I found the problem that clang compiler introduces warnings when building PostgreSQL. Attached patch fixes it.
=== Compiler version === clang version 10.0.0-svn372772-1~exp1+0~20190924181208.2504~1.gbpb209ff (trunk)
Older versions of clang may not generate this warning.
=== Warning ===
timestamp.c:3236:22: warning: implicit conversion from 'long' to 'double' changes value from 9223372036854775807 to 9223372036854775808 [-Wimplicit-int-float-conversion] if (result_double > PG_INT64_MAX || result_double < PG_INT64_MIN) ~ ^~~~~~~~~~~~ ../../../../src/include/c.h:444:22: note: expanded from macro 'PG_INT64_MAX' #define PG_INT64_MAX INT64CONST(0x7FFFFFFFFFFFFFFF) ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ../../../../src/include/c.h:381:25: note: expanded from macro 'INT64CONST' #define INT64CONST(x) (x##L) ^~~~ <scratch space>:234:1: note: expanded from here 0x7FFFFFFFFFFFFFFFL ^~~~~~~~~~~~~~~~~~~ 1 warning generated. pgbench.c:1657:30: warning: implicit conversion from 'long' to 'double' changes value from 9223372036854775807 to 9223372036854775808 [-Wimplicit-int-float-conversion] if (dval < PG_INT64_MIN || PG_INT64_MAX < dval) ^~~~~~~~~~~~ ~ ../../../src/include/c.h:444:22: note: expanded from macro 'PG_INT64_MAX' #define PG_INT64_MAX INT64CONST(0x7FFFFFFFFFFFFFFF) ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ../../../src/include/c.h:381:25: note: expanded from macro 'INT64CONST' #define INT64CONST(x) (x##L) ^~~~ <scratch space>:252:1: note: expanded from here 0x7FFFFFFFFFFFFFFFL ^~~~~~~~~~~~~~~~~~~ 1 warning generated.
===
This warning is due to implicit conversion from PG_INT64_MAX to double, which drops the precision as described in the warning. This drop is not a problem in this case, but we have to get rid of useless warnings. Attached patch casts PG_INT64_MAX explicitly.
