MDEV-37530 fixes

[ParadoxV5]

Fix issues in MDEV-37530 (#4430) found when releasing v12.3.0 (preview).

Use array, not std::initializer_list

• The Jira issue number for this PR is: MDEV-38410

Some environments appear not to retain the backing array of a static std::initializer_list in the MDEV-37530 release candidate, and eventually crash when reading overwritten data.

This commit resolves the stealth issue by reverting to conventional arrays, while maintaining convenience through deductive overloads.

Avoid C++17’s <charconv> entirely

• Some of our platforms (namely SUSE 15, which uses GCC 7.5) support C++17 syntaxes, but not all libraries, <charconv> among those.
• We are yet to figure out what to do with our 10 choices of string→int converters, so now is not the time to settle preferences on <charconv> or whatever superset that covers all of their cases.

Warning fix: make int→uchar explicit

Arithmetics start from int and so auto-promotes.

Use boolean OR, not bitwise OR

Unlike Boolean operators, the order of operands of bit-logic operators is not guaranteed.

DBUG_ASSERT() uses the boolean value and does not show the integer value anyway.

PR quality check

• This is a new feature or a refactoring, and the PR is based on the main branch.
• This is a bug fix, and the PR is based on the earliest maintained branch in which the bug can be reproduced.
• I checked the CODING_STANDARDS.md file and my PR conforms to this where appropriate.
• For any trivial modifications to the PR, I am ok with the reviewer making the changes themselves.

[CLAassistant]

All committers have signed the CLA.

[ParadoxV5]

Yeah PR view doesn’t know that MDEV-37530 is a PR based on main

[ParadoxV5]

Previous discussion: #3724 (comment)

[vuvova]

Looks ok. I'd prefer a more reusable C++ str->int approach, see inline comment.

And let Brandon review the std::initializer_list changes.

[vuvova]

may be (quoting from my yet-unpushed commit)

template<typename T> inline char *str2int(const char *s, size_t l, int base, T *val)
{
  char *end= const_cast<char*>(s) + l;
  int err;
  longlong v= my_strntoll_8bit(&my_charset_latin1, s, l, base, &end, &err);
  if (!err && static_cast<longlong>(*val= static_cast<T>(v)) != v)
    err= EDOM;
  return (errno= err) ? NULL : end;
}

template<> inline char *str2int<ulonglong>(const char *s, size_t l, int base, ulonglong *val)
{
  char *end= const_cast<char*>(s) + l;
  int err;
  *val= my_strntoull_8bit(&my_charset_latin1, s, l, base, &end, &err);
  return (errno= err) ? NULL : end;
}

it's type safe and universally usable (could be in a common header). and doesn't add a new str->int, but wraps existing one.
modeled after str2int (that I want to get rid of), that's why errno. may be not the best approach.