Autocomplete of users incorrectly escaping some characters

What does this MR do and why?

Refactor markdown escape processing (!108376 - merged) introduced changes in how markdown escaped characters were converted. Which broke how escaped characters in references are handled.

For example, when searching for @bprescott_, the autocomplete finds the correct person, but outputs @bprescott\_. This has the effect of our username reference filter to detect only @bprescott.

We now detect additional circumstances when an escaped character, such as <span data-escaped-character>~</span>, should be converted back to ~ or left as is.

How to set up and validate locally

Create some (or all) of the following labels:

c_bug_
d _bug_
e*bug*
f *bug*
f *bug*
g`bug`
h `bug

The following markdown should then render properly when pasted into a comment field and previewed

~c_bug\_

Label ~c_bug\_ and _more_ text

~"d \_bug\_"

~"e\*bug\*"

~"f \*bug\*"

Label ~"f \*bug\*" **with** more text

~"g\`bug\`"

~"h \`bug\`"

MR acceptance checklist

This checklist encourages us to confirm any changes have been analyzed to reduce risks in quality, performance, reliability, security, and maintainability.

• I have evaluated the MR acceptance checklist for this MR.

Related to #389197 (closed)