On 22/09/2020 14:51, Jürgen Groß wrote: > On 22.09.20 14:17, Olaf Hering wrote: >> With commit 68a8aa5d72 the build randomly fails. Up to commit >> c7e3021a71 no errors (like this) are seen. For some reason nothing >> below tools/ was touched in this range, so the bug may already exist >> for a while. >> >> attempt on host#1: >> [ 81s] ln: failed to create symbolic link >> '/home/abuild/rpmbuild/BUILD/xen-4.15.20200922T102230.68a8aa5d72/non-dbg/tools/libs/ctrl/../../../tools/include/xenctrl_dom.h': >> File exists > > This is very weird, as the link is created via "ln -sf", so an existing > target should not make the command failing. > >> [ 81s] >> /home/abuild/rpmbuild/BUILD/xen-4.15.20200922T102230.68a8aa5d72/non-dbg/tools/libs/ctrl/../../../tools/libs/libs.mk:82: >> recipe for target >> '/home/abuild/rpmbuild/BUILD/xen-4.15.20200922T102230.68a8aa5d72/non-dbg/tools/libs/ctrl/../../../tools/include/xenctrl_compat.h' >> failed >> [ 81s] make[6]: *** >> [/home/abuild/rpmbuild/BUILD/xen-4.15.20200922T102230.68a8aa5d72/non-dbg/tools/libs/ctrl/../../../tools/include/xenctrl_compat.h] >> Error 1 >> >> attempt on host#2: >> [ 178s] xenstat_qmp.c:26:10: fatal error: _paths.h: No such file or >> directory >> [ 178s] >> me/abuild/rpmbuild/BUILD/xen-4.15.20200922T102230.68a8aa5d72/non-dbg/tools/libs/stat/../../../tools/Rules.mk:153: >> xenstat_qmp.opic] Error 1 > > This shouldn't happen, too. > > Is this really a normal upstream build, or do you have any additional > patches applied?
I've also observed the first symptom sporadically. They've all been in my main dev tree, which absolutely includes the possibility of an unclean tree. So far I haven't observed it from a clean tarball in a mock chroot. ~Andrew
