'\u' means 'u'
TFAE: \n, \12, \012, \uA, \u0A, \u00A, \u000A, \x0A (but not \xA)
'\u000AA' means '\nA'
'\xA' means 'xA'
A backslash followed by a newline character (vs escape) means '\\'
    Works as a pattern but not in a string
A backslash followed by a carriage return character (vs newline) means '\uD'
    Works in patterns and in strings
A backslash followed by a space character means ' '
'\7' means '\u0007' (vs '7')
'\777' means '\?7' (i.e. \77 followed by 7)

Summary:

\u codes can have 1 to 4 hex digits (longest match)
\x codes have exactly 2 hex digits
Octal codes can have 1 to 3 digits (first digit <= 3) (longest match)
Escaped newline (line feed) results in a backslash (doesn't work within quotes)
Escaped nrtfb work as usual
Escaped other character (including carriage return and space) results in that character
Octal codes take precedence over escaped digits

Confirmed in LexScan.flex
Note: default case is matched as \\. (should probably be \\ (. | \n))
LexScan.flex appears to disallow \u codes with fewer than 4 digits and all \x codes in strings
Not clear on why backslash line feed is accepted