When computing bucket_capacity= records/histogram->get_width(), round
the value UP, not down.

Part#3:
- make json_escape() return different errors on conversion error
  and on out-of-space condition.
- Make histogram code handle conversion errors.

Fix the description

Change it to LONGBLOB.
Also, update_statistics_for_table() should not "swallow" an error
from open_stat_tables.

.. part#2: correctly pass the charset to JSON [un]escape functions

MDEV-26595: ASAN use-after-poison my_strnxfrm_simple_internal / Histogram_json_hb::range_selectivity

Add testcase

Item_func_decode_histogram::val_str should correctly set null_value
when "decoding" JSON histogram.

Correctly handle empty string when [un]escaping JSON

Escape values when serializing to JSON. Un-escape when reading back.

Do not put Histogram objects on MEM_ROOT at all

Provide buffer of sufficient size.

Aslo add more test coverage

- Use String::c_ptr_safe() instead of String::c_ptr
- Do proper datatype conversions in Histogram_json_hb::parse
- Remove Histogram_json_hb::Bucket::end_value. Introduce
  get_end_value() instead.

Basic ideas:
1. Store "popular" values in their own buckets.
2. Also store ndv (Number of Distinct Values) in each bucket.

Because of #1, the buckets are now variable-size, so store the size in
each bucket.

Adjust selectivity estimation functions accordingly.

- Fix bad tests in statistics_json test: make them meaningful and make them
  work on windows
- Fix analyze_debug.test: correctly handle errors during ANALYZE

Remove Histogram_*::is_available(), it is not applicable anymore.
Fix compilation on Windows

Factor the code that updates count, count_distinct,
count_distinct_single_occurrence into class Basic_stats_collector

Change from Histogram_builder and its descendant Histogram_builder_json
to  Histogram_builder (the interface), and Histogram_binary_builder,
Histogram_json_builder.

In Histogram_json_builder, do not forget to collect the right bound
of the right-most bucket.