FTBFS on s390x and armhf

[Andreas Hasenack]

a) s390x FTBFS: https://launchpad.net/ubuntu/+source/blender/3.4.1+dfsg-2ubuntu1/+build/26480653

In file included from /<<PKGBUILDDIR>>/intern/cycles/kernel/../util/half.h:7,
                 from /<<PKGBUILDDIR>>/intern/cycles/kernel/../kernel/device/cpu/kernel.h:8,
                 from /<<PKGBUILDDIR>>/intern/cycles/kernel/device/cpu/kernel.cpp:46:
/<<PKGBUILDDIR>>/intern/cycles/kernel/../util/math.h: In function ‘bool ccl::isfinite_safe(float)’:
/<<PKGBUILDDIR>>/intern/cycles/kernel/../util/math.h:357:55: error: implicit conversion from ‘float’ to ‘double’ to match other operand of binary expression [-Werror=double-promotion]
  357 | return (f == f) && (x == 0 || x == (1u << 31) || (f != 2.0f * f)) && !((x << 1) > 0xff000000u);
      | ~~^~~~~~~~~~~

I checked upstream git repo and that code is still the same in the main branch. I also didn't find any open or closed issues dealing with this type of build error. This warrants more investigation, maybe an upstream issue.

b) armhf FTBFS: https://launchpad.net/ubuntu/+source/blender/3.4.1+dfsg-2ubuntu1/+build/26480650
/<<PKGBUILDDIR>>/source/blender/blenlib/intern/string.c: In function ‘BLI_str_format_int_grouped’:
/<<PKGBUILDDIR>>/source/blender/blenlib/intern/string.c:1128:23: error: conversion to ‘int’ from ‘size_t’ {aka ‘unsigned int’} may change the sign of the result [-Werror=sign-conversion]
 1128 | const int num_len = BLI_snprintf(src, sizeof(src), "%d", num);
      | ^~~~~~~~~~~~
/<<PKGBUILDDIR>>/source/blender/blenlib/intern/string.c: In function ‘BLI_str_format_uint64_grouped’:
/<<PKGBUILDDIR>>/source/blender/blenlib/intern/string.c:1138:23: error: conversion to ‘int’ from ‘size_t’ {aka ‘unsigned int’} may change the sign of the result [-Werror=sign-conversion]
 1138 | const int num_len = BLI_snprintf(src, sizeof(src), "%" PRIu64 "", num);
      | ^~~~~~~~~~~~

[Rik Mills (rikmills)]

FTBFS on 32bit is expected, as far as I am aware and won't be fixed. Build on s390x is fixed in the latest mantic uploads. So marking this as fixed.