Bug 2228368

An abidiff check between libloc-0.9.16-2.fc38 and libloc-0.9.16-3.fc38
<https://artifacts.dev.testing-farm.io/d0b37bc4-5366-4d8c-8f70-b4d82e9e12f2/> reports a change in size of time_t type on i686:

  [C] 'function time_t loc_database_created_at(loc_database*)' at database.c:705:1 has some indirect sub-type changes:
    return type changed:
      underlying type 'typedef __time64_t' at types.h:223:1 changed:
        typedef name changed from __time64_t to __time_t at types.h:160:1
        underlying type 'typedef __int64_t' at types.h:47:1 changed:
          entity changed from 'typedef __int64_t' to compatible type 'long int'
            type name changed from 'long long int' to 'long int'
            type size changed from 64 to 32 (in bits)

build.log does not show any difference in compiling src/database.c. All compiler options are the same:

libtool: compile:  gcc -DHAVE_CONFIG_H -I. -include ./config.h -DSYSCONFDIR=\"/etc\" -I./src -Wall -Wchar-subscripts -Wformat-security -Wmissing-declarations -Wmissing-prototypes -Wnested-externs -Wpointer-arith -Wshadow -Wsign-compare -Wstrict-prototypes -Wtype-limits -fno-semantic-interposition -ffunction-sections -fdata-sections -DLIBLOC_PRIVATE -fvisibility=hidden -O2 -flto=auto -ffat-lto-objects -fexceptions -g -grecord-gcc-switches -pipe -Wall -Werror=format-security -Wp,-U_FORTIFY_SOURCE,-D_FORTIFY_SOURCE=3 -Wp,-D_GLIBCXX_ASSERTIONS -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -fstack-protector-strong -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -m32 -march=i686 -mtune=generic -msse2 -mfpmath=sse -mstackrealign -fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection -c src/database.c  -fPIC -DPIC -o src/.libs/libloc_la-database.

A difference in libloc sources is small and does not change time_t in any way. (There a lot of included header files from glibc and from openssl that I did not yet checked.)

A difference in root.log shows these upgrades:

glibc-headers from x86-2.36.9000-19.fc38 to 2.37-4.fc38
openssl-devel from 3.0.7-3.fc38 to 3.0.9-1.fc38

I verified with "print sizeof(time_t)" when loading the libraries in GDB that the size indeed changed from 8 to 4 bytes. I checked also the very fist libloc build, libloc-0.9.15-2.fc38.i686, that used to be 8 bytes.

I will try to reproduce the regression in mock environment and to minimize the reproducer, but my preliminary suspect is glibc.