bug#66692: [cuirass] Spurious failed builds following patching u-boot sources

[Maxim Cournoyer]

Hi,

I've noticed build failure events from Cuirass such as:

 --8<---------------cut here---------------start------------->8---
. [   ?: Cuirass                ] Build u-boot-sifive-unmatched.x86_64-linux on 
master is broken.
R. [   ?: Cuirass                ] Build u-boot-am335x-boneblack.x86_64-linux 
on master is broken.
 . [   ?: Cuirass                ] Build u-boot-a20-olinuxino-lime.x86_64-linux 
on master is broken.
 . [   ?: Cuirass                ] Build u-boot-wandboard.powerpc64le-linux on 
master is broken.
 . [   ?: Cuirass                ] Build 
u-boot-qemu-riscv64-smode.aarch64-linux on master is broken.
 . [   ?: Cuirass                ] Build u-boot-rpi-2-efi.powerpc64le-linux on 
master is broken.
 . [   ?: Cuirass                ] Build u-boot-qemu-arm.x86_64-linux on master 
is broken.
 . [   ?: Cuirass                ] Build u-boot-rpi-2.x86_64-linux on master is 
broken.
 --8<---------------cut here---------------end--------------->8---

There are a few more, and they were triggered by applying a patch (that
only touches the patman tool) in this commit:
<https://git.savannah.gnu.org/cgit/guix.git/log/?id=faeebdc37e9137d62e11175e83988f77496a6641>.

The build failures are not captured in the logs, which appear truncated,
e.g. from <https://ci.guix.gnu.org/build/2294052/details> (log empty --
builder: sjd-p9, worker N38vG7OW) or from
<https://ci.guix.gnu.org/build/2294218/details> (builder:
hydra-guix-107, worker 4CWieLpQ) which ends with:

--8<---------------cut here---------------start------------->8---
/tmp/guix-build-gcc-cross-arm-linux-gnueabihf-11.3.0.drv-0/build/./gcc/nm -pg 
_arm_fixsfsi_s.o | gawk 'NF == 3 && $2 !~ /^[UN]$/ && $3 !~ /.*_compat/ && $3 
!~ /.*@.*/ { print "\t.hidden", $3 }' > _arm_fixsfsi.visT
/tmp/guix-build-gcc-cross-arm-linux-gnueabihf-11.3.0.drv-0/build/./gcc/nm -pg 
_arm_fixunssfsi_s.o | gawk 'NF == 3 && $2 !~ /^[UN]$/ && $3 !~ /.*_compat/ && 
$3 !~ /.*@.*/ { print "\t.hidden", $3 }' > _arm_fixunssfsi.visT
mv -f _arm_truncdfsf2.visT _arm_truncdfsf2.vis
/tmp/guix-build-gcc-cross-arm-linux-gnueabihf-11.3.0.drv-0/build/./gcc/nm -pg 
_arm_floatdidf_s.o | gawk 'NF == 3 && $2 !~ /^[UN]$/ && $3 !~ /.*_compat/ && $3 
!~ /.*@.*/ { print "\t.hidden", $3 }' > _arm_floatdidf.visT
mv -f _arm_negsf2.visT _arm_negsf2.vis
mv -f _arm_addsubsf3.visT _arm_addsubsf3.vis
/tmp/guix-build-gcc-cross-arm-linux-gnueabihf-11.3.0.drv-0/build/./gcc/nm -pg 
_arm_floatdisf_s.o | gawk 'NF == 3 && $2 !~ /^[UN]$/ && $3 !~ /.*_compat/ && $3 
!~ /.*@.*/ { print "\t.hidden", $3 }' > _arm_floatdisf.visT
/tmp/guix-build-gcc-cross-arm-linux-gnueabihf-11.3.0.drv-0/build/./gcc/nm -pg 
_arm_floatundi
--8<---------------cut here---------------end--------------->8---

These appear to be some problem with the CI software (probably not
hardware, as many workers across different architectures suffered the
same problem) rather than packaging problems.

I didn't see anything useful in /var/log/cuirass.log or
/var/log/cuirass-remote-worker.log berlin.  I haven't checked the logs
on the builders.

-- 
Thanks,
Maxim