collations

Analysis:
When we have negative index, the value in array_counter[] array is going to
be -1 at some point ( because in case of negative index in json path, the
initial value for a path with negative index is -<size_of_array>, and as we
move forward in array while parsing it and finding path, this value
increments). Since SKIPPED_STEP_MARK, is maximum uint value, it gets
compared to some int value in the array and eventually equates to -1
and messes with path.
Fix:
Make SKIPPED_STEP_MARK maximum of INT32.

buf_read_ahead_linear(): Correct some calculations that were broken
in commit b1ab211d (MDEV-15053).

Thanks to Daniel Black for providing a test case and initial debugging.

Tested by: Matthias Leich

btr_cur_need_opposite_intention(): Check also page_zip_available()
so that we will escalate to exclusive index latch when a non-leaf
page may have to be split further due to ROW_FORMAT=COMPRESSED page
overflow.

Tested by: Matthias Leich

best_access_path(): Simplify the logic for computing fanout.
This fixes up commit 4329ec5d

Command lsof can fail on Debian install.
Revert logic more like old one to make sure that there is no failing
and still does don't boundce on shellcheck.

Debian install scripts if-clauses are not formatted
as they should be. This commit formats Debian
Pre and Post script if-clauses.

Commit fixes several ShellCheck found problems in Debian
Pre- and Postscripts.

Debian script mariadb-server-10.6.postrm contains shellcheck
Fixed problems are:
  https://www.shellcheck.net/wiki/SC2166 -- Prefer [ p ] && [ q ] as [ p -a q...
  https://www.shellcheck.net/wiki/SC2006 -- Use $(...) notation instead of le...
  https://www.shellcheck.net/wiki/SC1091 -- Not following: /usr/share/debconf...

Debian script mariadb-server-10.6.postinst contains shellcheck
Fixed problems are:
  https://www.shellcheck.net/wiki/SC2129 -- Consider using { cmd1; cmd2; } >>...
  https://www.shellcheck.net/wiki/SC1091 -- Not following: /usr/share/debconf...
  https://www.shellcheck.net/wiki/SC1072 -- Expected test to end here (don't ...
  https://www.shellcheck.net/wiki/SC1073 -- Couldn't parse this test expressi...
  https://www.shellcheck.net/wiki/SC1009 -- The mentioned syntax error was in...

Debian script mariadb-server-10.6.preinst contains shellcheck
Fixed problems are:
  https://www.shellcheck.net/wiki/SC2231 -- Quote expansions in this for loop...
  https://www.shellcheck.net/wiki/SC2166 -- Prefer [ p ] && [ q ] as [ p -a q...
  https://www.shellcheck.net/wiki/SC2001 -- See if you can use ${variable//se...
  https://www.shellcheck.net/wiki/SC1091 -- Not following: /usr/share/debconf...
  https://www.shellcheck.net/wiki/SC1007 -- Remove space after = if trying to...

There is several misindentation inside Debian post and pre
installation scripts. False indentation with space
as indent space should be 2 and indentation with tabs.

Derived table creation code would call Field::make_new_field() which would
memcpy the Field object from the source table, including Field::read_stats.

But the temp. table as a whole had table->stats_is_read=false. Which was
correct but not consistent with Field::read_stats and caused an assertion.

Fixed by making sure that Field::read_stats=NULL for fields in the new
temporary (i.e. work) tables.

The problem, introduced in patch for MDEV-26301:

When check_join_cache_usage() decides not to use join buffer, it must
adjust the access method accordingly. For BNL-H joins this means switching
from pseudo-"ref access"(with index=MAX_KEY) to some other access method.

Failing to do this will cause assertions down the line when code that is
not aware of BNL-H will try to initialize index use for ref access with
index=MAX_KEY.

The fix is to follow the regular code path to disable the join buffer for
the join_tab ("goto no_join_cache") instead of just returning from
check_join_cache_usage().

Do not try to decide merge/materialize for derived if it was already decided
(even if it is a view).

RocksDB (in a submodule) has to include <cstdint> to use uint64_t
but it doesn't. Until the submodule is upgraded, let's replace
problematic types with something that's available

select_insert::store_values() must reset
has_value_set bitmap before every row, just like mysql_insert() does.
because ON DUPLICATE KEY UPDATE and triggers modify it

Temporary fix to avoid the server crash.