DOUBLE_PREC_HB histogram produces wrong estimates for COMPRESSED columns

--source include/have_sequence.inc

CREATE TABLE t (a VARCHAR(16) COMPRESSED COLLATE ucs2_sinhala_ci);

INSERT INTO t SELECT REPEAT(CHR(seq%26+97),16) FROM seq_1_to_100;

set histogram_type= DOUBLE_PREC_HB;

analyze table t persistent for all;

ANALYZE FORMAT=JSON SELECT a FROM t WHERE 'c' BETWEEN 'l' AND a ;

set histogram_type= JSON_HB;

analyze table t persistent for all;

ANALYZE FORMAT=JSON SELECT a FROM t WHERE 'c' BETWEEN 'l' AND a ;

# Cleanup

DROP TABLE t;

        "table": {

          "table_name": "t",

          "access_type": "ALL",

          "r_loops": 1,

          "rows": 100,

          "r_rows": 100,

          "r_table_time_ms": 0.006661524,

          "r_other_time_ms": 0.006350519,

          "filtered": 3.846199989,

          "r_filtered": 0,

          "attached_condition": "'c' between 'l' and t.a"

        }

        "table": {

          "table_name": "t",

          "access_type": "ALL",

          "r_loops": 1,

          "rows": 100,

          "r_rows": 100,

          "r_table_time_ms": 0.005778671,

          "r_other_time_ms": 0.005598088,

          "filtered": 93,

          "r_filtered": 0,

          "attached_condition": "'c' between 'l' and t.a"

        }