Hello, Björn Höfling <bjoern.hoefl...@bjoernhoefling.de> ezt írta (időpont: 2019. febr. 5., K, 21:40): > > On Tue, 5 Feb 2019 20:49:35 +0100 > Andreas Enge <andr...@enge.fr> wrote: > > > I cannot reproduce this; could it be some non-determinism in the build > > process? Both packages build without problem for me with commit > > 406ef442580e6a1b68ca99091fef2935dc4fbb4f . > > I can confirm it is indeterministically failing: I tried it a week ago > or so and sometimes it builds and sometimes not. > > For example, the latest Hydra evaluations fails: > > https://hydra.gnu.org/build/3346840 > > I don't understand the error-log or which of these messages is a real > error and which not. Anyone else? >
The showstopper here seems to be this: ../RULES/local.cnf:46: OBJ/x86_64-linux-gcc/local.cnf: No such file or directory 4273make[1]: *** No rule to make target 'autoconf.m4', needed by 'configure'. Stop. 4274make[1]: *** Waiting for unfinished jobs.... this part looks like this on a working build: ln: failed to create symbolic link './priv.c': File exists ln: failed to create symbolic link './acgeneral.m4': File exists ln: failed to create symbolic link './aclocal.m4': File exists ln: failed to create symbolic link './acoldnames.m4': File exists ln: failed to create symbolic link './acspecific.m4': File exists > Björn I will have a further look at this tomorrow, comparing earlier parts of the log. The ext2 header also warnings on succesful builds, so that is not the problem. I also believe that perl is not the problem. Best regards, g_bor