Improve numeric_power() tests for large integer powers.

Two of the tests added by 4dd5ce2fd fail on buildfarm member
castoroides, though it's not clear why. Improve the tests to report
the actual values produced, if they're not what was expected.

Apply to v11 only for now, until it's clearer what's going on.

src/test/regress/expected/numeric.out

@@ -1668,10 +1668,10 @@ select 1.000000000123 ^ (-2147483648);
  0.7678656556403084
 (1 row)
 
-select 0.9999999999 ^ 23300000000000 = 0 as rounds_to_zero;
+select coalesce(nullif(0.9999999999 ^ 23300000000000, 0), 0) as rounds_to_zero;
  rounds_to_zero 
 ----------------
- t
+              0
 (1 row)
 
 -- cases that used to error out
@@ -1687,10 +1687,10 @@ select 0.5678 ^ (-85);
  782333637740774446257.7719390061997396
 (1 row)
 
-select 0.9999999999 ^ 70000000000000 = 0 as underflows;
+select coalesce(nullif(0.9999999999 ^ 70000000000000, 0), 0) as underflows;
  underflows 
 ------------
- t
+          0
 (1 row)
 
 -- negative base to integer powers

src/test/regress/sql/numeric.sql

@@ -907,12 +907,12 @@ select 3.789 ^ 35;
 select 1.2 ^ 345;
 select 0.12 ^ (-20);
 select 1.000000000123 ^ (-2147483648);
-select 0.9999999999 ^ 23300000000000 = 0 as rounds_to_zero;
+select coalesce(nullif(0.9999999999 ^ 23300000000000, 0), 0) as rounds_to_zero;
 
 -- cases that used to error out
 select 0.12 ^ (-25);
 select 0.5678 ^ (-85);
-select 0.9999999999 ^ 70000000000000 = 0 as underflows;
+select coalesce(nullif(0.9999999999 ^ 70000000000000, 0), 0) as underflows;
 
 -- negative base to integer powers
 select (-1.0) ^ 2147483646;