Source: clickhouse Version: 18.16.1+ds-5 Severity: serious Tags: ftbfs Hi,
clickhouse fails to build from source in bullseye/sid: https://tests.reproducible-builds.org/debian/rb-pkg/unstable/amd64/clickhouse.html /build/1st/clickhouse-18.16.1+ds/dbms/src/Functions/intDiv.cpp:55:75: error: no match for 'operator/' (operand types are '__m128i' {aka '__vector(2) long long int'} and 'libdivide::divider<unsigned int, 0>') 55 | _mm_loadu_si128(reinterpret_cast<const __m128i *>(a_pos)) / divider); | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~ In file included from /build/1st/clickhouse-18.16.1+ds/dbms/src/Functions/intDiv.cpp:9: /usr/include/libdivide.h:2039:3: note: candidate: 'template<class T, int ALGO> T libdivide::operator/(T, const libdivide::divider<T, ALGO>&)' 2039 | T operator/(T n, const divider<T, ALGO>& div) { | ^~~~~~~~ /usr/include/libdivide.h:2039:3: note: template argument deduction/substitution failed: /build/1st/clickhouse-18.16.1+ds/dbms/src/Functions/intDiv.cpp:55:75: note: deduced conflicting types for parameter 'T' ('__vector(2) long long int' and 'unsigned int') 55 | _mm_loadu_si128(reinterpret_cast<const __m128i *>(a_pos)) / divider); | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~ I was able to reproduce this on my own system as well. Would be great if you could fix this soon as it's blocking the poco transition. One simple workaround would be to remove the #if __SSE2__ block (line 48-60) in intDiv.cpp as it's only a SSE2 optimization and the generic case is below. Cheers Jochen -- System Information: Debian Release: bullseye/sid APT prefers unstable APT policy: (500, 'unstable') Architecture: amd64 (x86_64) Foreign Architectures: i386 Kernel: Linux 5.4.0-3-amd64 (SMP w/8 CPU cores) Locale: LANG=de_DE.UTF-8, LC_CTYPE=de_DE.UTF-8 (charmap=UTF-8), LANGUAGE=de_DE.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Init: systemd (via /run/systemd/system) LSM: AppArmor: enabled
