./tests/tool-info/random-with-startup-tool/simpockets.tbl.orig': Operation not permitted debian/rules build After I'd done a sudo chattr +i to it. > https://github.com/LinuxCNC/linuxcnc/blob/master/src/Makefile#L496 > > I checked and find . -name '*.o' does not match .orig (of it did there > would be a lot more files disappearing) > > The buildbot is building Pi4 .debs several times a day (I have > forgotten why you are also doing this?) without problems. Since when? Seb shut that down around 20 months ago because that odriod was always crashed, matching my experience with it, I have a c2 thats bricked from trying to get around the UEFI BS in its BIOS, just to install linux. Its never showed up in an apt update since, so I've been building my own, first on the pi3 for several months, and on that pi4 since last summer after I installed it, well before I had the heart attack in late Sept. I have, on this list, offered to buy Seb a pi4 since it seems to be rock solid here, zero crashes since I put a $35 ups on it, offered at least twice but have never had a reply, yea or nay. And from the man page for dpkg-buildpackage, the options given in the wiki doc on building Linuxcnc, do cause a make clean as shown above. That can be shut off, but should it? And as shown by the backtrace clip above, its the make clean stage, issued by dpkg-buildpackage thats complaining it can't delete that file. And in further checks last night, the wheezy build continues, but to the rpi4 its a fatal error and bailout. So if a fresh git clone doesn't fix it, I'll add that git checkout line that puts it back just before the runtests invocation in my build scripts. My build scripts are an exact copy of the wiki, boiled down into a pair of scripts, with a git pull as the first active line added, one script that does everything to runtests, and the second builds the installable debs that I run if runtests passes all. Neither of these git clones is precious, so I can blow then away and do new clones if you think that will fix it. But I'm getting less and less respect for git as this plays out. Thanks Andy. 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 [email protected] https://lists.sourceforge.net/lists/listinfo/emc-developers
